Stat growth rates from levelling up - collecting data? calculator + sheet

What's this

As we know, wolves gain stats when levelling up. I previously read on the forums that this was random and assumed as much/possibly influenced by their role, but then tried adding up the stats my wolves gained on levelling up, and noticed:

• All the pups (litter of 6) from breeding to a stud with high strength and smarts ended up gaining more strength and smarts than other stats, even though each one filled a different hunter role, and the mother had high speed/agility
• Similar (litter of 5), but with a different breeding to a stud with high wisdom and smarts, so that's another data point
• Random stat gain should result in fairly averaged-out stat growth but all of the wolves I checked had high gains in some stats and low in others

All of these are somewhat statistically unlikely if stat growths were completely random. This suggests:

1. Stat growth from levelling up is not entirely random
2. Stat growth proportions (and not just the birth stats) are hereditary

I think it might be something interesting to look into! (At least, I don't think I've seen anyone post about this before?) I don't really intend to research this hardcore, but it would be nice to have some confirmation for breeding purposes.

What are we trying to find out

• Well, first of all is this actually true? Do stat growth proportions pass down from parents to children? (Which is where data collection comes in)
• How is this inherited? Does it mix it up between the two parents, possibly with minor randomness, or pick one parent's build to pass down?

Tools

I tried manually typing in data from the user log, but this is an extremely painful way of collecting data and I strongly unrecommend it. Instead, I've coded a calculator/formatter and made a spreadsheet for recording this stuff:

Click here for the calculator

Click here for the spreadsheet template

Here are some visual instructions for using both of these things

Feel free to use them for personal purposes as a tracker or something. I say "data collection", but really I just spent a couple of hours coding this out of frustration and didn't want the effort to go to waste haha.

hi! so i've actually been tracking the stats gained from level ups only for my pack for some time now, and have an entire spreadsheet for it!

link to spreadsheet

what i've gathered is that there are two stat categories (ie. strength, agility, speed, wisdom, smarts) that each wolf favors individually, regardless of pack role and other external factors. and every time a wolf levels up, there is a base percentage that each stat category has that determines how many stats it receives upon clicking the "level up" button, with some give-or-take thanks to rng.

so in my spreadsheet, i have raw data gathered in columns c through g for each wolf, which is colored in to show how heavily each stat was favored in that particular level up. then, in columns h through l, i have a formula going that calculates what percentage of the stats distributed went toward each stat category during that particular level up instance.

what i've found is that in general, each wolf favors two stat categories that have a base percentage of about 30% each in stat distribution upon level up instance, and the other three stats have about 13.33% each in stat distribution upon level up instance. then there are rng modifiers that slightly alter the otherwise expected level up stat gain results, which is why sometimes a stat category that isn't favored may gain more stats than a stat category that is favored. however, this randomness starts to stabilize and become more consistent with a wolf's favored stat categories after it hits level 10 (from what i've seen).

as for breeding, my theory is that these favored stat categories get passed down to pups (rather than being chosen randomly upon wolf creation) regardless of what starting stats a pup was born with. so, for example, if you bred wolf 1 with favored stat categories strength + agility to wolf 2 with favored stat categories strength and speed, most likely their pups will be guaranteed to favor strength, and either agility or speed, regardless of if wolf 1 and/or wolf 2 had really high wisdom that was manually raised through scouting, hunting, herbalism, etc.

i'm testing my breeding theory at the moment, since i now have two litters of pups that have become adults and are currently being leveled through hunting. unfortunately this is kind of a slow process, since you'll never be too sure of what exactly a pup's favored stats are until you've leveled them up until about level 10, and in certain cases you may even have to go a little beyond that to be really sure. that said, my stat breeding theory has held up so far with the two pups that i have managed to level beyond level 10, and i'm curious to see if there's some sort of randomness component in which pups can randomly choose to favor a stat category that isn't favored by either parent (kind of like how personality inheritance works).

hope this is helpful!

This is incredibly interesting. I hadn't given it much thought, just assumed that it either favoured the wolfs role or was totally random. Certainly doesn't hurt to investigate it though!

