Jump to content

Getting power from ethanol actually does not worth it and my feedback on it


Recommended Posts

I set my power control station to use lead for tune-up microchips and it seems to work. However, I noticed my dupes stop tuning up the generator as soon as the smart battery disables the automation grid, *and* the dupes don't resume where they left off when they come back to the task. So that means they are forever only half tuning up generators and wasting their time. That seems like a bug and makes me inclined to not seriously consider generator tune-up when evaluating any build.

Link to comment
Share on other sites

I would be careful about lead.  Early mid game, it is plentiful and cheap.  However, as far as I can tell, it is finite on the asteroid.  And lead has some nice properties that makes it good for certain conditions.  For example, I have a tamed volcano and I used a small amount of molten lead for my thermal sensor.  It has a good thermal conductivity and low specific heat, which means it comes to temperature faster than most gases (I believe).  However, you do need to make sure your area says above a certain temperature.  My point is that you don't want to run out of lead. 

If you have a metal volcano, I would use that do make power control chips.

Link to comment
Share on other sites

17 hours ago, PhailRaptor said:

If true, it's likely a bug.  Lead is classified as a Refined Metal, and Refined Metals are what Dupes use to manufacture at the Power Station to use in Tune Ups.

I use lead exclusively for tune-ups.

Link to comment
Share on other sites

19 hours ago, DarkMoge said:

Right, limited... The falling from the sky makes it infinite.

As far as I am aware, lead cant be used to tune generators.

It can though. All my tune ups are done with lead, works just fine.

Also, there's often an abundance of it - like more than you'll need ever even if you're a sadist that plays to 5k+ cycles, you wont run out of lead.

Then, as you correctly alluded to - space. It's plentiful.

THEN there's also every other metal mined out of your asteroid, which unless you're pissing it up the wall, should last a very, very long time.

Link to comment
Share on other sites

18 hours ago, Snotfroth said:

I set my power control station to use lead for tune-up microchips and it seems to work. However, I noticed my dupes stop tuning up the generator as soon as the smart battery disables the automation grid, *and* the dupes don't resume where they left off when they come back to the task. So that means they are forever only half tuning up generators and wasting their time. That seems like a bug and makes me inclined to not seriously consider generator tune-up when evaluating any build.

I think this is a similar bug to dupes abandoning coal delivery jobs when the generator gets disabled by automation.  I'm fairly sure they didn't use to do this, so it's a new problem.

Link to comment
Share on other sites

I use pips to plant tons of trees and start from there. So now I have free wood!

Next, dig up near surface and build your processing and powerplant there. Set it all up, make nice chimney.

Profit!

Enjoy your free PW+power.

On 25/08/2019 at 2:21 PM, blash365 said:

I have 4 Petroleum generators and they still overheat fairly quickly. My co2 is not that cold though. How do you get it to 40F?

 

And how do you cool your steam turbine down?

I exploit that lumber spawns at 20C I think and i just have my stills and generators take a swim in it and let logs soak as much heat as they can. Lumber if routed properly in termally conductive areas can keep your ethanol distillery at 40-50C without any aquatuners or steam engines.

Link to comment
Share on other sites

Not sure why folks think you can't tune with lead... My entire electrical grid outside of temperature-extreme zones is nothing but lead and all the applicable power stations are set to tune with lead and only lead... and the last map, that's all they were ever set to use.  My stuff got tuned, so clearly you can tune with lead.

 Also, @Snotfroth, that's an unfortunate issue,  I don't remember if they fixed this or not, but at one point the buff timer ran regardless of if the generator was in use or not, so tuning generators that weren't used meant you were basically just destroying metal.  If I'm remembering correctly, the behavior was modified to not tune shut-down generators to prevent this.   Honestly, I think the better solution would have been to allow dupes to tune generators regardless of if they were running and base the buff duration entirely on up-time and not duration starting at application... a generator that stays off should either never become untuned, or at the very least, should have to sit there a lot longer than the default duration before the buff expires.  But... I haven't checked the buff to see if they changed it to that or not, sorry.

 

23 hours ago, DarkMoge said:

*brevity snip*

 Unfortunately, I'm going to have to disagree with your numbers. Here's why.

Universal requirements (common to both sides, assumes default schedule)
Dupe, 1
Input: 100 g/s O2, 1kcal, 8.33 g/s H2O (assuming flush toilet, one use)
Output: 450 dupe-seconds (d-s) labor, 19.5 g/s PH2O (assuming a flush toilet, one use) (Note, I'm ignoring dupe CO2 output for sieving purposes, it's negligible, and I'm pretty much done extending this spreadsheet to account for every little thing.)

So let's get translating all that to water and power, shall we? 

 Since I'm applying this to both sides, it doesn't matter what food route I choose to go. Both sides are equally penalized or benefit equally.  So 1kcal is 3.74 bristle blossoms, which eat 33.33g/s H2O.  The pumping costs for this water is 0.8W

Total food cost: 0.8 W, 33.33g/s H2O

 in order to supply that 100g/s, you end up running an electrolyzer at 11.26% capacity, which outputs your 100g/s O2 and 12.6 g/s H2 while costing 13.51 W and 112.6 g/s H2O.  That 12.6 g/s H2 runs a H2 Gen at 12.6% capacity which outputs 100.9W..  A lazy pumping set-up for the electrolyzer that simply will not back-up until your output vents do uses 3 pumps and 3 filters (Note: These numbers are derived from the fact that for 20 dupes you roll 2 electros with 3 pairs of pumps and 2 filters per pair... or 3 filters per electro.  Yes it's lazy.  I never said I optimized this part. The savings are minimal and seeing as both cycles get hit with this so that doesn't matter. Using the same utilization as the electrolyzer (11.26%), 3 pumps and 3 filters cost 121.62 W)

Total O2 Cost: -13.51+100.9-121.62=-34.23 W

PH2O sieving: The toilet water will run 0.39% of a sieve, eating 0.468 W.while getting you 259g of dirt.

Grand total for 1 dupe
-34.23 W (O2)-0.8W(food)-0.468W (Sieve) = -35.03W
-8.33 g/s (bathroom) -112.61 g/s (O2)+19.5 g/s (Sieving)-124.84 g/s (food)= -226.29 g/s H2O (or ~135.774 kg H2O / cycle)
468g dirt/cycle
450 d-s labor

Now let's apply those numbers to the wheel cycle.  Since we're not monsters, we don't lock the dupe in the same room with the wheel.  All tasks assume a 10-second travel penalty, so since this is his only task, we lose 10 d-s to travel, leaving 440 d-s to operate the manual generator.  The manual generator creates 0.66W/d-s, so you get 293.33W.

Total output
293.33 - 35.498 = 257.8 W
468g dirt

Total Input: 135.774 kg H2O

Wheel Cycle Efficiency (Output/Input) -> 257.8W/135.774 kg H2O -> 1.89 W per KG H2O per cycle



Ethanol cycle:  (Hold on to your belts, this is going to be as easy to follow as it was transcribing it off the spreadsheet in some sort of semi-sane order, followed by trying to put running net totals in-line to make it easier to follow...)
Same dupe-based input... Units that follow are kg/cycle to simplify things

8 trees
Input: 560 kg PH2O, 80 kg dirt, 120 d-s labor (3 d-s per branch * 5 branch/tree*8 trees)
Output: 2666.67kg lumber,

This enables 4 distillers:
Input: 2400 kg lumber (Net 266.67 kg), 960W
Output: 800 kg Pdirt, 400 kg CO2, 1200 kg ethanol

Which in turn enables 1 petro gen
Input: 1200 kg ethanol (net 0)
Outputs: 450 kg PH2O (net -110 so far), 300 kg CO2 (net 700), 2000W.(net 1040)

That 800kg Pdirt gets fed to 4 sage hatches
Input: 600kg pdirt, 32 d-s labor (5 second groom + 3 seconds jerk hatch ignore time)x4 hatches
Output: 600kg coal, 200 kg pdirt pass-through to composts

Turn the coal around to a coal gen:
Input: 600 kg coal (net 0)
Output: 600W (net 1640), 12 kg CO2 (net 712)

Speaking of power.- excess lumber -> 37.03% Wood burner
Input: 266.67 kg lumber (Net 0)
Output: 111.11 W (net 1751.11), 37.77kg CO2 (net 749.77)

Fire up the scrubbers! 4.17 of them to be more or less precise.
Input: 749.77 kg CO2 (net 0) 2499.25 kg H2O, 499.85W (net 1251.26)
Output: 2499.25 kg PH2O (net 2389.25)

Speaking of sieves, we're running at 80.03% capacity (Don't forget the 11.7kg PH2O from the dupe)
Input ~2400kg PH2O (net 0), 96.03W (net 1155.23)
Output ~2400kg H2O (net -100), 96 kg Pdirt (net 296)

Composts - 4.93 of them (But we're rounding the labor up)
Input: ~296 kg Pdirt, 25 d-s (5 d-s per compost)
Output: ~296 kg dirt

Liquid Pump - All PH2O input is free, 560kg pumped to trees, 2400 pumped to sieves, 49.3% pump capacity
Input: 49.3% of 240W ->  118.4W (net 1036.83)

That's just about all of it... except labor.  Now, we started with 450 d-s of labor, lost 120 to the trees, 25 to the composts, 32 to hatches, and 30 to travel penalties, so total labor used so far: 207 d-s, which leaves us 243 d-s to do with as we please.  What shall we do with it?  This is supposed to be a fair comparison, so  Isn't it obvious?

The Wheel.
Input: 243 d-s labor
output: 162 W. (net 1198.83)

Now, after we go through and add everything up, what does the ethanol cycle give us with this configuration?  Other than confusion?
After knocking off the rounding errors and adding the universal dupe costs I just realized I hadn't included, we get:
Inputs: (Net -100 - 135.77) =  235.77 kg H2O
Outputs: (Net 216+0.4 = )216.4 kg dirt, (Net 1198.83- 35.03=)1163.8 W

Ethanol Cycle Efficiency - 1163.8 W / 235.77 kg H2O -> 4.94 W / kg H2O


 Now, I'm more than happy to admit there's probably some stuff that looks like math errors up there given what a pain it was to transcribe,, but I'm confident the running totals for the ethanol cycle and the wheel cycle final numbers are correct... So unless I completely brainfarted, it should all be there...

 So there you have it, 1.89 W/kg < 4.94 W/ kg.  Ethanol cycle wins.  And now, after having spent more hours than anyone sanely would compiling and transcribing all this so I could post it at 6AM, I'm going to bed.  :)

Link to comment
Share on other sites

4 hours ago, storm6436 said:

lazy pumping set-up for the electrolyzer that simply will not back-up until your output vents do uses 3 pumps and 3 filters

wow, I could never imagined a setup that inefficient, the worst I could think of is 2 pumps with 1 filter. There are many different ways to optimize an electrolyzer. If optimized, electrolyzers could even be used to delete 150 kDTU of heat per second while also giving out oxygen at 37C, but its really hard, because of how inconsistent they can be in their work time. The setup that I usually use is 1 pump which is automized to only ever pump hydrogen, 0 filters per electrolyzer which ends up working at 50%-70% of its maximum capacity. But it took me quite some time to come up with a design like that...

4 hours ago, storm6436 said:

Yes it's lazy.  I never said I optimized this part.

Trying to use numbers for 3 pumps + 3 filters is way too unreasonable. Its like if I added pumps with filters to my ethanol chain calculation saying that I need to 1 gas pump with 1 filter per 1 distillery and 2 more pumps with filter for generator itself. Which would add 6 pumps and 5 filters to the setup, because I am too lazy.

4 hours ago, storm6436 said:

Total O2 Cost: -13.51+100.9-121.62=-34.23 W

So, I do not agree with that, because oxygen production is power positive rather than power negative.

4 hours ago, storm6436 said:

Since I'm applying this to both sides, it doesn't matter what food route I choose to go. Both sides are equally penalized or benefit equally.  So 1kcal is 3.74 bristle blossoms, which eat 33.33g/s H2O.  The pumping costs for this water is 0.8W

it does matter, because using food that is made out of water screws the final ratio of how much water the setup requires.
x/y is not equal to (x+z)/(y+z), but you treated it as if its equal.
In your calculations, you are comparing 1 dupe versus 1 dupe + ethanol chain... 1 dupe would normally require 67.5 kg of water/cycle and 1 dupe + ethanol chain would require 177.5 kg of water/cycle. Second is 263% of the first. Than, you screw the calculation by adding 3.75 bristle blossoms which adds 75 kg of water/cycle to both side. 1 dupe side is penalized much harder, because you more than double how much water 1 dupe needs. and you end up with a ratio of 252.5/142.5=1.77 which screws with your final calculation.
Also, the fact of having 1 dupe on both sides is kind of weird. In my calculations, I am using the setup where you would have the same number of dupes to consume same amount of water as ethanol chain.
Why did you pick bristle blossoms? why not frost burgers? or, maybe, meal lice? You would get completely different numbers for the food cost. Generally, I did not include food cost in my calculation, because food is usually abundant and there are plenty of setups where you can start making food with practically no cost and there are setups where food is very expensive (frost burgers)

4 hours ago, storm6436 said:

Input: 600kg pdirt, 32 d-s labor (5 second groom + 3 seconds jerk hatch ignore time)x4 hatches

since when? I made sure to make a test world and try to build a setup with minimal travel times to test it... My first idea was, maybe, just maybe, hatch ranch take less work than other ranches.
ranches.thumb.PNG.e5520d7196c5a031e7a44fcf372d6bf5.PNG

500 seconds of work time... According to your numbers, Camille should be able to handle 500/8=62.5 hatches which almost 8 full ranches... But during the live test, she was able to handle only 19 hatches in those 500 seconds... Well, we can also add the ranching bonus of 110%. 19*2.1=39.9 hatches. Though, in more realistic base, we would probably have a rancher that handles 32 hatches or maybe even less, depending on how much downtime we give them, in your calculations, you give them 3 hours downtime and 3 hours sleep time for 450 sec worktime as an example. Anyway, handling 1 hatch is about 12.5 sec....

But honestly, you could make more efficient setup and just get yourself 20 ungroomed hatches within an enclosed room and they would give as much coal as 4 groomed hatches.


With so many wrong number, your all other numbers end up being wrong too.

