Semi-starvation ranching of Pacus

[blakemw]

Pacus have undergone many developments over the years. One idea I've not seen explored, is semi-starved pacu.

The Key Principles

• When a Pacu eats from a Fish Feeder, they gain the "Ate from Feeder" buff which awards +5 happiness for 2 cycles, regardless of the amount of food.
• A Fish Feeder cannot be set to below 1 kg, but a Conveyor Meter can split off algae into 1 gram pieces and a Sweeper can load those gram pieces into the Fish Feeder, the normal 7 kg of algae a pacu eats per cycle, could provide the basic "Ate from Feeder" buff for 14,000 cycles.
• A Pacu starts adulthood with 450 kcal, and uses 100 kcal a cycle. This means it has 4.5 cycles of happiness even without eating food.
• A happy Pacu lays an egg every 1.5 cycles, so it can lay 3 eggs from its starting kcals.
• A glum Pacu lays an egg every 15 cycles, so it can lay a 4th egg soon before dying.
• So in 25 cycles, the Pacu should replace itself and export 3 eggs, and 1 Pacu Fillet.
• If cooked that would be total of 18400 calories or roughly 736 calories/cycle, 4 pacus would very nearly feed 3 dupes.

Distribution of gram algae

Gram algae is made using the Conveyor Meter, though you don't need 1 Conveyor Meter per Fish Feeder, as the Conveyor system has predictable packet branching behaviour, so you emit 4 packets and have the Conveyor system distribute them.

 

Though it must be noted that Conveyor Meters are finicky about emitting only 1 packet, even a 0.2 second green signal can cause the Conveyor meter to emit two packets, that's because it can immediately emit the packet, and then the signal is still green which resets it again. You can use a Power Shutoff to make sure the Conveyor Meter is always powered off when it receives a green signal, then it can only reset once (it doesn't need to be powered on to acknowledge the reset signal). Personally, because in this case the gram quantities are so miniscule and there are no bad consequences to emitting an extra packet, I just make sure that at least the desired number of packets are emitted.

The Greedy Critter Problem

The Greedy Critter Problem is that critters don't take turns eating limited amounts of food, there is a "simulation order bias" which is generally that the oldest (first created) critters get dibs on the food.

However if two pieces of food are added in rapid succession, the critter which starts eating the first piece of food is locked in the eating animation and cannot claim the second piece of food, allowing a second critter to also eat. However by the time a 3rd piece of food is added, the first critter is finished and ready to eat it again. So it seems that it's only possible to reliably feed two critters with one feeder.

However a pacu can only immediately claim the food if they are within 9 tiles of the food, so if the pond is at least 10 tiles wide then when the greedy pacu randomly swims to the far side of the pond that will definitely allow another pacu to have a turn eating. Of course, you need to emit a lot of extra gram algaes to give more opportunities to eat.

In my current real game, on the Marshy asteroid I collected most the water in the bottom of the map and have two fish feeders, spaced 15 tiles apart, I'm emitting 4 algae every 40 seconds (the algae stockpile will last over 100,000 cycles at this rate) and all the Pacu are eating just fine.

So big ponds can definitely work.

The effect of late game critter lag, and the Fast Track mod, need further investigation. It does seem that in the late game the critter lag makes it more possible for critters to take turns, while Fast Track doesn't seem to have any real effect.

 

 

For example in this communal Pacu tank from my current game at cycle 530 there tend to be between 7 and 10 pacu, and between 2 to 6 tend to be eating at any one time.

My thinking is that if you want many pacu in a large pond, at least 10 tiles wide, you're better off just letting the pacu spread out so they'll take turns more reliably, I don't think there's enough benefit to confining them to a small area, even though it is possible for more than 2 pacu to eat it seems very non-deterministic.

Semi-Starvation Pacu Box

I cannot claim to be an eggspert in ranching Pacus, quite the contrary really, I just enjoy investigating potential exploits. I'm also not good at pacu boxes, I mostly use big open ponds.

But here's one of the I think simpler setups (and because it's simpler, not as good in some ways), also it's not fully compactified.

So what we have here, is 4 chambers, each chambers houses 2 Pacus, so that's 8 Pacu in total, which would feed nearly 6 Dupes. And in principle this build should only need 4 g of algae per cycle, though for practical reason it might be more like 9 or 13 g per cycle, anyway 100 kg of algae would feed 6 dupes for 11100 cycles, so it's economical on algae.

Naturally there are purely Conveyor solutions to replacing the Pacus, some simpler and less good, some more complicated and more good. But for this simpler one, I'm using Dupe labor to hug the Incubator to get a "variable Fry production rate" and also to drop the Fry in the fish release. Basically each Fish Release is set to 2 Pacu, and the Incubator only gets turned on if any of the Critter Sensors detects fewer than 2 Pacu, since the box houses 8 Pacu and Pacus live 25 cycles, it should need a new Pacu every 3.125 cycles, a Powered Incubator can crank out a Pacu Fry every 1 cycle, so with this solution, in the long term, the Box should only go without a Pacu for at most 1 cycle before the powered Incubator spits out a new Fry.

I'm sure a completely dupeless solution could just arrange for an egg to be hatched every 3.125 cycles (on average), using a Conveyor Shutoff or something to drop an egg on schedule, and I'm sure after a couple of generations the system could have a Pacu dying every 3.125 cycles like clockwork. Probably anyway.

Conveyor Layer is pretty simple, one set of rails which twice splits to distribute Algae, and another set to carry Filet and Eggs, I use a Conveyor Filter to pull off the Filet, this is obviously a crime because in the long run it'll use like a 0.32 watts on average while I could use a solution which will use no watts at all.

Automation involves a Critter Sensor in each box, which sends a green signal if there are fewer than 2 Pacu. That directly opens the door, and also gets merged onto a wire which activates the Incubator.

The other automation relates to releasing 1 gram alges. Basically, I use an AND Gate with two Timer Sensors, one sends a green signal once a second (0.2/0.8), and the other sends a green signal for 12 seconds (0.02 cycles) then a red signal for 1.98 cycles. This means that 13 algae packets will be emitted. This could be done slightly more accurately, but it really doesn't matter.

Other uses of semi-starvation

Pacus are the only critter which can get the "groomed" buff from being fed tiny portions of food. But for any critter it resets the starvation countdown, that used to be more useful than it is now because critters used to get reproduction even while starving and ungroomed, so resetting starvation would allow any glum critter to lay an egg during its lifetime. Now critters only gain reproduction while starving if they are groomed, and they do so at the bad glum rate making it a bad use of grooming labor, unless you can exploit the "Ate from Feeder" buff.

This means that resetting the starvation timer is only really good for keeping a critter alive in the hope of future times of greater abundance - except of course Dreckos, but you don't have much control over their portion size.

Conclusion

The semi-starvation Pacu ranch is one of the zero-labor solutions to producing food almost out of thin air, and overall its calories-per-tile is fairly good for a zero-labor food solution and it also doesn't need nearly as many critters as non-fed pacu boxes, which only produce 1 fillet per lifetime.