I'll go through my log and collect the data for my current Lv20 hunters - this data only goes back to Lv13, but perhaps something can be gained from it. In addition to that, I'm currently training up new hunters which are directly descended from my older ones. So I can track their data from Lv1 and draw up some comparisons when they start hitting higher levels. I won't have data for the fathers in most cases, but I can consider their stat totals and role.

It's interesting to consider my Lv20 hunter's stats here, too - 4 of them were customizer wolves (so they started with rounded stats) and for the most part their stats which aren't relevant to their role round out pretty evenly, and in places where they don't it can be explained by gains from short-term scouting. The only exception to this is my finisher, who inexplicably has significantly higher smarts that I'd expect. Will have to track their levelling data before I can elaborate any further on this though.

Here's my spreadsheet link for reference

@woollydaisie
Wow that's detailed, thank you!!

That tracks with what I've observed, including the two favoured stats thing! I wasn't entirely sure if it really was that there would always be two top stats, since it seems to be the case in NBWs I've levelled, but among my gen 2s there are a couple who currently have only one prominently higher stat at level 13...maybe they're late bloomers?

Have you noticed if it's more likely that a particular parent is more likely to pass down their favoured stats? The two cases I mentioned in my first post passed down the father's favoured stats, and consistently throughout the entire litter, so I wonder if this is random or it's possible to tell which parent will pass it down. (e.g. always the father, always the one with higher stats, etc. of course though it could still be random but applied to the entire litter) Definitely interested to find out if there's the randomness component too, since that might...complicate things...but so far it seems kind of unlikely, or at least a somewhat rare occurrence.

My guess is: since the amount of stats a wolf can gain is already fixed at 420, I think it's possible that the max stat growth amounts are generated when the pup is created, and simply pulled from randomly as the pup levels up, so there's no chance of the growth going in unexpected directions. The max stat growth proportions COULD be inheritable...unfortunately it already takes a very long time to hit level 20,  so getting enough data to make more precise predictions will probably take months. Will keep gathering data though!

@Srinyx

Looking forward to seeing what you find! Same, my older wolves' data has mostly disappeared so it's kind of patchy. Though at least, the numbers at higher levels are more representative of the growth rates, so it should still tell you what the wolf's good at!

@Srinyx

i used to think that it was role-based too! but then one of my chasers suddenly had better stats than my finisher, and i realized that even though my finisher had better strength in the beginning, after a while it was actually far from his strongest stat xD but it'd definitely be interesting to see what you find!

@aleator

interesting! i've definitely had wolves where it was clear that they favored one stat, but the second stat didn't become clearer until between the levels of 10 and 15, so it's possible they could be late bloomers and rng just hasn't favored their second favored stat xD but since they're gen 2, there's definitely the possibility that maybe more genetics is at work behind the scenes.

which actually reminds me, i had two theories going into tracking my wolves' stats and wanted to see which one followed better, which is why i started tracking their level up stats in the first place:

theory a: there are "categories" of favored stats with set percentages of stat gain upon level up, which is made slightly random by a secondary hidden percentage / number / rounding. the first category would be the favored stats, of which there are two, the second category would be average stats, of which there are either two or three, and, in the case that there are only two average stats, then there's possibly a third category of weak stats, of which there is only one. my original thought for this was that the two favored stats have a base percentage of 30%, the two average stats have a base percentage of 15%, and the weak stat has a base percentage of 10% (or, alternatively, the two favored stats have a base percentage of 30% and the three average stats have a base percentage of 13.33%). in this case, pups would inherit stat categories in a similar way to which they inherit markings from their parents.

theory b: wolf stats have set stat percentages (instead of categories) upon creation, with gen 1 stats being relatively close to to 30% / 30% / 15% / 15% / 10% or 30% / 30% / 13.33% / 13.33% / 13.33%. then, when bred, gen 2 puppies get stats with averaged percentages from their parents (with some level of randomness) in the same way that wolf marking opacities get inherited.