 

4 hours ago, storm6436 said:

Wheel Cycle Efficiency (Output/Input) -> 257.8W/135.774 kg H2O -> 1.89 W per KG H2O per cycle

1 dupe:
input: 67.5 kg of clean water. 1 kcal.
output: 6.7 kg of polluted water, negligible amount of CO2. At least 450s of worktime, can be upped to 500s of worktime or even 525s.

I have calculated for 475s worktime, your 1 dupe nets 394W, not 257.8W... It could be higher.
Even if we ignore the 6.7 kg of water return.

394W/67.5 kg of H2O -> 5.83 W per kg H2O per cycle.

Though, your number for ethanol efficiency is also wrong, though, I am not willing to try an recalculate it again... Not to mention, you do not include the cost of cooling the system down... Even if you cool down your distillers and generators with lumber from trees, you still need to cool down dirt and water or your trees will overheat and you will no longer be able to cool down things with lumber.
 

4 hours ago, storm6436 said:

Ethanol cycle wins.  And now, after having spent more hours than anyone sanely would compiling and transcribing all this so I could post it at 6AM, I'm going to bed.

you probably should not do calculations with a sleepy head. I do not mean to be offensive, but doing lazy calculations as well as doing calculations that are mathematically incorrect, leads to false results.

4 hours ago, storm6436 said:

Not sure why folks think you can't tune with lead

Well, because, I have not tested it for the latest update. I did a test now, it does seem to work with lead which makes it more appealing. Takes about 150s to tune a generator.

4 hours ago, storm6436 said:

Now, I'm more than happy to admit there's probably some stuff that looks like math errors up there given what a pain it was to transcribe,, but I'm confident the running totals for the ethanol cycle and the wheel cycle final numbers are correct... So unless I completely brainfarted, it should all be there...

which you did. x/y is not equal to (x+z)/(y+z)

Link to comment
Share on other sites

11 hours ago, DarkMoge said:

wow, I could never imagined a setup that inefficient, the worst I could think of is 2 pumps with 1 filter. There are many different ways to optimize an electrolyzer. If optimized, electrolyzers could even be used to delete 150 kDTU of heat per second while also giving out oxygen at 37C, but its really hard, because of how inconsistent they can be in their work time. The setup that I usually use is 1 pump which is automized to only ever pump hydrogen, 0 filters per electrolyzer which ends up working at 50%-70% of its maximum capacity. But it took me quite some time to come up with a design like that...

 

 Probably because those two electrolyzers in my setup are sitting to either side of an AETN that's sitting in a hydrogen chamber with metal walls.  There's also sufficient open space in the hydrogen to add wheezeworts + automation for phosphorite delivery in case I need to twiddle with the temperature of outgoing air.  It's basically a giant convection-based air conditioner.  All externally facing surfaces are insulated igneous and the doors have igneous tempshift plates to slow the leakage of cold. Total time to ponder this?  About 30 seconds.  Pump pairs are done because each pair can saturate the line they're feeding.  Since 2 pumps on an electrolyzer will manage, on average ~95-98% uptime due to gas flux, the third ensures full output of H2.  Without that third pump, you do not have 100% uptime, and given that electros preferentially ignore H2 in favor of O2, if you want close to the rated output, 3x*2x1) is the barest minimum setup that ensures clean H2 to your generator and the AETN while ensuring the electrolyzers can literally never back up without power or output failure. I  went 3x2x2 only because I got tired of randomly dropping random gasses into my base because a dupe got bored and decided to stand in a doorway and chill.  Sure, I could lock the doors, or just replace them with more insulated tiles, etc... but *shrug* Most of my designs lean toward either 'needlessly realistic' or enabling my particular flavor of selective OCD lazy.  It's an air conditioner, not a LOX room, volcano containment vault, or a sour gas condenser, A small decrease in efficiency there is largely ignorable, especially when I do sometimes tweak or otherwise fiddle with it for fun.  The doors enable that.

Oh, Full disclosure: The physicist in me is a bit unhappy that the units I used were incorrect.  And by that, I mean time operated * wattage = J produced, not W... but since I kept it consistent, the comparison doesn't change, just the units being compared, so whether it's W or J per kg doesn't matter in the intermediate calculations when we're talking average over a cycle, which means J/kg at the end is really J/(kg*cycle) -> j/kg*600sec -> W/kg   And since that error was done on both sides, it does, in fact, divide out and leave the ratios between the two unchanged.   Given it was long after midnight and the results are unchanged, I think it's excusable.

 

11 hours ago, DarkMoge said:

Trying to use numbers for 3 pumps + 3 filters is way too unreasonable. Its like if I added pumps with filters to my ethanol chain calculation saying that I need to 1 gas pump with 1 filter per 1 distillery and 2 more pumps with filter for generator itself. Which would add 6 pumps and 5 filters to the setup, because I am too lazy.

 Except... why would you use pumps at all with a distillery?  You don't pump any self-sorting gas unless it's trapped in a pocket you cannot protect otherwise.  Either let it spill into a common gathering cistern where you're already pushing your CO2 to for scrubbing or toss two scrubbers under the set-up itself.  Well, option 3 still involves the common cistern, but an automated pump to move the CO2 to slicksters, but that's slickster dependent and requires 1 pump to move literally your entire base's output of free CO2, not just one facility.

 

11 hours ago, DarkMoge said:

 

So, I do not agree with that, because oxygen production is power positive rather than power negative.

 

O2 generation the way I do it?  Is certainly power positive... when you tune, otherwise, it basically just pays for itself 93.75% of the time.  Is it inefficient? Sure. Does it satisfy my odd OCD that irks me when things back up that shoulldn't?  Yeah.  Yeah it does.  The whole point behind the setup I posted was that it was *completely* an apples to apples comparison.  Both sides take the same punch to the face in terms of water consumption from said dupe, etc.  I think you're sorta missing the point.

 

11 hours ago, DarkMoge said:

it does matter, because using food that is made out of water screws the final ratio of how much water the setup requires.

x/y is not equal to (x+z)/(y+z), but you treated it as if its equal.
In your calculations, you are comparing 1 dupe versus 1 dupe + ethanol chain... 1 dupe would normally require 67.5 kg of water/cycle and 1 dupe + ethanol chain would require 177.5 kg of water/cycle. Second is 263% of the first. Than, you screw the calculation by adding 3.75 bristle blossoms which adds 75 kg of water/cycle to both side. 1 dupe side is penalized much harder, because you more than double how much water 1 dupe needs. and you end up with a ratio of 252.5/142.5=1.77 which screws with your final calculation.
Also, the fact of having 1 dupe on both sides is kind of weird. In my calculations, I am using the setup where you would have the same number of dupes to consume same amount of water as ethanol chain.
Why did you pick bristle blossoms? why not frost burgers? or, maybe, meal lice? You would get completely different numbers for the food cost. Generally, I did not include food cost in my calculation, because food is usually abundant and there are plenty of setups where you can start making food with practically no cost and there are setups where food is very expensive (frost burgers)


 To address two of your comments directly:  What I'm doing is comparing 1 dupe using the wheel chain you claim was superior... to 1 dupe using the ethanol chain.  That 1 dupe?  That dupe is Z.  The Wheel is X and Ethanol is Y.  And why did I use bristle blossoms?  Because they're decent enough mid-grade food that most people will use at some point.  They're not exactly appropriate for the food tier you'd have running ethanol, since by that point you'd be running some mix of sleet-wheat, etc... but bristle blossoms in specific?  They have no other cost than water and no fertilizer.  I didn't have to model "Oh, we have to cook this" or any other additions steps.  If I'd done meal lice or ... basically any other choice, it would necessarily complicate the calculation needlessly... and I'd already spent more time than was warranted on that spreadsheet.
 
The point you seem to be missing is precisely that (x+z) does not equal (y+z), and that the comparison is kept apples to apples.  If you are comparing (x+a) and (y+b), you aren't comparing the difference in two processes, you're comparing the difference of four processes.  That's why when you benchmark hardware, you use as close to the same setup for both systems as you can get.  You don't, say, benchmark a new processor released yesterday with a 10 year old video card against last year's model with a brand spanking new video card... and then turn around and claim "Look, last year's model is better!"  There's a bit more that goes into that in terms of how exactly you're running the benchmark, but that should be obvious enough.

 In short, if (X+Z) is better than (Y+Z), that's indicative that if you change the value of Z, the comparison between the two is not going to change significantly, if at all, provided the setup mathematically boils correctly to (X||Y)+Z, because X and Y are constants.   It's a vital part to mathematical logic that if x+z=y+z, then x=y.  That's one of the first proofs anyone is taught in a 200-level math logic course, generally right after or before you're taught how to prove something is even, or odd, or what numerical set they belong to. 

I should point out that I'm not doing (x+z)/(y+z), as you asserted.  I'm doing (x+z)>(y+z)||(x+z)<(y+z) ...  that's a boolean OR.  Process x plus common factors is either greater or lesser than process y plus common factors.  Only one side can be true.  Given your objections so far involve dickering over the value of Z, you're not making a convincing argument because every W you subtract from Y is  W subtracted from X, and every kg H2O you shuffle moves similarly.  That's the entire point of using common factors.

 Let's play a game.  I've presented X+Z and Y+Z... now, let Z=a.  You insist that Z=a+b, where b is literally any real valued number, positive or negative. then X+Z -> X+(a+b) and Y+Z -> Y+(a+b) ... If X+Z = Y+Z, then X+a+b=Y+a+b.  In fact, it doesn't matter if they're equal, or if it's an inequality.  Changing Z does not change the math.  But... what if Z=a/b?  X+Z -> (aX)/b and Y+Z-> (aY)/b.   You'll note nothing changes.  Odd how that works. 

 Is it obvious why food choice doesn't matter yet?

 

11 hours ago, DarkMoge said:

since when? I made sure to make a test world and try to build a setup with minimal travel times to test it... My first idea was, maybe, just maybe, hatch ranch take less work than other ranches.

ranches.thumb.PNG.e5520d7196c5a031e7a44fcf372d6bf5.PNG

500 seconds of work time... According to your numbers, Camille should be able to handle 500/8=62.5 hatches which almost 8 full ranches... But during the live test, she was able to handle only 19 hatches in those 500 seconds... Well, we can also add the ranching bonus of 110%. 19*2.1=39.9 hatches. Though, in more realistic base, we would probably have a rancher that handles 32 hatches or maybe even less, depending on how much downtime we give them, in your calculations, you give them 3 hours downtime and 3 hours sleep time for 450 sec worktime as an example. Anyway, handling 1 hatch is about 12.5 sec....

But honestly, you could make more efficient setup and just get yourself 20 ungroomed hatches within an enclosed room and they would give as much coal as 4 groomed hatches.


With so many wrong number, your all other numbers end up being wrong too.


 Probably because I sat and watched my sage hatch ranch for 5 cycles and averaged the times observed.  I was pretty sure I'd included a disclaimer on that line, but I scrolled up and saw that it's not there.  See, I'd written that post basically 3 times and due to either my own idiocy in accidentally clicking something or other factors, page shifts ate the post.  The first two times I'd written it, before I said "Screw doing all this math in the post, I'd included a disclaimer saying "This is based off observational data off of the same rancher doing the same tasks, not pulled directly from game files.  Individual groom times may be shorter or longer and skill level may or may not affect time."   But, point being, groom tasks were directly observed to be 5 seconds and depending on location of the hatch being called and how much of a jerk they are (hence the hatch jerk buffer time), grooming one hatch has an expectation value one can arrive at through statistical analysis.  Given it's just 4 hatches, the room isn't very wide so there's not a lot of time wasted like there would be for a full 4x24 ranch.

 Not to get too far afield, but there's a reason why I'm giving the dupes the same time schedule, and that schedule is 2 break, 1 bath, 3 sleep... because that's the default schedule... and depending on the person reading this post, that's the schedule most likely to closest to the one used.  Now, you might want to be "But no, I can totally squeak out 500 seconds!  but... that doesn't matter.  Every extra second you squeak out for the wheel is an extra second the ethanol dupe gets to spend on his wheel.  Net gain for your case: zero.  Why?   Given the paragraphs above dealing with X+Z vs Y+Z, it should be obvious, but just in case it's not... be patient.  Read the next paragraph and the points that follow.

 But let's address how realistic that time is, right?  8 seconds per hatch, (5+3)4 ... let's double the ranch size to a full 8, which does imply that the jerk hatch delay should increase somewhat accordingly since we're doubling the area they can be in, so (5+6)8 -> 88 seconds.  It might be non-linear, but I really don't feel like doing non-linear extrapolations at the moment when doing so is pointless.  You'll see why it's pointless here in a minute.  Add in task trip penalty, 98 seconds.  Let's add a second ranch, 196 seconds.  On the face of it, that appears pretty low... but here's a few points to ponder:

1. 1 rancher can handle 2 full ranches worth of grooming tasks + fill in elsewhere (or handle a partial 3rd ranch),... given I'm not modeling interrupting tasks, or other tasks at all, this isn't quite as unreasonable as it first sounds... but before we dive down that rabbit hole, keep reading.

2. Each second you quibble over is only subtracting (2/3) of a watt off of the Ethanol cycle.  Can you guess what where this is going?

3. If we assume 2 ranches eat a whole cycle, one ranch eats half a cycle, so half a ranch is 1/4 of a cycle.  Flipping? It's 5 seconds per flip. 3-4 compost bins? 15-20 seconds, so 1/30 to 2/45th of a cycle., even adding trip penalty doesn't significantly alter this.  Those trees? 3 seconds per branch, 5 branches, tree, 8 trees, as observed, 120 seconds.  That's slightly more than a quarter of a cycle... So using the largest values to give you the complete benefit, 0.25+0.26+0.04= ~0.56 
That's slightly more than half a cycle, so at that point, it's quite clear the dupe will have extra time after doing all the necessary tasks to sustain the ethanol cycle.  But in the interest of fairness, let's pretend he doesn't.   Let's pretend that somehow he has zero time left over, that, say, doing the trees takes 2.5x the observed time.  Sound fair?

4.  Ethanol cycle loses 162W, which then changes the output to: 1163.8-162= 1001.8 W  .... How's that reflect on efficiency? 1001.8W / 235.77 kg H2O = 4.249 W/ kg H2O.  That's still more than twice the wheel.  Your assertion still fails.  

