Learning attribute is no longer important?

[Inter]

For years players have made the observation that Learning is the most important, and perhaps the only important attribute a dupe may possess.  The reason for this is that, until the most recent update (QoL mk3), higher levels in the Learning attribute meant a faster rate of improving other Attributes.  Since returning to this game last year after a hiatus, I have struggled to verify these claims.  
    An additional advantage was claimed by the community when the Job Board was originally introduced.  High Learning meant dupes could advance their skill mastery more quickly, a function that could stack with any Interests that dupe may have possessed.  This time I *was* in fact able to find a positive correlation on this claim.  My findings were reinforced (and guided, tbh) by the game's mouse-over tool tip that Learning increases not only research speed but also "+10% skill leveling" per level of the Learning attribute.  

Fast forward to April 16 2019, the release of the "Quality of Life upgrade mk 3".  The Job Board has been replaced by the Skills Board.  The former system of mastering skills over time of wearing that skill's hat has been replaced by an independent system of experience points (XP) gain to earn Skill Points (SP) which can be spent toward instantaneous mastery of chosen skills.  I will soon prove to you here that Learning has no bearing on these XP & SP gain rates.  In conjunction, my proof should indicate that the "+10% skill leveling" per level of the Learning attribute no longer exists.  Lastly, I hope to illustrate that other attributes are improved independently of how much Learning a dupe possesses.  

I started a new colony when this QoL mk3 update was released.  The colony is now 189 cycles old and has 12 dupes.  Dupe #7 (by chronology of hire) is my dedicated researcher.  His Learning attribute is level 18.  For primary comparison, my dupe #3 (ie. one of the original three from Day 1) has a Learning attribute level of 0 (zero).  Secondary comparisons may be drawn as follows:
Dupe #1: Learning = 7
Dupe #2: Learning = 7
Dupe #3: Learning = 0 (see this paragraph above)
Dupe #4: Learning = 4
Dupe #5: Learning = 9
Dupe #6: Learning = 0
Dupe #7: Learning = 18 (see this paragraph above)

First I will make the easy proof: that the new Skills Board and its XP & SP systems are utterly unrelated to the Learning attribute.  Here are how many Skill Points (total, both spent and unspent) the same above seven dupes have earned since the Skills Board was erected in this colony somewhere around cycle 10:
Dupe #1: 11 total Skill Points
Dupe #2: 11
Dupe #3: 11
Dupe #4: 11
Dupe #5: 12
Dupe #6: 12
Dupe #7: 11
That is not a typo.  Dupes # 5 and 6 have earned enough experience points (XP) to be in possession of 12 total skill points, one more than their predecessors and one more than my dedicated research dupe, the one whose Learning attribute is at level 18.  
    Now, to be fair, the other five less experienced dupes are all within 3,000 XP (about 23% or less) short of earning their next Skill Point.  But I find it fascinating that two dupes, hired sometime between my original three and Dupe #7, one of them having Learning level 9 and the other Learning level 0, could be one Skill Point ahead of the other five.  This without a doubt proves that Learning attribute level has no bearing on rate of XP gain and in turn rate of SP gain.  This means the mouse-over info tip of "+10% skill leveling" no longer applies to any function in the current version of the game.  It is obsolete.  

Next is the slightly trickier proof: that levels in the Learning attribute do not affect the rate of improvement of other Attributes, as people have claimed in past years.  I can happily attest that Athletics and Tinkering still enjoy a partial comorbidity thanks to long term use of the Manual Generators, a machine I am quite fond of: my new colony still has 19 of them in operation on different circuits.  (Don't judge.)  
    Using the same seven dupes, here are their Athletics and Tinkering attribute levels:
Dupe #1: Ath = 13, Tin = 11
Dupe #2: Ath = 14, Tin = 15
Dupe #3: Ath = 13, Tin = 10
Dupe #4: Ath = 12, Tin = 10
Dupe #5: Ath = 15, Tin = 16
Dupe #6: Ath = 10, Tin = 12
Dupe #7: Ath = 15, Tin = 13
    Out of these seven dupes, only two of them have mastery in skills under the "Operate" interest: Dupes # 5 and 6.  Both of them have mastery in the skills "Improved Tinkering I" and "Electrical Engineering", which together grant a free +4 to the Tinkering attribute level for those two dupes.  The other five dupes earned their Tinkering attribute level strictly from operating machinery in/around the base.  I would like to add that Dupe #2 is my dedicated chef and therefore spends much of his day at the Microbe Musher and the Grill, both of which contribute toward Tinkering advancement.  