5. If you're going to change the inequality, you're going to have to find a math error that cuts the ethanol cycle's efficiency by a factor of 2.61 just to make them equal.  While I'm happy to own up to the possibly being math errors, I don't see that totaling to 615.36 kg H2O total consumption or that my net wattage is off by 717.9W. (Which would be an error of 61.8% on wattage and 38.3% for H2O... that'd be starkly obvious to the point you wouldn't be the only person contesting the ethanol cycle numbers.)  Or... and this is a big "or", you'd have to more than double the amount of time allotted to the wheel on your cycle.  Do you have 1148.4 d-s to devote to the wheel in a single cycle?  No?  Can you subtract 83.58 kg of water from the wheel cycle without similarly subtracting it from the ethanol cycle?  Actually, no you can't because of how the test is constructed.  A change in X automatically entails a change in Y.

11 hours ago, DarkMoge said:

1 dupe:

input: 67.5 kg of clean water. 1 kcal.
output: 6.7 kg of polluted water, negligible amount of CO2. At least 450s of worktime, can be upped to 500s of worktime or even 525s.

I have calculated for 475s worktime, your 1 dupe nets 394W, not 257.8W... It could be higher.
Even if we ignore the 6.7 kg of water return.

394W/67.5 kg of H2O -> 5.83 W per kg H2O per cycle.

 Tisk tisk, hiding your math is unbecoming.  Using the bathroom eats 5kg water and generates 11.7 kg PH2O that sieves at -0.28W. You're still using 112.613 g/s H2O for the electrolyzer, or 67.567kg H2O.  Given 67.567+5=72.567, not 67.5, your math is already in error and understating the H2O requirement by ~7% while ignoring a quarter watt deficit, but lets ignore that and let's keep going...

 That wheel only produces 2/3 J per second operated.  (400W source / 600 seconds per cycle = 2/3 J/s) As per my second paragraph all the way at the top, calling it 2/3W or 2/3J doesn't change the numbers, it just annoys OCD-about-units types... so calculating this out: as (2/3)*x where x is total available work time... you get:
450->300 W
500->333 W
525->350 W

Factoring those trip penalties only lowers it, so you're overstating your wattage output by a significant margin. (26%- 30%)

But let's pretend you're not.  Since you gave hard numbers in at least one spot, Let's pretend that somehow 475 d-s really does translate to 394W... this gives us two paths, we can either assume you're increasing the efficiency of the wheel, or we can assume you're altering the yield off the O2 cycle.  I'll be charitable and assume you're not fiddling with the wheel output because that'd be too obvious.

475 d-s -> 316.67W
394 W-316.67 W = 77.33 W
 So you're insisting that you can get 77.33W off O2? Okay, let's go with that.

For reference, my original numbers

Quote

-34.23 W (O2)-0.8W(food)-0.468W (Sieve) = -35.03W

and

Total O2 Cost: -13.51+100.9-121.62=-34.23 W

So H+P+E (H2 gen + pump/filter + electrolyzer) = +77.33?  Okay.  So, we both agree that the electrolyzer is 13.51W...And (12.6g H2 / 100 g H2)*800W = 800*0.126 = 100.6
100.6+P-13.51=+77.33
100.6+P=90.84
P=-9.96 W
So. some linear combination of pumps and filters yields 9.96W
a*b*p+c*d*f=9.96
p=240, f=120, a and c are integers, b and d are necessarily less than 1.  a<=1.

Sanity checking this setup immediately fails.  Why?  At the barest minimim, you're running 1 pump the full cycle, so 240W.  Now, we're only counting the dupe's usage, so 100g/s, which if you pretend your pump gathers nothing but oxygen and magically fills the pipe with 1kg/s packets, that's 10% of 240W, or 24W.  24>9.96.  That might seem uncharitable, and but it really isn't.  1 pump only pulls 500g/s, we both know this.  Electros that back up will prefer O2 output over H2 output.  So, if you're going to use the 12.6g H2 rate, the electro cannot be allowed to back up, therefore the pump must, by assertion, pump 1kg packets, otherwise everything else becomes trash.  Given that 1kg packets is similarly trash, it's pretty obvious this path is non-physical and simply cannot be.

 Now, that might seem a little unfair, and if I stopped there, it would be.  It then makes sense, if we pretend that gas flow over the electro can be adequately serviced by 2 pumps and is not turbulent, allowing you just 2 pumps while still getting the full 112 g/s H2.  You can't, but let's pretend the turbulent flow doesn't exist so then you can.

 Now you have 2 pumps, both running the full cycle: 480W.  The expectation value for each packet is then 444 O2 and 57g h2.  Now, because of gas quantization, every time a pump cycles, it will gather only one gas.  Which gas is dependent on whatever's inside its reach, starting with its preferred square and comparing that to whatever is in the buffer.  Modeling that?  That's more math than I care to do, much less care to type out here... so we take the easy way and approximate things by taking the cycle average.  We know the total amount introduced and the total amount being removed.  By previous assertion, we're pretending no H2 shortfall and we're also assuming the system is running at equilibrium, therefore the amount removed equals the amount added.  Since we have 2 pumps pulling the full amount, we divide the input by two, and because of gas quantization, we recognize that we must pump of these fractional packets, not 1.  Sooooo... Our expectation value for the average packet is 444g O2 and 56g H2.  Given this is statistics, over sufficiently short time periods, you will experience variations.  This isn't any different than Quantum mechanics in that regard.  If you expect X over a time period t, then that means all the values during that time period average to x.  If you unwind time, ignoring the most recent packets first, shortening t, the average value will fluctuate, perhaps wildly, depending on the distribution/standard deviation of your sample... but I'm getting sidetracked.

We're only using 100 g/s out of the packet of Oxygen, so that's  0.225225 of each O2 packet, let that equal 'a' and pulling the 12.6g H2 out of the packet of H2 means we're using 0.221053 of each H2 packet, which we'll let equal b.. Since both pumps have the same wattage and necessarily must run the entire cycle, (a+b)240=total power required to pump just what we're looking for. That's 107W.  That's even further afield from our first guess at decyphering your math.

 So let's set that aside and try to find an easier bar for your math to satisfy.  How we figure out how you can separate the H2 from O2 (minimum 1 filter)?  That filter processes every packet, 120W, regardless of size since you're feeding the gen with it.  That means you eat the full 120W penalty the entire time it's active for any packet that applies.  Now lets be extra generous and scale that by our utilization numbers above.  Every packet is either oxygen (thus incurring a 0.225225 utilization) or hydrogen, thus incurring a 0.221053 utilization.  Reusing our math from the last paragraph, (a+b)*120=total power req.  Since 120 is half 240 and most people can divide by two in their head, we skip the math and go straight to 53.5W.  That's still higher than your number of 9.96 W by a factor of  5.37.

 Now, maybe I misunderstood what you were getting at so let's take a moment to check assumptions.  Let's start over, starting at your end point and working back.

 You assert a net gain of 394W.  Okay. 

D+W=394 where D is all costs associated with the dupe, W is wattage gained from work. 

W=T*G, where T is time spent operating the device and G is the fractional wattage generated on a per second basis. 

G is derived by taking source wattage / seconds per cycle, or 400 / 600 -> 2/3
I'll waive the travel penalty here, allot you the entire 475 d-s you're insisting you have.
(2/3)*475=316.67

 Okay, so D+316=394 -> D=77.33 ... which is what I was using.

D= O2+food+conversion costs
 I didn't see anywhere in your math anything dealing with food or conversion, so I took those as 0...  so O2=77.33

G+P+E=77.33
12.6g H2 runs 800W * 12.6% = 100.8
We both agree the Electrolizer runs 13.51W just for the dupe... so
100.8+P+(-13.51)=77.33
100.8+P=77.33+13.51
P=77.33+13.51-100.8=-9.9567.

Nope. At this point, I've been more than charitable   Your math doesn't add up.  You simply cannot pump *and* filter 100g/s O2 and 12.6 g/s for 9.9W total power.  Again, 2 pumps minimum, and even if we simplified a different way to try to trivialize the math, you get 888g/s O2 on one pump, 112 g/s H2 on the other.  Which then means the O2 utilization coefficient = 0.112613 and the H2 pump's coefficient is 0.1125.  (a+b)*240 = 54W.   Whatever model you used does is simply invalid as a representation of what people would experience in game.

But... in the interest of going so, so very far out of my way to give you every last inch of the benefit of the doubt... let's just pretend your math was actually correct.

Your figures for the common dupe cycle
D=77.33 and not -34.23W.  That's a change of 111.56W for the common dupe.
You're claiming 67.5kg usage for the common dupe, I went with 135.774 , that's a change of -68.274 kg for the common dupe.

I'm really hoping you've been paying attention because the next step is as obvious as it is inevitable because the common dupe is... common.  He exists the same in both cycles.  Remember X+Z and Y+Z?  All you did is alter Z.  I'm not even going to go hunt through the math I did to correct the ethanol subtotals.  No, I will eat the entire conversion cost for resources that do not exist, and also the pumping costs for non-existent water, and literally every other cost I wouldn't have to pay because I'm changing the expected values for the 1 dupe both cycles share.  In fact, I'm not even going to argue that the extra 25 seconds of work you introduced could be used by the ethanol dupe for anything other than standing at the water cooler talking to himself for a +1 morale buff.  And so the ethanol cycle will be presented at it's utterly weakest, using entirely your numbers.

Original figures for ethanol cycle efficiency:  ( Ethanol Cycle Efficiency - 1163.8 W / 235.77 kg H2O -> 4.94 W / kg H2O
1163.8 W + W_dupechange = 1163.8+111.56 = 1275.36 W
235.77 kg H2O + H20_dupechange = 167.496 kg H2O.

Ethanol cycle efficiency using your altered common dupe #s and eating full penalties without revision becomes:
1275.36/167.496=7.61427 W/kg H2O.   And that's without tuning anything, the same as your untuned wheel.  Given the generators I'm toying with are all much larger base capacity other than that pesky wood burner, the wattage output of ethanol favors tuning... but accounting for dupe time to do that makes things more complicated than it already is, so I've skipped doing so.  Proving minimum baseline is sufficient.

What was your revised efficiency?   5.83 W/Kg H2O. 

 So, do you see the beauty of (X+Z)(Y+Z) comparisons yet? :)  When I said it didn't matter, it's because it didn't matter, not because I was making some rhetorical or mathematical sleight of hand.  What you change common to one must be similarly changed in the other.  The ratios will always be preserved unless you can find a change that affects one and not the other.  Your best bet was *not* to increase labor time.  No, the best argument you could have made would have been to go through the input/output streams for the ethanol cycle and find a typo or miscalculation that either gave me a few hundred watts I should've have or artificially lowered my water usage by a significant amount (Those specific numbers I mentioned up stream, BTW)  ...  And I say that was your best bet because the wheel cycle is so dead simple that the only pace I *could* have screwed up was the common-dupe portion.  Your next best bet if you insisted on toying with labor would have been to look at what the ethanol dupe was doing, which... well, you tried to, but even when I granted you your presumption as true, even when I went further than you'd pointed out and demanded his entire work cycle be nothing but those three tasks, ethanol still won.

Your alternative still spends water and its return on water spent is less efficient.  If you have the water capacity to spare and you need energy, the ethanol cycle will give you more wattage per unit water overall than your alternative at the very least.  Maybe more than anything that comes before it if you factor in the dirt/coal renewal, whose value depends solely on the what the player needs.  The petrol path is certainly similar enough, but it's not going to get you all that dirt... and your dupe is going to spend a little less than half a cycle consuming 480w at the refinery, so depending on how the math shakes out, it might not be nearly as good as it sounds, especially if you need more dirt than the sieving just the petrol output would.  Petrol output that's identical to the ethanol output, I might add.

 Which ultimately suggests:  if you have extra water, ethanol is a good idea.  If you have extra water and you need renewable dirt, ethanol is your best friend because it will provide more dirt for similar wattage efficiency than straight petrol will... and it's flexible enough that you can scale the dirt production at the expense of wattage.  And there exists no other alternative that can do this for you with the same efficiencies.  Solar doesn't print dirt.  Steam doesn't print dirt directly.  NG kinda does print dirt, but it's pretty slow at it and in order to scale it up, you have to have enough NG and past a certain point, you're wasting W just to get dirt... if you're going to sacrifice W for dirt, ethanol is better at it.  Coal runs into the same problem, except it's even slower at making dirt than NG is and its resource is starkly finite without space travel.

 And that's why the ethanol cycle doesn't need buffed. It already has a purpose and it serves that purpose well.

Postscript:  Posting while tired is generally a bad idea for much the reasons you suggested... but you failed to find any substantive errors in what I posted and having looked over it again, unless I missed the errors a second time, I didn't see any.   As someone who did systems analysis for ~15 years before going back to college and then getting derailed off an Econ major into a math/physics dual major, It was nice to deep-dive into modeling things in this game.  As a result, I've filled in a large gap in what was already a fairly involved spreadsheet I'd been using to tweak/calculate design concepts on top of developing a few techniques to selectively tweak the input/output of said spreadsheet.  Thankfully, I let it do all the math, so the biggest danger was transcription errors, not math ones.

 Also, despite some of the tone I may have used that might suggest otherwise, I truly and honestly appreciate the opportunity to recheck my math and validate existing models and how they differ from observed conditions.  Sometimes I get wrapped up in the math and forget not to be a b-face to other people, so for what it's worth, my apologies for any offense I may have offered.

Link to comment
Share on other sites

2 hours ago, storm6436 said:

The point you seem to be missing is precisely that (x+z) does not equal (y+z)

You are missing what I am trying to point out again... What you are doing in your calculation is (X+Z) divided by (Y+Z) and saying that its equal to X divided by Y. Most of your "does not matter" have this exact math error that you apply multiple times.

lets go with an example...

X=6
               ---------->  X / Y = 6/9 = 2/3 = 0.66666
Y=9
------
Than, we add Z = 3

(X + Z) / (Y+ Z) = (6 + 3) / (9 + 3) = 9/12 = 3/4 = 0.75

And I hope we can both agree that 0.6666 is not equal to 0.75... But in your calculations you are pushing that it is.
In my attempt to point out flaws in your mouth, I bring up bristle blossom, because it is Z. You are calculating ration between X and Y and you are adding Z to both sides.
 

2 hours ago, storm6436 said:

Every extra second you squeak out for the wheel is an extra second the ethanol dupe gets to spend on his wheel.  Net gain for your case: zero. 

And you are applying same error here. In your case, you argue that your ethanol cycle is 1275.36W... Lets go with that number to point out the error of that statement... I argue that my dupe is worth 394W.
X=394W
Y=1275W
X/Y = 0.309

Now, you want to argue that adding extra second to both sides does not mean anything, but it is our Z... 25 seconds running on a wheel is about 16W if we round it.

(394 + 16)/(1275 + 16) = 0.317

Huh, our ration completely changed. Now to continue why it matters... Lets say we added another duplicant to both sides running on a wheel 475s, with my calculations its + 394W to both sides.

(394 + 394)/(1275 + 394) = 0.472

How does this not matter?
In your previous calculations, you were adding water to support bristle blossoms to both sides and than you are calculating their water consumption ratios.
I do not know how else to explain the error in applying math in this fashion. You are not doing comparison of X < Y, in that case adding Z to any side would make no difference. In your calculations, you are comparing ratios of power to water. More accurate formula for what you are doing would be:

X1 / (Y1 + Z) ? X2 / (Y2 + Z) 
and claiming that it is equivalent of X1 / Y1 ? X2 / Y2
Lets take some more random numbers to make an example

X1=1
Y1=2
X2=3
Y2=9
Z=3

than 1/2 > 3/9  ~~ 1/2 > 1/3
but 1/(2+3) < 3/(9+3)     ~~ 1/5 < 1/4

Thats exactly what you are doing with adding bristle blossoms and it will also happen when you change duty cycle of a dupe.

2 hours ago, storm6436 said:

Given 67.567+5=72.567, not 67.5, your math is already in error and understating the H2O requirement

Well, you do realize that if I calculate the extra water return, my ratio actually gets better not worse? I could also use outhouses, because its much more simple. But anyway.
394/67.5 = 5.83
Lets add the water return from water sieve. We only need a sieve, because you can feed the output from sieve as input for your toilets. We produce 11.7 kg of polluted water... At full capacity, sieve processes 3t of water... It needs to work for slightly more than 2 seconds to process it.
11.7/3000=0.0039
The sieve at full capacity costs 120W to run.
120*0.0039=0.468
Which is less than even 1W... Lets generously round up to 1W and than the ratio becomes
393/60.8 = 6.46

Actually, I should have went with that number, because of how much better it is for my side of the argument.

 

2 hours ago, storm6436 said:

And why did I use bristle blossoms?  Because they're decent enough mid-grade food that most people will use at some point. 

Well, I want to argue that bristle blossoms are the worst food, but I have been forced by the game to use bristle blossoms multiple times on ravenous hunger difficulty... I would than *Upgrade* to meal lice and get rid from my bristle blossoms... And than upgrade to omelette... and than to endgame foods.

 

 

2 hours ago, storm6436 said:

and that schedule is 2 break, 1 bath, 3 sleep...

Well, you see, I applief same logic to get 475s, because 1 segment of bath will be treated as extra segment of work unless the player purposely adds showers.
 

2 hours ago, storm6436 said:

Sanity checking this setup immediately fails.  Why?  At the barest minimim, you're running 1 pump the full cycle, so 240W. 

That sanity check fails, because you do not run a pump for anything other than hydrogen, you let hydrogen self filter, there are many ways to build that and let your O2 go straight into your base. Here is the setup that I use in my bases. I could not think of any better way of explaining other than giving a video.

https://www.dropbox.com/s/4jcfe7mlgbplgwc/2019-08-31 05-25-25.flv?dl=0

Hope it helps. You can see the amount of hydrogen rapidly growing within the reservoirs. And you can see pumps only ever pumping hydrogen.

Lets do this argument differently... I want to see you make a box with ethanol production chain that uses domestic trees and all the things that you have described and I want to see it run for multiple cycles without increasing in temperature. Using infinite gas stacking to stack CO2 is not allowed, because it gives me ocd and no venting of the CO2 to space.... Just lock 1 dupe inside the box with ethanol production chain and let it live there... You can give it no stomach setting to not overload you with setting up a food farm.

 Maybe, just maybe, I cant imagine the working setup that does not overheat, because, so far, cooling was not a part of your calculations. Also, auto sweepers and auto loaders were not a part of it either.

2 hours ago, storm6436 said:

  As someone who did systems analysis for ~15 years before going back to college and then getting derailed off an Econ major into a math/physics dual major, It was nice to deep-dive into modeling things in this game.

Do we need to talk about our education? On my side, I completed masters degree in specialized computer systems. I have worked in both programming and making them. I am working to obtain phd in electronics

Link to comment
Share on other sites

I'm not doing it in (X+Z)/(Y+Z).  X and Y are the yield and cost ratios from process-unique factors and as I've said at least twice now, I'm doing a comparison. Sure (X+Z)||(Y+Z) isn't necessarily the best way to express the comparison, given "||" is a boolean OR, but doing it in the correct notation is even uglier.  Still, you can keep insisting I'm doing X/Y and I assure you, you'll remain wrong until the universe eventually fizzles out, collapses or whatever it's going to do, if it's going to do anything at all.

Look at it this way:  X = (Process_yield/Process_cost), Y is the same but for the second process.  Z is the common factors that both suffer from.  Z = the common dupe stats I calculated beforehand.  At no point am I dividing X by Y or Y by X.   It's a strict comparison unless I choose to normalize... ie. set X=X/X=1, *then* y=y/x ....  but all that does is give you the favorability ratio for those two specific numbers and the resulting number literally cannot be used for any other purpose precisely for the reasons you object to x/y...., which is to say, if y/x =1.2, then you know that process y is 20% more efficient than the other. Flip that.  1/1.2 = 0.83,  What does that tell you?  That process x is 83% of process y.  Nothing more.  Normalizing to compare differences is simple benchmarking and it is done by literally every scientific and engineering field on a daily basis.  That is literally how this is done by people who probably pay more in taxes than either of us earn gross at the moment.  In fact, if you couldn't normalize things and obtain useful information, quantum mechanics and all subsequent fields of study simply wouldn't work.  Yet they do.

 But... I'll break it out a little bit more, again X,Y = their respective unique factor ratios.
The (X+Z)||(Y+Z),compare resolves merely into X=Y, X<Y. or X>Y.  , they're either equal or the inequality points in one direction.  This isn't QM or QFT, so no superpositions allowed. :p

 Now, let's reframe this.  Say we decide to go with an alternate food source.  What does new food do?  It doesn't change X or Y.  It can't because food isn't part of the process, it's part of dupe upkeep... and dupe upkeep belongs to the common factor, therefore that changes Z.  As I said in my very first post: your accounting is wrong.  Whether it changes the wattage used from Z or the water consumption used by Z doesn't matter for this example.  For the sake of completeness, let A indicate the positive difference, if any, between the original common rate and your new calculated rate.  Similarly let B indicate a decrease the same way.  To spell it out, if Z was 10W/2kg and you come up with some new common thing that's 20W/1kg, then A=15W/kg.  Similarly, if the new calculated rate is 5W/2KG, then B=-5W/2kg.  Does this make sense?

The two values then change accordingly:
(X+Z) -> (X+Z+A)
(Y+Z) -> (Y+Z+A)

Making the comparison:
(X+Z+A)||(Y+Z+A)

Using B proceeds exactly the same except you end up with B instead of A on both sides.  Can you explain how you change the result by altering Z into A or B?  If you think you can, keep reading:

Let's assert X>Y to start.  Doesn't matter which process X is, it's just better.  Hell, it could be the bellybutton lint generator option vs the cat with butter on its back cycle for all it matters. 
X>Y, by assertion means if you add same factor to both sides, the inequality remains the same.  (X+Z)>(Y+Z)
(X+Z+A)>(X+Z), by construction.  (ie. A is a positive number)
(Y+Z+A)>(Y+Z) by construction.

Therefore.... let's add these terms one at a time.  Starting with X+Z+A to the original assertion that (X+Z)>(Y+Z)
(X+Z+A)>(X+Z)>(Y+Z)

And then the other:
(X+Z+A)>(X+Z)>(Y+Z+A)>(Y+Z)

You'll note that there's literally no way adding A changes the fact that X+Z>Y+Z.  Since B works similarly but in reverse, X+Z+B<X+Z,.  Let's explore how that reflects though.
X+Z+B<X+Z, therefore Y+Z+B<Y+Z
Therefore:
(X+Z)>(X+Z+B)>(Y+Z)>(Y+Z+B)

  *OR, if B is sufficiently negative:
(X+Z)>(Y+Z)>(X+Z+B)>(Y+Z+B)

(X+Z+B)>:(Y+Z+B) remains true regardless. . It just swaps their relationship to their original state (ie. whether X+Z+B is greater or lesser than Y+Z and vice versa.  But the comparison at that point is between X+Z+B and Y+Z+B.. You literally cannot effect the relationship of the processes by fiddling with common factors.  Again, this is a basic math proof taught at the 200 level.  Your food suggestion? Common factor.  Better food just makes Z bigger or smaller, so A or B in the previous example... which I've just proven mathematically that such things literally will not ever change the relative value. Adding more labor time?  Common factor.  Complaining about oxygen efficiency?  Common factor.  Almost everything you have suggested to include the post I'm replying to here is a common factor, so therefore A or B, and therefore mathematically there will be no meaningful change.  Will the ratio's change? Sure.  You might start with X>Y, where X/Y=1.2 ... but say you didn't have a Z at all but introduced one, that's the same thing as changing Z with Z+A or Z+B.  so X+Z/Y+Z, where, say, Z is 2, then we go from 1.2 to 1.06.  Closer right?  But 6% higher efficiency is still noticeable and therefore worth doing.  Let's say Z=1,000,000,000,000, a trillion.  Now, most calculators will see those numbers and tell you the answer is 1.  They're wrong, but they'll tell you that because they don't have the digits. A billion and one divided by a billion is certainly not 1.  Unless Z=infinity, they will never equal 1... and technically you get infinity over infinity which requires you to run L'Hopital's rule to see if it's undefined or actually resolves to an answer.  I'll spare you the calculus..  At best you'll get to the point where you'll just shrug and say they might as well be equivalent.

Still, you did have one suggestion that was relevant but poorly considered: fiddling with what the ethanol dupe can do.  If he can't tend all the hatches, then we lose up to almost 600W... but even with losing 600W of generator time, I don't quite lose 600W.  Why?  Because the systems that service that generator then also shut down and I *save* their cost, leaving me -600W (gen) + (support costs) ...  but that's exactly why I turned around, outlined how that dupe *certainly* had enough time to do those tasks and then basically had him stand around with his thumb in his ear the entire time he wasn't doing those three tasks.  That scenario is the absolute worst case the ethanol cycle runs into if manpower demand is met. 

Minor rehash: 8 trees 4 hatches, 4 compost bins. 600 seconds to it all in, but only 450 seconds where the dupe is active.  We know 1 dupe can do 2 full ranches in a day plus a little more... round that to only two to give your argument maximum advantage and that still means those 4 hatches take a quarter cycle, at most, and that's accepted by basically everybody, not just me. At t=10, he starts hatches.  T=160, go to composts, T=170, start flipping, T=190, move to trees, T=200, start trees.  Are you going to tell me that 8 trees takes 250 seconds? It doesn't but I pretended it did anyway.

 What would we need for 8 trees to take 250 seconds? Well, we know the dupe starts immediately in range of the first 3 branches... so let B= the time per branch. Say we have 3 branches on the left side of the tree, 2 on the right, that's a travel of 2 squares to hit the other side... and 1 square to move onto the next tree.so, 1 tree proceeds as left 3branches->2sq->right 2 branches->1sq.to next tree.so the first 7 trees are worth 5 branches each and 3 squares, the last tree is only worth 2 since there's no tree to go to next. So 35 branches and 21 spaces + 5 branches and 2 spaces.... we get 40 branches and 22 spaces.  How long does it take a dupe to move one space?  Depends on athletics, right?  Well, let's just presume something absurdly slow, like 1 second per tile.  Even anemic day 1 dupes can manage faster than that, I'm pretty sure.  Not by much, maybe, but 1.0000001>1.

 So, 40 branches, 22 spaces becomes 40b+22s=250.  -> 40b=228, b=5.7.    So in order to eat the entire day, it would require the dupe to travel 1 second per tile and take 5.7 seconds per branch.

 Well, farming skill is supposed to reduce timers by 5% per level.  If the base time is 6 seconds, someone with 1 farming skill can do it in 5.71 seconds.  Does the game round?  I dunno, I doubt it... but you're not sending unranked farmers after trees if you're making ethanol.  Skill level 2 can easily do it.  At base rate 6 seconds, you need to be at skill cap to hit 3 second harvests.  Given that I just checked all my farming dupes skills and only 1 was at 20, this proves that any dupe with 1 rank in farming can complete the task, but we already had that since we needed 1 rank in ranching to do the hatches.

So what this proves is that even with anemic, a dupe whose first two skill choices went to farming and ranching could complete the task. Anyone doing ethanol should already have one of those. It's mid-late game like you said. My guys were doing it in about 3 seconds flat, so that would still leave them 107s, or 97 seconds after travel to run on the wheel.  And i passed that up to make the point that the process constant for Ethanol is clearly larger. I literally did nothing but complete the necessary cycle tasks and still beat the wheel.  Nothing else matters unless you can go through the ethanol cycle's process list and find me inputs or outputs that do not match operational values.  The math on the wheel is too simple to screw up.  Everything else is just Z.

But... then I read the rest of your post.  There's no real complimentary way to tell someone all the math they think they did only serves to confirm their bias because they do not understand the logical structure they're operating in.  Sorry.  Like I started off with, this isn't X/Y.  It's X>Y or X<Y.

 Literally everything you typed out is mathematically irrelevant and based on how you proceeded in your post, you are clearly having issues understanding how the accounting works.  Go ahead, change toilet vs outhouse.  Doesn't matter, that's Z.  1 pump air, 20 pump air, even using debug to create the air and the hydrogen directly at zero cost, doesn't matter, that's Z.  In fact, the worse you make Z, the closer the comparison becomes only X vs Y and as outlined above, if you set Z=0, which is where progressively worse common factors head as they get worse and worse, it doesn't change the relative value.  If X>Y, X will always be greater than Y.  Always, no matter what value of Z.