By comparing the seven dupes' Learning, Athletics, and Tinkering attribute levels, I think it becomes clear that levels in the Learning attribute do not affect the rate of improvement of other attributes.  The vast differences of Learning attribute levels is in no way reflected by the relatively even playing field of the Athletics + Tinkering attribute levels.  Further, I find no other correlations with other attributes to suggest this proof is false.  

So what does all of this mean?  By my assessment, it means the Learning attribute has completely lost its meaning as the number one, go-to, must-have, indispensable, quintessential, never-fail attribute that vocal players in the past have focused on and sworn by when making decisions about which dupe to hire when a new Printing Pod batch arrives.  So far as I can understand it now, based on the above findings, the *ONLY* thing the Learning attribute now affects is rate of research at the associated research stations.  (It may also influence the speed at which geyser analyses are performed but I've had insufficient opportunities in this new colony to make a confident attestation in that regard.)  
    I don't know about other players but, as for myself, my rate of advancement through the Research node tree has always eventually outpaced my base's capacity to integrate the newly discovered technologies.  Thus, I see no reason to predicate hiring decisions based on the Learning attribute at all anymore.  Feel free to hire stupid dupes.  

Please share and discuss if you have observed different results.  Thank you.  

 

[johnerson]

I wish there was more clarification on this info. I've seen no difference between any leveling whether skill points or attribute increases. They all seem the same. This is with two dupes running on wheels all day with drastically different learning attributes. Tinkering doesn't go up faster. Neither does skill point experience. Nothing.

Additionally I too have heard and seen many people saying in QoL mk3 that it affects skill points increases, and other have said attribute gain rate. People have claimed to test it, yet I see literally no difference under any circumstances. Before QoL mk3 learning was definitely a major factor.

[Gurgel]

Interesting. I had not noticed either way. But normal play is too varied to really be able to tell reliably. It requires a controlled experiment. Probably two dupes, one with high learning and one with low learning and them doing nothing except the hamster-wheel until both have gotten a skill increase. Or something like that.

While it is certainly possible to set up something like that in debug (large water tank, large petroleum tank for energy, electrolysers, large store of non-perishable food), and several dupes in their own "bubbles" with nothing to do except run on the wheel, I don't think I can be bothered at the moment. 

 

[Yunru]

Easiest way to test is to open up the save editor and max out one dupe's learning.

[JohnFrancis]

This has been brought up a few times recently, learning still works when it comes to using the hamster wheel at least.

Two dupes, "Smart" +5 learning, "Dumb" 0 learning. No points in athletics, tinkering or any interests in them. Locked in small room with nothing to do but run on a wheel. Smart hit Tink 4 by cycle 6.48, Dumb hit Tink 4 by cycle 9.24. Put another way Smart got to level 5 tinkering just before dumb got to level 4 tinkering. Start and end save game files attached.

Notes, bottomless stomach increase the time it takes for a dupe to finish a meal. However -3 strength slows down a dupe cleaning a toilet so it all evened out. Athletics level up a fair bit slower than tinkering on the wheel.

Learning test End.sav

Learning test.sav

[0xFADE]

Good to know it still works.  Maxing research tree is still the best then.

[mahesar]

22 hours ago, JohnFrancis said:

This has been brought up a few times recently, learning still works when it comes to using the hamster wheel at least.

Two dupes, "Smart" +5 learning, "Dumb" 0 learning. No points in athletics, tinkering or any interests in them. Locked in small room with nothing to do but run on a wheel. Smart hit Tink 4 by cycle 6.48, Dumb hit Tink 4 by cycle 9.24. Put another way Smart got to level 5 tinkering just before dumb got to level 4 tinkering. Start and end save game files attached.

Notes, bottomless stomach increase the time it takes for a dupe to finish a meal. However -3 strength slows down a dupe cleaning a toilet so it all evened out. Athletics level up a fair bit slower than tinkering on the wheel.

Learning test End.sav

Learning test.sav

This matches my recent observations as well. It seems that learning still helps attribute gain, but no longer has an effect on xp gain.

[Craigjw]

All dupes have an inherent -50 to intelligence and have the dumb as a stump trait.  I would have thought that this is obvious from their behaviour.

[Inter]

On 4/27/2019 at 6:07 PM, Yunru said:

Easiest way to test is to open up the save editor and max out one dupe's learning.

Any advice on how to do that?  I have not done the things (nor care to) that allow true sandbox mode (ie. developer debug sandbox with super speed).  Is the save editor something that can be accessed without enabling true sandbox?

[Yunru]

24 minutes ago, Inter said:

Any advice on how to do that?  I have not done the things (nor care to) that allow true sandbox mode (ie. developer debug sandbox with super speed).  Is the save editor something that can be accessed without enabling true sandbox?

It's an external program: https://robophred.github.io/oni-duplicity/#/