As for the thermal dodge, nice try.  I didn't include thermal because the discussion was already going to be hard enough for most people to follow.  Not including those numbers is irrelevant if all we're talking about is W/kg H2O.  You'll note this isn't W*DTU/kg H2O or W/DTU*kg H2O.

If we locked both cycles in an insulated square and magically ensured they never ran out of air that did not affect the total energy contained in the box, they'd both die from heat exposure, the only difference being how long it would take.  Why?  Because we're talking only unique things to each process.  Neither process includes a cooling solution, and even if you wanted to go that direction, it's irrelevant.  Why?  Because then it isn't X>Y or X<Y, it's X+A>Y+B or X+A<Y+B., which I pointed out right at the start is not measuring the effectiveness of one process against the other at all.  That said, ethanol does have heat deletion capabilities and I did briefly discuss one of them that in one of my early posts.  There's more than one though. I can delete heat feeding water to the trees and delete more feeding lumber to the distillers.  That's built into the process... so you can, in theory, use ethanol to delete heat for other processes.  Can't do that with a wheel.

 Though, also, nice try with the sweeper arm bit, too.  I did start off explicitly saying (or at least I believe I did, remember the post deletion issues) that the ethanol cycle was automated.  How many watts does it take to move all that Pdirt?  0.25 second per activation?  120*0.25/600 = 0.05W.  Even half a second per activation, it's 0.1W  The ethanol cycle leads by 717.9W ...do you really think it takes 7179 half-second sweeper arm activations to run the ethanol cycle for one day?  That's how many it would take for the wheel to break even.

Heck, let's pretend the ethanol plant drops 1 chunk of dirt every second.  So 4 activations a second. That's  600 activations. Thing is, I can automate sweepers to trigger only every so many seconds... Sweeper arms can move 1000kg, so at 333 g/s, I can move  enough to feed the entire ranch once every 450 seconds if I wanted to burst-process the Pdirt... but here's the thing, that generator only needs 1kg/s.  I'm generating 1.33kg/s Pdirt from just the distillers..  I can afford to dribble it in, and let's pretend that impairs the coal generator so I don't get the full 600 W.  Who cares?  If these facilities are located near each other, I might lose, what, maybe 20-30 seconds of output?  95% of 600 is 570.  I lose 30W.  Even if the feeding side uses another 600 activations, ethanol still has another 5979 before it loses.  That 30W? That's 300 activations. 5679 more to go!.  Each branch? 40 activations.  Delivering the branches?  Less than that due to incoming flow vs outgoing placement.  Still, let's pretend it's 40. 5599 to go.  Can you find 5599 more activations?  You're not going to get that feeding the composts.  That's not time sensitive, I can drop that all at once, one activation per pile. 5595 to go. Removing pdirt from the sieves?  I only need one sieve.I'm already moving enough to run the coal gen. 1 activation. 5594 left.  The wood gen? It's 200kg. As long as I drop that before the last third(ish) of the day, that's all it was going to run anyway. 1 activation.  5593.  Delivering coal? I can stutter that out, and depending on battery levels, it might not even run anyway.  The whole point to the coal was that you *could* burn it if you needed. I could go with 0 activations, but I'll be nice and spot you 10. 5583 to go.  And that's it.  That's every piece of automation. I still have 558.3W left, or possibly more depending on the length of the electrical pulse.

Look, again, unless you can find an error (or errors) that swing some combination equivalent to pulling 717.9W or adding 615.36 kg of water consumption from ethanol, there's literally no way you can say the wheel is better, no matter what angle you try to come at it from... because without there being an error in my process math, anything the wheel gets to do, ethanol gets to do as well.  You just end up changing Z. 

External PH2O enters the cycle at only one point unique to the ethanol cycle.  Just one.  External H2O?  Is only a concern for servicing the dupe. It's not necessary to the operation of the ethanol cycle beyond keeping that dupe alive. Put another way, if the dupe in both cycles was an android that didn't need to eat, drink, or excrete... and could work the full 600 seconds without pause or complaint... ethanol still wins and I've proven it mathematically at least 2 or 3 different ways.

You originally said ethanol was underpowered... it's pretty obvious to everyone other than you that we're well past having proven it isn't.  All you needed to do is find those errors.  You presumably have the ability to check the numbers, to find those errors, yet you haven't.  Instead, you keep insisting that shuffling Z will get you the victory you desire... except it won't.  You look to thermal to claim victory, except it kills you, too.  At best, that's a draw. (Side note, I have 700 W to play with and 2 sources of heat deletion.  Want to bet I can run a thermoregulator and 1 pump... and become more heat neutral than your solution?  Or even actually heat neutral or better?  It's not unreasonable to think that with the right pipe config and insulation, it could be done  I mean, I clearly have the power budget to spare here)  And then you went to automation, and well, we see how that's gone.  I didn't include those numbers because they're negligible to the end result.

 It isn't unreasonable to state that if you haven't found the errors then you aren't going to. You've had more than ample time and independent sources to verify my numbers.  More importantly, after a third look through my numbers, I can tell you:  They're not there.   Your wheel loses.  So at this point, after all the legitimate mathematical proofs I've supplied and the number of times I've explained how the comparison actually works, you're not arguing with me, you're arguing with mathematics itself.  And while your wheel loses, math doesn't.  Your argument is invalid.  Good day.

Link to comment
Share on other sites

6 hours ago, storm6436 said:

To spell it out, if Z was 10W/2kg and you come up with some new common thing that's 20W/1kg, then A=15W/kg.  Similarly, if the new calculated rate is 5W/2KG, then B=-5W/2kg.  Does this make sense?

The problem is that if you look at it that way, you are screwing ethanol to always win. why? because, you insist that 1 dupe is a common factor and you are comparing dupe against dupe + ethanol... Remove dupe from both sides and you get 0 against ethanol. And than what ever you do, becomes irrelevant, because the numbers on ethanol chain are always positive, it would always be higher than 0. So, you cant talk about when ever there is a better use for water in that case.

And well, you cant separate the terms like that. Why? because ethanol is actually ethanol + dupe, it does not work without dupe labor input. The thing is, ethanol is a system with inputs: polluted water and dupe labor. And the whole argument is that there is a much better use for those inputs.

And so, I will keep insisting that you are using your math wrong. Because I know that you are using it wrong.

In our case, the model, we have two systems: A and B.

Let me give you a few examples where doing math the way you do proves to be an error. It is like basic math.

Example 1: We have 2 countries, country A and county B. Country A has 25 people in it and has size of 1 km^2. Country B has 60 people in it and has size of 2 km^2. Next year, both countries grow in population by 25 people.
X1=25
X2=60
Y1=1
Y2=2
Z=25

If the question is which country has higher population. The answer will always be country B, because it does not matter when ever we are comparing 25+25 against 60+25 or 25 against 60.
Now, if the question is which country has higher population density. The answer changes. This year, it is:
A=X1/Y1=25 people/km^2
B=X2/Y2=30 people/km^2
B>A
Next year:
A=(X1+25)/Y1=50 people/km^2
B=(X2+25)/Y2=42.5 people/km^2
Between the years, it changed from being B>A to A>B

You are asking 2nd question and applying logic of 1st question to it.
 

Example 2: 
We have two bases in ONI. Base A and base B. Base A has 1 dupe in it and generates positive 200W. Base B has 3 dupes in it and generates positive 900W. After that, both bases decided to tame a volcano and produce extra 850W from the steam turbine for free...

If you ask the question which base generates more power than its always base B, because it does not matter if we compare 200+850 to 900+850 or 200 against 900.

But if the question is how much power each base has per dupe.
Before taming a volcano:
A=200/1=200
B=900/3=300
So B>A
After taming volcano:
A=(200+850)/1=1050
B=(900+850)/3=583.333
So, it changed to being A>B

again, you are basically asking question 2 and applying logic of question 1 to it.

Link to comment
Share on other sites

Example 3:
We have 2 systems in ONI. System A and system B. System A requires 1 dupe to ran it and produces 400W. System B requires 3 dupes to ran it and produces 1000W. At first, we were feeding our dupes automated meal lice production (meal lice autoharvest time is 7 cycles) than we decided to change it to having a pip ranch which requires extra dupe time to ran. Lets say, 1 rancher can support 4 dupes with food.

If we ask a question of which system produces more power per dupe.
While they were fed meal lice:
A=400/1=400
B=1000/3=333.33
So, A>B

now, there are multiple ways to aproach question 2 and get different results, which one is correct depends on the specifics of the situation, lets go with 2 of them:
question 2.a:
We added 1 full rancher to both sides:
A=400/(1+1)=200
B=1000/(3+1)=250
So, it changes to B>A

question 2.b:
System A only requires 1/4 of a rancher to run it.
System B only requires 3/4 of a rancher to run it.
So, values become:
A=400/(1+0.25)=320
B=1000/(3+0.75)=266
it remains A>B
if we dig deeper into the formula of question 2.b, we can discover that applying this logic to it makes it so that a rancher can never change the result A>B. Because, in this application, it does actually become a common term:
A=400/(1*(1+0.25))
B=1000/(3*(1+0.25))
We can remove the (1+0.25) from both sides and it does not change it.

But the variation 2.b is incorrect, because we cant just add a dupe, a rancher and say that he does not eat food as well. It has different impact on both systems, because it is a variable that is dependent on the other variables within the system.

Again, you basically ask question 2. And apply logic of question 1 to it while trying to sell it as 2.b.

If you still cant see the faults of your "it does not matter", "it is a common term" than I give up trying to explain it to you. You can keep on believing that you are right and I wish you a good day.

Link to comment
Share on other sites

7 hours ago, storm6436 said:

In fact, if you couldn't normalize things and obtain useful information, quantum mechanics and all subsequent fields of study simply wouldn't work.  Yet they do.

In fact, I am the one who tries to normalize this information by trying to give water some value. How much value can we get from water by feeding it to dupes and how much can we get by feeding it to ethanol chain. My argument is that feeding water to dupes running on hamster wheels is a better use of water than feeding it to ethanol chain. I picked hamster wheels, because any noob can place a hamster wheel and make their dupe run on it. They can start optimizing that system as they learn the game and they end up with a better use of their water.

Link to comment
Share on other sites

On 8/30/2019 at 4:29 AM, DarkMaster13 said:

I think this is a similar bug to dupes abandoning coal delivery jobs when the generator gets disabled by automation.  I'm fairly sure they didn't use to do this, so it's a new problem.

Yeah this is annoying.

Link to comment
Share on other sites

5 hours ago, DarkMoge said:

You can keep on believing that you are right and I wish you a good day.

he can tell you the same thing :)

and it seems to me you are talking about different things. He systematically evaluates labor efficiency in the ethanol cycle with a screaming wheel. you are trying not to recognize his logic as a legal way of evaluating.

in my opinion, the ethanol cycle is too complicated to implement, but has a lot of advantages. the 6 sage hatchs farm processes p-dirt into coal and supplies meat / eggs (4kcal/cycle), Slicksters love very much hot co2 from distillers and gives oil and meat (ungroomed animals process co2 three times more than groomed ones). the petrol generator gives more than 1000W of free energy, if you heat it and place it under the turbine, you get clean water, additional energy from the steam and  Slicksters love very much hot co2 that are pumped out by the door pump. and no problems with cooling and co2 removal.

did not consider the effectiveness of the structure in figures, but it is two orders of magnitude more interesting than the hamster wheel.

ps. although debug makes any scheme effective

Link to comment
Share on other sites

 

19 hours ago, DarkMoge said:

And so, I will keep insisting that you are using your math wrong. Because I know that you are using it wrong.

In our case, the model, we have two systems: A and B.

Let me give you a few examples where doing math the way you do proves to be an error. It is like basic math.

Example 1: We have 2 countries, country A and county B. Country A has 25 people in it and has size of 1 km^2. Country B has 60 people in it and has size of 2 km^2. Next year, both countries grow in population by 25 people.
X1=25
X2=60
Y1=1
Y2=2
Z=25
If the question is which country has higher population. The answer will always be country B, because it does not matter when ever we are comparing 25+25 against 60+25 or 25 against 60.
Now, if the question is which country has higher population density. The answer changes. This year, it is:
A=X1/Y1=25 people/km^2
B=X2/Y2=30 people/km^2
B>A
Next year:
A=(X1+25)/Y1=50 people/km^2
B=(X2+25)/Y2=42.5 people/km^2
Between the years, it changed from being B>A to A>B

You are asking 2nd question and applying logic of 1st question to it.
 


  That's a pretty bold assessment.  Shame we're off with an immediate invalid assertion based on the same error you've continually made and keep carrying forward: Inability to do proper accounting... only this time you actually provide enough substance to work with.  See, it appears you've misidentified whose math is wrong and based off this post, it's because of one of two most likely possibilities: you don't understand math nearly as well as you think you do or you're deliberately misunderstanding everything I say for any number of possible motivations, most of which involve an aversion to admitting you're wrong..The distant third most likely excuse is both of those at the same time.  Why do I say that? Because it looks like you're literally reaching for any combination of numbers you can stuff into an X/Y format that "proves" me wrong without bothering to stop and consider the construction you're abusing to get there.  Doing so highly suggests you haven't once reconsidered your interpretation despite the literally thousands of words and math proofs I've provided. 

Now, here's where I think you're messing up.  Your example asks a question I didn't in a way that I didn't and your set of Ys is not measuring the same thing mine was.  Your Ys?  Land area. They're constant.  My Y? The ratio for one process.  Your X's?  Population.  My X's? The ratio for the other process.  Similarly, my X and Y are not a function of time. At all.  Your X's are. That's the first clue you've constructed your example wrong if you're trying to impugn my math.  But the biggest and most obvious clue you are mistaken in what you think is your understanding of the math, it's right in front of us, in this paragraph.  What are the units involved?

Your X: People 
Your Y: Area
Your X/Y: people per area


My X: W/kg 
My Y: W/kg
My X/Y: unitless.

 Do you see the clue? Our end units are different.

Why would my X/Y be unitless when yours isn't?  Because my X/Y is a normalization function.  Provided X and Y are positive and X>Y, then X/Y>1, otherwise 0<X/Y<1.  The whole point of normalized values is to provide easy and direct comparison between data.  So if X/Y>1, let's say it's 1.5, This tells you just by looking how bigger X is than Y.  (1.5-1)*100=50 ... which is exactly how much bigger X was than Y in percent.  Do yourself a favor, go hit wikipedia (or just google in general) and look up normalization, specifically in regards to statistical analysis.  It's clear you don't understand why you do this or the limitations of what you can do with those numbers once they're normalized.  Your X/Y isn't normalization, it's a calculation of a new data point derived from one of the two things being considered... the fact that you used X and Y the way you did seems deliberately obtuse, as I observed earlier.

 Still, let's chew on your examples. 
Example 1:
What's the analog to population in my cycle comparison?  That would be the resource amounts.  I have no analogue for area.  None of my resources are static.   That, combined with the criticism above (like how your results are time dependent, mine aren't),. is sufficient to just move on.  Your analogy here has failed.  Because your construction is *NOT* identical to mine, you've proven nothing.

One final criticism for example 1, what you've done here is an X+A vs Y+B comparison.  No, I don't mean your variables are set up that way, I mean that your results do not hinge one one thing being different.  They hinge on two separate qualities between the two different comparison points. (Hence X+A)  These comparisons are utterly useless because they tell you nothing about what your weak point is and require you to dive further into the data to explain the observation.  This is why you control what's actually varying to the point that the only things that change are what's different between the two.  It's almost like... oh, right, X+Z vs Y+Z where X and Y are those "Only things" I just referenced.  Inconvenient, that.  Proper comparisons *must* rely on *only* the differences.  In my comparisons both the water and the wattage were unique to each side and were derived entirely from factors unique to the processes.  At no point did one cycle steal anything from the other.  They both had all the same opportunities.  I didn't, say, arbitrarily decide to multiply the ethanol cycle's yield (and only it's yield) by 2... which is exactly what you're doing in this first example with the stats for the country that's half the size.

 In case the subtleties of this dissection are lost on you:  The question I actually asked in the cycle comparison did not depend on a function of time.  If it was true at one point, it's true for all points.  Why could I get away with that?  Because I built cycle-averages into the construction of the calculations.  Let's pretend a moment that the comparison was actually close instead of the blow-out it was.  I didn't care about momentary fluctuations in the answer.  I didn't care which cycle was better for the first 30 seconds of the day, the last 30 seconds of the day, or at any single arbitrary timeslice that did not encompass the entire cycle.  Because, like your examples manage to accidentally point out, sometimes the system you're modeling with has some variation built into it and you don't want to know which one is best in 10 minute increments, you want to know which is better overall.

 In the end when you model conceptual industrial processes, .you don't start off doing industrial process engineering.  You don't model resource delivery delay, you don't model thermal flow across a proposed layout, etc.  That's a *lot* of math and at this point you don't even know if doing all that math is even worth it because there's a chance you could work days to get all the to the last 10% of all that ugly math and find out the whole idea is worthless.  So you start with things like cycle averages, like I did.  The math is a lot simpler.  You don't care that the process needs... say, 30 seconds to spin up.  You don't care if it takes 2 days to prime.  You're modeling the average day of operation because if the average day can't do what you want, there's no point in even looking at edge cases.  None.

 Similarly... efficiency is either unit-less or a ratio between two different resources that are related in the regard that one is an input, the other is an output.   Examples: a heat engine, say a refrigerator, is able to remove 10W of heat at the cost of 30W of power.  10W/30W = 0.33 <- no units.  Similarly a gas engine has a number of inputs other than just gasoline. and a number of outputs.  Some of these comparisons aren't terribly useful.  How many quarts of oil the engine cycles per second is largely only relevant to the engineers designing it, provided the finished product matches expectations of how the engine acts over its lifespan.  But how many gallons of gas or diesel it uses to produce X horsepower at Y RPM is more useful to more people... and how many gallons it takes to drive a mile?  Everyone cares   Your example 1 X/Y technically resolves into an efficiency rating.  How many people per area.  My X/Y does not because X and Y are already efficiency ratings and the division just gives you the ratio between them.

 Lets take the engine a step further since the other examples you gave don't get any closer to the question I asked with that comparison.  Maybe this will help..  I compared two different engines in the same circumstances. whether you want to imagine them both on stands or both mounted in the same body and then driven/run for the same amount of time,  The point is the comparison I made could be stated as 'Engine A gets X mpg.  B gets Y mpg.  Which is higher or are they equal?"  And you're complaining that 15 MPG is lower than 30MPG.   And what your examples do (and thus what I think you think I'm doing) is mount those engines in two different bodies with wildly different amounts of drag, drive them for the same amount of time... and then you're pointing out that the high drag car that used the 30mpg engine got beat by the 15mpg engine... And you're missing out on the fact that the 15 mpg engine wasn't trying to push a 3 ton brick, while the other one was.  The way that test is put together ensures the crappy motor wins..

Still, I'll humor you and show you how I'd work your first example. Since this wasn't necessary for Ethanol, I didn't do any of the following analysis, but since you're asking a question I didn't in a way that I didn't.

Growth is 25 people per unit time.
Area is a fixed resource.so let Ax=1, Ay=2,
Let X be the calculated population density of country A
Let Y be the calculated population density of country B.
Let X(1)=25, Y(1)=30     <- functions now, not just variables.

The general form of population density that changes linearly with time: (StartPop+GrowthRate*time)/Area

This means X(T)=(25+25*T)     <- since Ax=1, there's no need to divide by 1.
Similarly Y(T)=(60+25T)/2  which simplifies to: 30+(25/2)T

Without even doing any math at all, you can see one starts off bigger than the other... and you can tell that the other will speed past it because the other is growing twice as fast.

 Seeing as there's no need to ask which country has a higher density every unit time like you did, we'll cut to the chase and solve the entire system directly:.. by setting them equal to each other.  They use the same variable, and we can do this because we're not solving for X or Y.  We're solving for T. Why are we solving for T?  Because this will give us the time they're equal to each other.
25+25T=30+(25/2)T
25T-(25/2)T=5
(25/2)T=5
T=10/25 = 2/5

 Now we know precisely when the population density inversion happens.  Just by reading the problem, we knew Y starts larger, so when we hit T=2/5, we know they're equal, and after that we know Y is smaller.  Country B (Variable X) will be larger from T=2/5 to T=infinity.. There's no need to even contemplate total populations or do literally any other math if all we're asking is which density is higher.  You now know everything worth knowing about this example without having to iterate between time steps like you did.   This analysis methodology works on literally any function across it's range that is fully defined and continuous.

You'll note the comparison was still X vs Y, exactly as I did it for the ethanol cycle. I didn't do X/Y because there's no reason to do so here. I'm not comparing efficiencies, I'm just comparing densities.  You *COULD* normalize and do a comparison if you were curious, but that's going to give you a unitless number that's useful for only comparing the two densities at precisely and only that moment. I'm not going to repeat myself, you can scroll up and reread the normalization bit if you haven't gotten it by now.

All the math I've posted is trivially simple.  It's not even high school level difficult. All it does is sum up all the various inputs/outputs and spit out the net yield for every resource. And the comparison only concerns itself with power per unit water, because, as you originally asserted so many posts ago, you figured it would be more efficient to make a dupe run a wheel the whole day than run ethanol.  And it's not.  Not only is the wheel less water efficient, the wheel only produces power, while Ethanol can be modified to suit needs to provide H2O, PH2O, Pdirt, or dirt at the expense of power.  The wheel can't do any of that.

Shifting gears:

Big picture? Everything I've provided, all the data I've produced and posted is useful in a cost/benefit analysis.  Everything boils down to simple economics.  If you have a small amount of spare water you don't want to chill (103.3kg/cycle) and you want some combination of power, PH2O, Pdirt, or dirt, ethanol is your best friend compared to the wheel.  And, if my premises were correct in my last post, if dirt is *anywhere* in your list of things you need, it's your best friend, period.  Accept no substitutes. 

You might note, I'm arguing economics now, not math.  Overall efficiency in one resource does not matter if the perceived value of that resource is not high enough to justify the effort conserving it... This is part of why water is so cheap to buy and is so freely available in basically *any* developed nation.    To complicate that a little, this is all a cost-benefit analysis of a single process/system construct that is very much plug and play into a larger running system.  If you are drowning in water, getting rid of it is a good thing.  In fact, the more your water supply exceeds your demand, the lower the acceptable efficiency can be without spelling doom.  And ultimately, if an option is the best way to get a resource you *don't* have, efficiency of a resource you already have plenty of gets to sit quietly  in the corner unless this new process is so outrageously inefficient you'd be a moron to do it in your circumstances.   As I have said repeatedly, the existence of more efficient options for one resource does not impact the overall utility of any other option. You cannot ignore opportunity cost.

Bigger picture, relevant to your first post in this thread?  You said the process needs a buff.  Why would you buff a solution that's *already* the best in class for specific applications?  I'd have to build out similar cycle models for the other energy generation methods to double check this, but in terms of value per unit input (ie. water and dupe labor), ethanol is already strongly placed against literally every other power cycle.  It doesn't just give you dirt better than any other solution, including space flight, it gives you energy, PH2O, H2O, and Pdirt, too.  It's the only one with that sort of scaling and flexibility... and the only one that provides that volume of dirt.  Economically speaking, if you need Pdirt/dirt, it's a no-brainer.  Nothing competes...  so, in that light... why would you ever buff that?  It's already a binary solution.  Need dirt/PDirt? Build it. Don't need dirt/Pdirt? Building it is optional and how good an idea it is to build it anyway depends on how closely you can configure ethanol to suit the rest of your needs.  In the end, nobody is holding a weapon to anyone's heads and forcing them to build a functioning ethanol cycle.  People should build it if it makes sense, and I've done a pretty darn thorough exploration of all the data, publicly, in this thread so people could decide for themselves if it makes sense. You say there's better alternatives to the ethanol cycle when it comes to water usage, what are they, what do they produce, and *when* are they preferable?

 You asserted that the wheel was better.  It's not and I've provided ample proof to that effect.  You didn't find flaws in the proof.  You instead spent 90% of your replies basically muttering to yourself about how aliens shouldn't wear hats.  The other 10%? Useful because they made me check my math. 

 When you said a buff was needed, and gave suggestions, your initial justifications were a bit weak, but they were arguable, so argue them I did. .  Lowering the PH2O requirement for the trees would make the cycle water positive.  The cycle started that way and was deliberately nerfed to avoid that. Why would they undo that?  And given that other than dupe labor, water is the only price you pay here, why would they allow such a broken design to exist?.  Provide 1 dupe, get 1.3KW untuned + infinite dirt and water.  Surely you can see how ridiculous that is?  Changing power requirements on distillers, etc might very well make ethanol more power efficient than straight petrol (Disclaimer: I'd have to check my math on that :p )  What else is there to say that hasn't already been said here or the preceding paragraphs?  It simply does not need a buff and every suggestion you've made to essentially boils down to something that sounds like you just want an "I win" button you can mash so you don't have to think about things.

 

17 hours ago, DarkMoge said:

In fact, I am the one who tries to normalize this information by trying to give water some value. How much value can we get from water by feeding it to dupes and how much can we get by feeding it to ethanol chain. My argument is that feeding water to dupes running on hamster wheels is a better use of water than feeding it to ethanol chain. I picked hamster wheels, because any noob can place a hamster wheel and make their dupe run on it. They can start optimizing that system as they learn the game and they end up with a better use of their water.

 By trying to give water some value? I'm sorry, I'm having some issues wrapping my head around that.  Making the statement of W/kg H2O immediately assigns value to water.  It explicitly states that 1kg water is worth however many W when used in the process.   More than that, it also states that 1kg water in that process is worth all the other resources you get in exchange for that.  By the very definition of value used in economics, what I did was explicitly an evaluation of value.

 The math that you haven't bothered to try to understand shows explicitly and in great detail precisely that the value of water is less with a wheel than it is with ethanol in direct terms.  If you've forgotten, your wheel is worth precisely 1.89 W per KG H2O per cycle and because you have a dupe, you cannot eliminate water usage. 

Meanwhile, Ethanol in the configuration I posted gets you:
4.94 W / kg H2O
0.917 kg dirt / kg H2O

 Want all the dirt or Pdirt you can handle?
2.39 W / kg H2O
~4.22 lg dirt / kg H2O

There are other configuration options, but really, do I need to keep beating that horse?  Even when you choose the least power efficient configuration, ethanol still wins

There are plenty of clues in this post alone that you entire way of looking at this is backwards.  Grab one and use it.  If you were wrong about this, what else are you wrong about?  Go explore, see what works that you swore couldn't. Actually have fun doing it because you're learning new things instead of blinding yourself to the possibilities.  Otherwise, all there is is frustration because you can choose to ignore reality, but you cannot ignore the price of ignoring reality.

Speaking of frustration, do you know what's been honestly frustrating about all this? It's not that your premises are wrong. They are.  It's not that your assertions are wrong.  They are.  And it's not even that the most important parts of your math are wrong, because they are too. Nobody can be correct all the time.   It's the fact that I've written several thousand words, plus proofs, and you can't admit you're wrong.  Despite the sharpness in my wording at times, I've still given you the benefit of the doubt the entire time.  I assumed things like this was an honest conversation and you weren't trolling,. (Reconsidering that one, honestly)   I assumed that if I could find the right way to lay it out and explain it, you'd be able to say "Oh, that's where I messed up.  My bad" and honestly mean it ... not even publicly, but just to yourself... Whether or not you want to admit it you are, in fact, wrong.  And not being able to admit when you're wrong is a pretty bad problem to have.. 

Link to comment
Share on other sites

30 minutes ago, storm6436 said:

And, if my premises were correct in my last post, if dirt is *anywhere* in your list of things you need, it's your best friend, period.  Accept no substitutes. 

I agree that ethanol generates more dirt than any other system we can currently build in game. In my first post, I basically gave the ethanol chain a name "dirt generator 2000". Though, when I start to look for something else out of that process, I am starting to be upset.

Looking for pH2O? The process is water negative unless you use wild arbor trees... Which means that either wild trees are overpowered or domestic trees are underpowered. Looking at all other wild plants... having them as wild does not seem to be as of a big deal to me, so I feel that it is domestic trees that are underpowered. Hence, my suggestion to buff them.

Looking for CO2? Well, we actually do not have enough ways to play with that quantity of CO2... We only have slicksters to play around with and to support slickster farm, ethanol is just too much... Hence, maybe, there should be a way to play around with CO2... maybe, some sort of carbon compressor that lets you make diamonds out of CO2.

Looking for power? I value W or J too little and H2O too much in comparison to ever take that trade. Because water is duplicants... I even value duplicants way more than just a dupe running on a hamster wheel. And my argument was that I could try to squeeze out more out of dupe running on hamster wheel than from ethanol chain. But that aside... I am tired of talking about it.
 

1 hour ago, storm6436 said:

The cycle started that way and was deliberately nerfed to avoid that. Why would they undo that?  And given that other than dupe labor, water is the only price you pay here, why would they allow such a broken design to exist?.  Provide 1 dupe, get 1.3KW untuned + infinite dirt and water.  Surely you can see how ridiculous that is? 

I have not been playing the time, before nerf. But that tells me that there is inherently something wrong with the whole process than. Why would we argue that reducing cost of maintaining the domestic trees would make the ethanol overpowered and provide free water when it can provide free water with wild trees? If you believe that being water positive with domestic trees is wrong than you should probably try to push for the idea of nerfing wild trees to the point where ethanol becomes water negative with them.

Free resources is not a foreign idea to the ONI, we have geysers providing water and power for no cost, just build once and get free goodies. We also have critters that end up providing free resources. So, the whole idea of free resources is not ridiculous. The question becomes how much of free resources and the method of obtaining them.

1 hour ago, storm6436 said:

Why would my X/Y be unitless when yours isn't?  Because my X/Y is a normalization function.  Provided X and Y are positive and X>Y, then X/Y>1, otherwise 0<X/Y<1. 

I believe, I know where you have lost me. What you think I was referring to with X/Y is the comparison of 2 final ratios of efficiency. My variables have units and they have units for a reason... Because this whole time I am trying to point out error in calculation in the different area. If I was talking about the comparison of the efficiencies, than any sort of Z would not matter.
A / B = d
if d > 1 than A > B
if d < 1 than B > A
when ever d is higher or lower than 1 does not change if we add any Z to both sides.

But that is where you lost me, I was not referencing that. And my math examples are attempts to try and point out your errors. You see, what you say is that food choice or how long dupe works in a cycle does not matter, because there is a dupe on both sides.
What I am referring to is:
A= X1 / Y1
B= X2 / Y2
now if we add Z somewhere in there, it will impact our result. And where do you add Z? If X stands for W and Y stands for water. You change food to bristle blossom, changing Y by adding more water requirement and claim that it does not matter. You change how much power we get from electrolyzer, changing X and claim that it does not matter.
What I am trying to point out is that you fiddle during the calculation of the efficiency, getting the wrong A and B to compare later.

2 hours ago, storm6436 said:

Speaking of frustration, do you know what's been honestly frustrating about all this?

 

2 hours ago, storm6436 said:

You instead spent 90% of your replies basically muttering to yourself about how aliens shouldn't wear hats.  The other 10%? Useful because they made me check my math. 

I have spent hours upon hours and typed thousands of words. And every time, I feel like you do not read my posts or do not try to understand them. There are times when I am wrong, but for the past several posts, I have been trying to point out a fault in how you did your calculations. And what do I get in return? I get claims that 90% of everything I say is bs and claims that everything I say is wrong. 
When I simplified my attempts to point out the error in your statement to the level of school math just so you can see it. You still seem to be deliberately ignoring it and still believe that I am talking about something else. How do I explain to you that changing ratios inside the system changes its efficiency and than when we compare efficiencies, we get different comparison?
 

Link to comment
Share on other sites

16 hours ago, DarkMoge said:

I agree that ethanol generates more dirt than any other system we can currently build in game. In my first post, I basically gave the ethanol chain a name "dirt generator 2000". Though, when I start to look for something else out of that process, I am starting to be upset.

Looking for pH2O? The process is water negative unless you use wild arbor trees... Which means that either wild trees are overpowered or domestic trees are underpowered. Looking at all other wild plants... having them as wild does not seem to be as of a big deal to me, so I feel that it is domestic trees that are underpowered. Hence, my suggestion to buff them.

Looking for CO2? Well, we actually do not have enough ways to play with that quantity of CO2... We only have slicksters to play around with and to support slickster farm, ethanol is just too much... Hence, maybe, there should be a way to play around with CO2... maybe, some sort of carbon compressor that lets you make diamonds out of CO2.

Looking for power? I value W or J too little and H2O too much in comparison to ever take that trade. Because water is duplicants... I even value duplicants way more than just a dupe running on a hamster wheel. And my argument was that I could try to squeeze out more out of dupe running on hamster wheel than from ethanol chain. But that aside... I am tired of talking about it.
 

 I'd typed out a fairly long post only to accidentally lose it when I went  to check numbers and hit enter on the firefox search instead of middle clicking... *sigh*  I'm on a time crunch for the next few days so I don't really have the time to retype it all now :(

 Trying to summarize:  Central premise of economics is that an individual places different values on different resources, and no two individuals have precisely the same set of values.  I figured we valued things differently, but it didn't occur to me just how wildly different our values evidently are.

 I broke down the geyser sources on the current map I'm playing, (in kg/cycle) 1980 steam, 1716 salt, 1704 water, totaling 5.4T/cycle.  On the PH2O sources from NG and oil, I cleared my calculator, but from memory, if I tapped everything other than the leaky oil fissure and boiled to petrol (not condensing sour gas), I'd be looking at something like 10-13kw and a little over 3T PH2O/cycle.

 I tend to limit to however many dupes I can fit in 2 electrolyzers until I need the extra O2 for LOX.  In practice that's 16-20 depending on traits, but for performance reasons, I cap at 20.  That eats 1900kg/cycle (1200 kg from electros + 100 kg from flush toilets+(600 from showers) ...  Is that wasteful?  Depends.

 Early game, I'm algae+mealwood as long as I can make that last. Usually that means by time I'm running low on algae, I've tapped at least one NG geyser, set up my standby coal bank, and built my NG generator room that serves as my general ph2o reservoir, add a 20x15 insulated hot water reservoir with adjoining ~30x5 chiller, a facility for refining (usually 2 smashers with 2 metal refineries added later), and tack on a sewage treatment unit on top of the generator facility... I've also hollowed out and insulated pretty much the entire starting biome to use as my base proper and all the industrial is outside.  I'll build not not start my electrolyzer set up (the horribly inefficient AETN one mentioned upthread), and raid any ice biome that doesn't have enough sleet wheat to justify wild farming for all the ices.  When I do finally hit terminal algae, I just start the electrolyzer. 

This is usually the point where I jump to flush toilets and bristle berries.  Now that certainly sounds wasteful, right? Especially since I end up adding showers later as well.  But here's the thing, the last 6 maps I've played, by this point I'm not just drowning in hot water from whichever geyser I've tapped, but I'm drowning in cold water from all the ice.  Going a little water negative overall doesn't bother me, I actually need to scoot it out. 

More importantly, that doesn't cover sieving the PH2O from the 2 NG gens I'm likely to have or the toilets (No showers yet) I have piling up into the water equation.  Adding that in, since I have wattage to spare, and tapping that one salt geyser... depending on how may dupes I have at this point (still early/mid game for me so 8-12) still leaves me with a +/- 100kg/cycle surplus or deficit of water,   And I'm already sitting on 100+ tons of water at that point due to all the ice.  Yeah, it's not all water yet, but it'll melt before things run dry.  At that point,  I'm generally tech-ing up and finally killing slime biomes for space or gold.  I won't bother with my needlessly realistic approach as it's *way* over-paranoid for what they nerfed slimelung into.  It suits me to keep treating slimelung like it's anthrax on steroids.  Based off what you said about not playing before Ethanol was water positive, if that means you started after (as opposed to just not playing during), old slimelung was lethal if you messed up containment. Outbreaks of the original version were... impressive.  Back to the point, since I'm hitting swamps, I'm pulling a ton of PH2O, so I'm still water positive overall.

Eventually I have all my industrial spaces set up and all the ancillary bits... usually it's about this time that heat starts to be a problem Inside, my cold reservoir is 40-60F and the air piped in is around 60F as well, so no problems there.  Outside on the other hand... temps in some areas are over 100F, generator spaces likely 120F.  Most maps, I set up a central PH2O cooling reservoir fed from a slush geyser, and then pump that in loops to cover all the hottest spaces.  This map doesn't have a slush, so I'll have to improvise.  Will probably put all the P-ice in there instead and once it melts, I'll use that.  Once that gets too warm, I dump it back in the generator room to get sieved. 

Without a slush geyser, I'd need to immediately jump to my steam/tuner chamber, which is going to suck because I certainly don't have 4.8 KW to spare, even tuned.  Yeah, I can get by using just 1 which helps, but I'm keeping an eye at when I can use it all.  So, at this point I've already raided the oil biome for lead, now I pull oil, refine to petrol, and build a petrol chill chamber.  Once this chamber is cool enough from tuning, feeds the metal refineries (prior to steel I'm feeding them off my hot water res), it gets used to a small number of places.  It feeds the metal refineries,  runs the active chilling loop on my water chiller, and one other pump circulates through the PH2O chiller so I dump less or possibly none.  Tons of automation.

*THIS* is the point where I give ethanol the side-eye.  I can service 2 petrol gens off that refinery's output but I don't want to do that just yet, despite needing the PH2O to service what almost certainly is a water deficit now (due to adding showers).  Why?  Normally I still have a good amount of time in terms of breathing space before I get low on water in general and this is the part where I'm going to actually be worried about temperatures in the generator room.  Everywhere else will just fine, but 3 NG gens cranking away 24/7 does build up some heat.  My petro gen facility gets tacked on the side of the NG facility.  It's basically just an extension with a few modifications to control the CO2 spread to ensure the scrubbers don't miss anything.  Without a doubt, the moment those scrubbers go online, I am massively water negative.  Advantage?  It's hot water negative.  Still largely don't care, I have breathing room and heat is more my concern.  Add the extension + 3 petrol gens, build ethanol cycle.  Route tree feed to pipe via conductive across all the generators.  Generator temp problems solved and staying solved.  *NOW* I pump petrol to the other two and automate them for high load... and the moment that's done, I can run 4 tuners constantly.  Barely, but I can.  And now that I have an excessive amount of heat deletion, I tap another water source... and now I'm hugely water positive while still having 2 untapped sources worth ~2T/cycle.

 This is usually where my efforts shift to going to space because I need to ditch the petrol in the chiller for super coolant..

 Now, I'm pretty sure you had a few moments where you blinked at what I typed and was all "Why the hell would you do that?"  That's because I don't have a fixed value for water.  My value for any resource varies entirely on how much I have on hand, whether I'm positive or negative on it, if negative, how many cycles I expect to pass before I run out, and two other factors:  are there untapped sources left and what do I need to do to tap them.  If there are easy sources left, value goes down a lot. 

I'd like to think I was thorough but brief enough to give you an idea how my value for water changes as the game progresses. It's very high value in the first 100 cycles, but the moment the ice and an external source rolls in, the value goes down and that's when I shift to toilets/bristles.  Also, I'll point out that during all this, I set up stone hatch ranches (The sage comes once ethanol is up) to try to keep my coal bank as close to coal neutral as possible, and as the meat output goes up, I start uprooting bristles to try to maintain roughly the same amount of kcal in the fridges.  That's usually around 500k/700k. 25-35 cycles of spare food isn't bad :p  Either way, my net water usage from bristles starts going down within 50-100 cycles of starting them... Fertilizer, wild grown sleet wheat, wild pinchas, they all drop that water usage since I'm looking at maintaining kcal stored.   I'd like to think doing it that way makes the best use of things.  Eventually I hit frost burgers ... because at that point, why not?  If I have a chlorine geyser, every bit of that is renewable.

Quote

I have not been playing the time, before nerf. But that tells me that there is inherently something wrong with the whole process than. Why would we argue that reducing cost of maintaining the domestic trees would make the ethanol overpowered and provide free water when it can provide free water with wild trees? If you believe that being water positive with domestic trees is wrong than you should probably try to push for the idea of nerfing wild trees to the point where ethanol becomes water negative with them.
 

  It's not that it's "wrong" per se, but incredibly unbalanced.  Back in the day, NG+fertilizer synth balance was so out of whack that some folks would literally go from wheels to a large bank of fert synths and never need another power source because they could scale the synths all the way to 20kw power output... and automate the entire thing so it was literally a sealed box that crapped power and fertilizer.

 Making ethanol water positive would basically do the same thing, except it would cost less than 1 dupe's worth of labor.instead of eating dirt like the synths.  So instead of "I get 1.3KW untuned for 200kg water" it becomes "I get 1.3KW untuned per dupe."  With a little support equipment and a computer that isn't slow... 10 dupes who do nothing but ethanol would output 13KW, and tuned 19.5 KW which is actually understating it a bit because I just multiplied by 10.  Since ethanol is configurable, you can literally tune that output to more or less how much dirt you want after paying trees, and shuffle the rest into KW too.

 At that point, it's functionally the same infinite energy machine the fert synths were.  And most people would gravitate to doing precisely that because once you get the first set-up running, everything else gets easier.  It trivializes a huge portion of the game because then there's no question, it's the best source until you hit solar...  except, if you have 10 dupes to spare, why bother building solar? You already have 20kw for infinity.  If anything, skipping solar means your automation for steel doors suddenly gets a lot easier.  It's not "Keep the doors open as much as possible", it's just "Rocket landing/launching?"

 

Quote


Free resources is not a foreign idea to the ONI, we have geysers providing water and power for no cost, just build once and get free goodies. We also have critters that end up providing free resources. So, the whole idea of free resources is not ridiculous. The question becomes how much of free resources and the method of obtaining them.

I believe, I know where you have lost me. What you think I was referring to with X/Y is the comparison of 2 final ratios of efficiency. My variables have units and they have units for a reason... Because this whole time I am trying to point out error in calculation in the different area. If I was talking about the comparison of the efficiencies, than any sort of Z would not matter.
A / B = d
if d > 1 than A > B
if d < 1 than B > A
when ever d is higher or lower than 1 does not change if we add any Z to both sides.

But that is where you lost me, I was not referencing that. And my math examples are attempts to try and point out your errors. You see, what you say is that food choice or how long dupe works in a cycle does not matter, because there is a dupe on both sides.
What I am referring to is:
A= X1 / Y1
B= X2 / Y2
now if we add Z somewhere in there, it will impact our result. And where do you add Z? If X stands for W and Y stands for water. You change food to bristle blossom, changing Y by adding more water requirement and claim that it does not matter. You change how much power we get from electrolyzer, changing X and claim that it does not matter.
What I am trying to point out is that you fiddle during the calculation of the efficiency, getting the wrong A and B to compare later.


Using that structure, my comparison is (X1/Y1)+Z  = A and (X2/Y2)+Z=B  If you use something like Xo and Xi instead of X1, Y1, it'll make keeping track of things a lot easier because with that the big letter tells you which system or entity you're tracking, and the small letter tells you what part of that entity it is. Using that  style of notation, what I did was (Xo+Zo/Xi+Zi)+and (Yo+Zi/Yi+Zo)  I'll point out this works better on paper than text.  I'm just happy this isn't quantum or E&M.  You have variables written with super and subscripts, keeping track of what the notation means before you even get to doing math gets ugly, fast.  

 I might add more later, I'm out of time for now.

Link to comment
Share on other sites

7 hours ago, storm6436 said:

if that means you started after (as opposed to just not playing during), old slimelung was lethal if you messed up containment

 

7 hours ago, storm6436 said:

It's not that it's "wrong" per se, but incredibly unbalanced.  Back in the day, NG+fertilizer synth balance was so out of whack that some folks would literally go from wheels to a large bank of fert synths and never need another power source because they could scale the synths all the way to 20kw power output... and automate the entire thing so it was literally a sealed box that crapped power and fertilizer.

Let me clarify, I have played the game since the earliest days... Like, as soon as alpha was available, I purchased it. What I mean is that I did not play during every single update it ever got. It usually goes like this... I play it for a while, than I get bored... Than I remember that this game is a thing and check out new features. The game updates more often than I come back to it... And well, I am well familiar with building fert synth for power. And I am well familiar with slimelung being deadly... You know, I have played this game, before germs even existed, back than, digging through slime biome and letting your dupes live on pO2 from slime was the best strategy. Than, they quickly introduced slimelung, but took ages to introduce any sort of medicine.

Link to comment
Share on other sites

Later:
@DarkMoge So while I was out running errands, a few things occurred to me.and rereading the last section of your post bounced off a few of those things.

  Not trying to beat a horse to death here, just recapping a very small section for context, specifically the benchmarking engines example. Now, this is a little oversimplified but thematically it's accurate.  In real life, engineers can test two different motors on a stand and be pretty accurate in estimating how it'd handle honest road conditions.  This is because they make more than a few assumptions about what kind of vehicle it will be mounted in, the weight of the driver, etc. But by pretty accurate, I mean, within a few percent.  Part of the reason for this is because they have a ton of performance data from all the other motors they've built, and then for them it's a question of mathematically taking the performance curves observed on the stand and essentially using their data stockpile to translate that into a "This is what it should look like on the track" curve.   

Unfortunately we don't have that kind of data for ONI and extra-unfortunately, real life is a lot messier than the lab and the paper calculations regardless of if you're talking ONI or the motor..  What direction the wind is blowing, how frequently it shifts, and how hard can affect mileage, gravel on the track,driver's driving habits,  etc etc, There's a ridiculous number of factors .  All the performance data they've already gathered is subtly marred by that and it's only through significantly large data sets can you mostly overcome it.  You never quite overcome it though, you just get progressively closer until you hit a point where you figure that's close enough.

 But at the end of the day, once they've done all the math, someone mounts the motor in a car and they go do donuts for hours to collect more data.  A bunch of calculations are made to determine how things went and all the data gets fed into that database. 

Now here's where things get a little interesting. Car  Manufactures/sales types are required by law to post a fair amount of numbers that consumers care about on the vehicle for sale..  Mileage is one of them.  But how do they define mileage?  Sounds like a stupid question, it's miles per gallon right?   Not exactly.  Back before the industry standardized the tracks used to test mileage, the inconsistencies between posted and real-world efficiency  were wider and they varied by manufacturer. But the standardized track barely resembles driving in the real world, so posted mileage is never what we actually get.   You're almost certainly not driving the same way the test driver did, much less weigh the same (or even close) or fit any of the other test factors.

 Thankfully ONI is far more predictable than real life, though random gas deletion bugs and the like do spice things up a bit. Using the car analogy... was take the cycles out for a spin and used the same driver for both.  If I'd driven the ethanol cycle with an 800 lb driver and wired the wheel with an RC controller so it basically didn't have a driver, it'd be kinda silly to expect a fair comparison, right?  Much like mounting a 30mpg (on the stand) modern engine on a 3-ton brick and racing a 15mpg one on a roller skate. (Don't ask me how.)  Because those conditions are massively different, the results won't reflect real life at all.  

Using the same driver, same chassis/body, and only swapping motors out, the test becomes motor vs motor and the results aren't "This motor gets 30mpg" they're "Under these conditions, its reasonable to expect 30mpg"  So the unwritten caveat for efficiency comparisons everywhere is "Efficiency of Motor A driving 2klbs w/150lb driver in 90F summer air" compared to "Efficiency of Motor B driving 2klbs w/150lb driver in 90F summer air"  It's just when you read the posted mpg, they don't tell you what those conditions were.  With my math, I was extremely explicit by providing all the math.

So when I said it didn't matter what I did so long as it's done to both sides, I wasn't kidding and I wasn't wrong.  Conditions matter to what you see when you use it, whether it's a car or the ethanol cycle or even the wheel, but I wasn't "using" either, I was calculating the posted rating using standardized metrics to ensure a fair comparison.  Real world experience would vary a little because the game does account for things I didn't.  Shipping takes time.  I didn't model that. Lots of other things take time or might happen if you're running it on your own game that would jiggle the numbers a bit +X kg H2O usage here, -Y wattage there.  Dupe stops to pick his nose instead of grooming hatches.  Massive pip stampede through the kitchen.  That's also why I calculated the cycle average, not any instantaneous numbers.  Real world experience, if you took multiple cycles worth of data using the exact same config, and averaged them, that average should be remarkably closer to my numbers than the individual cycle's numbers.. The more cycles the closer. 

 Ultimately, by keeping the conditions the same, you get a baseline of what to expect if you build it yourself with the same conditions.  Will your conditions vary? Yeah, but if your only two choices are the wheel and ethanol, no matter what your conditions are, one is more efficient than the other.so if efficiency is important, the choice is clear. 

  Though, even if Ethanol is more  efficient per kg water, who cares?  If you don't need the watts, why do you care about W/kg?  The comparison is useless to you.  Similarly, let's pretend a moment that the wheel used the same amount of water as ethanol but had the same efficiency it did in the comparison anyway.  If you were had plenty of dirt, water, and electricity, why would you build either? Or, going back to the numbers I calculated, what if you needed wattage but you didn't have the water ethanol needs and, for some reason, the wheel was your only other option (or your best option... somehow)   Even if it's less efficient, the fact you can't pay the ante for ethanol means the wheel's your only choice, so wheel it is.

Economics and the study of how people make choices and comparisons is actually pretty interesting.  I *highly* recommend Basic Economics by Thomas Sowell.  The guy is very thorough and it's not nearly as bad as reading one of my posts. :D

Wrapping this up here:

Let's ask a different sort of question.  Instead of "Which cycle is more efficient?" let's ask "Which cycle is better?"  They're not the same question.

: Say you somehow had a dupe that magically refunded all the water costs for just the wheel cycle, or reduced it to *almost* zero.  Say, less than 1kg cycle.  Which is more efficient if it's zero?  Which is more efficient if the wheel only uses 0.1kg water?  Certainly that would *dramatically* change the numbers.

If H2O=0, you can't even divide W/kg H2O, you're dividing by zero
If H2O=0.1, then: 257.8W/0.1 kg H2O -> 2578 W/kgH2O.  Impressive numbers, right?

And that's the point where we run into an interesting facet of economics where a full discussion would involve talking about scarcity, market price, market demand, market supply,  etc... but I wasn't kidding when I said I was wrapping up, so I'll skip to the punchline:  Numbers lie  Or to be a little more specific, the math and the numbers it provides don't lie... We do.  To ourselves. A lot of people see numbers and think they know what those numbers mean, what question they're answering, etc when they don't.  Sometimes this isn't their fault.

This is why efficiency by itself doesn't tell you the whole story.  This is where the current value of resources comes into play. 

If you hold a resource at a constant value, if that value is sufficiently high you'll do anything to conserve it, anything to make more of it.  What if it's already abundant? Imagine someone who was wrapped up on their lead stockpiles, to the point they ignored steel and everything space-age.  Maybe even the rest of the metals, too.  This is why variable value (and also variable market prices) is such a vital thing to proper decision making. I mean,

 Similarly, as you can see in the 0.1 example, efficiency rating by itself doesn't tell you exactly much of anything.  It's not going to tell you that you're only going to get 400W, and it doesn't tell you how much water's involved either.  It could get you 1W and cost a few mcg of water, or it could cost 100kg and crap out 257800W, So, caveat emptor, a 1000mpg motor doesn't do you any good if all it can push is an ant.  That's why we don't rely on efficiency alone.  Motor-wise that's why you check horsepower, etc. 

ONI-wise, that's why you look at *everything* it takes in and gives out.  When you look at everything, you're already mentally gauging efficiency, comparing your personal value for each resource and trying to make the decision of which one is best. 

 Say we have 5 people who need a solution to a problem and they're all looking at a random process that really only has 1 input and 1 output.  What that the right logical representation for what these people are asking?  Oddly enough, structurally it's identical to (X||Y) that I've used a number of times so far... but what's packed into that X and Y isn't the same.

 So, for efficiency, if Xo=Output of X and  Xi=Input of X, then we can say the efficiency of X is Xe=Xo/Xi.

 What those 5 people are doing?   They're asking which is better... so what's Xb?  Well, it's the sum of how much they value each resource times the amount of resources.  So if we have n number of resources,, Xb=[SUM(V1*R1+V2*R2+....+Vn) / SUM(V1*R1+V2*R2+....+Vn), Vs are restricted to positive real numbers. 
 
 That might look like it should be 1 since it appears to be a sum divided by itself, but it's not.  If you have simplified/reduced all your figures, R1 is only going to be on the top or bottom, never both.  Same with the rest of them, so you'll end up with a number on top and a number on bottom..  Repeat this for process Y and then compare the two just like we did the cycles with (Xb||Yb).  Larger number wins. 

Most people even use the same logical construct to compare a single process.  To what?  To the perceived benefits of doing it or not doing it which don't have to reflect reality at all.  Perceptions are weird things.

 And to cap this, R1 through Rn aren't just resources you can stick in a storage container of some sort.  Space is a resource, so is time.  Anything a person values that's involved with the decision gets it's own R#.  So, when I said earlier I don't do wild arbor trees, it's because my value for cleared space and the amount of time involved to make it actually work properly is higher than my perceived benefit of using it since I'm not playing on a map with a forest biome at the moment. 

Anyway, hope this helps.


 

3 hours ago, DarkMoge said:

 

Let me clarify, I have played the game since the earliest days... Like, as soon as alpha was available, I purchased it. What I mean is that I did not play during every single update it ever got. It usually goes like this... I play it for a while, than I get bored... Than I remember that this game is a thing and check out new features. The game updates more often than I come back to it... And well, I am well familiar with building fert synth for power. And I am well familiar with slimelung being deadly... You know, I have played this game, before germs even existed, back than, digging through slime biome and letting your dupes live on pO2 from slime was the best strategy. Than, they quickly introduced slimelung, but took ages to introduce any sort of medicine.

 Didn't want to assume when you played so I figured I'd cover all possible bases.   I'm sure my methods make total sense even if my refusal to change them doesn't then.

 Though, you might not share my dislike of the fert synth spamming.  From my perspective, it trivialized the game because no matter what your map looked like or anything, the solution was always "Spam fert synths for the win."  That's boring, and even if I didn't *have* to spam fert synths, always having that option if things went to hell detracted from my sense of "I really have to consider my options and be smart about them."  3/4 of my interest in the game is the complexity of the possibilities and being able to model, implement, then fix/optimize using those complexities.  I'm kind of a sucker for Factorio and KSP as well.

Link to comment
Share on other sites

Worth noting, I corrected the math on the tail end of the reply to where you asked about the math earlier today..  (Xo+Zo)/(Xi+Zi).  I was in a bit of a rush and distracted.  Didn't stop to verify.  What I originally posted makes no sense mathematically.  I think I still explained it correctly elsewhere though.  It does bias the exact numbers, but since it's based off the conditions that define Z and Z is added to both sides, the resulting number is a valid comparison for the configuration/conditions being described.  Different conditions are obviously not part of the model, so the efficiency derivation at that point is merely a starting point or approximation.  The accuracy of the original direct comparison gets worse as you change conditions, but the overall result A>B or A<B will not change provided the comparison remains valid.

Ie. if A>B, then (Ao+Zo)/(Ai+Zi)>(Bo+Zo)/(Bi+Zi)

Notes due to construction:   If Zi = -Bi or -Ai. it invalidates the comparison.  Similarly, the relationship between the paired numbers (Ie.[Ao+Zo], [Ai+Bi], etc) must remain positive.  Negative values invalidate the comparison because negative outputs aren't outputs, they're inputs.  Negative inputs aren't inputs, they're outputs.  Beyond that, comparisons require the same units.  You can't compare W/kg to W*kg, just the same as you can't ask if twenty kg is greater than 2 seconds.

  However, it's also important to note that if the Zs are sufficient to flip both sides, it does reverse the comparison, but then you're not asking W/Kg, you're asking Kg/W, so it makes sense the comparison would flip.
 

Link to comment
Share on other sites

How do you all have so much energy to type 20+ paragraphs about this XD. Without an in-depth analysis it's obviously resource positive, but the scale to which you need to set it up exceeds reasonable expectations. Consider that half of your power is consumed creating the fuel itself. That would be enough to prove my statement. It depends on how much you value every other byproduct.

Fertilizer synth was pretty broken. You should at least require a multi step process like ethanol or sour gas. Perhaps it's not that ethanol feels weak, but that petroleum and natural gas feel strong. Would be nice to find a good middle ground.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

Please be aware that the content of this thread may be outdated and no longer applicable.

×
  • Create New...