at the moment, i'm leaning more toward theory a because i have two gen 2 wolves that were bred from a speed + smarts gen 1 with a wisdom + smarts gen 1, and while they both inherited smarts as a favored stat, both also favor speed, while wisdom is one of their weakest stats, which wouldn't follow with theory b if that stat was averaged between the parents. (they've actually both shown more growth in strength as a non-favored stat, which is weirdly the weakest stat for both their parents, also leading me to believe that there are three average stats and not two average stats and a weak stat, although again, it could just be a coincidence of chance).

i do find the theories worth mentioning though in case there's some mixture of both that hasn't been observed yet or it sparks an idea about a different way that stat genetics could work.

and, to answer your question, in the case of the two gen 2 wolves, they did both inherit their two favored stats from one parent, but then again, one of those stats was a shared favored stat between their both parents, so i'm still guessing it's probably roughly a 50/50 chance between the two parents.

and your guess is definitely something i haven't though of before! but the only thing that makes me skeptical of it is that storing an array of stats for every single wolf on site might be a lot, considering how much concern has been thrown around in the past about wolvden's servers and why certain mechanics are the way they are because of it xD i'll be the first to admit that i don't know too much about web servers in general, but if they were deciding to level up stats that way, it wouldn't track with the paywall of the scrying stone, the complete wolf deletion of chased and un-dynastied wolves, and seemingly reused image links for deleted wolves. in my opinion, and just from like, a coding and game dev style standpoint, it seems more likely that there's a single instance of stat generation + rng that happens when a level up button is clicked, since it requires less code and less stored data, and it follows with wolvden's general preference for putting an rng on most of their mechanics

and speaking of coding, has anyone else noticed the seemingly random order in which stats gains are reported upon level up?  (you can see it in the banner after clicking the level up button and in your user log) although it might not be anything, something makes me think that the order might be a little window into how the game decides to distribute stat points, although how exactly i'm not entirely sure xD i just think it'd be weird for the devs to make the order random for no reason, so there must be some reason, maybe :o

@woollydaisie

Your theories are interesting! I agree that the averaging is unlikely, since that would mean that after several generations wolves wouldn't display the 2 favoured stats phenomenon and tend towards having very average stats. I think something like theory a would probably be more likely, with some percentage range for strong stats and average stats, I guess? Given the 420 stat limit they probably would first roll for which ones are strong/average and then throw in a bit of randomness. 

re: having three average stats vs two average stats - I have a level 20 NBW that has one weak stat, and another gen 3 that doesn't really, so can't really say yet...but my gen 2 hunting team seems to show 3 average stats, so inclined to believe it's just that the range for the last 3 stats is probably quite large.

On storing stats - thinking about it from a coding standpoint (have experience here), it would only be an extra 5 integers at minimum (or an array stored as a string) so I don't think it's that much! The scrying stone paywall is due to resource consumption from running the breeding calculation every time and not storage (as mentioned before by staff), the entirety of wolf data is very much more than part of their stats, and images take up a lot more space than database data so it makes sense that those are deleted entirely.

The stats per level up don't have to be stored, for example: a wolf could have a potential stat gain of 132 strength, 78 speed, 51 agility, 72 wisdom, 85 smarts through levelling up, and this is rolled initially. Only those 5 values are stored, and for each level the stats are randomly rolled from this potential stat gains, e.g. for level 2 the wolf needs 4 stats, and this rolls as +2 strength, +2 smarts. This is added to their actual stats, and their remaining potential stat gains are reduced to 130 / 78 / 51 / 72 / 83. This goes on until level 20, where it hits 0 / 0 / 0 / 0 / 0 - it'd be simple, efficient and not prone to RNG mishaps...which are probably somewhat likely to happen, since 19 level ups is a very small sample size.

The stat gains being in a random order is probably for efficiency of randomisation. I think it goes like: take the array of stats, randomise its order, and then traverse the array in the new order, rolling stats as you go. This is efficient because 1. this would only require rolling 4 times at maximum, 2. ensures that wolves at a low level (too low to gain 5 different stats at once) can gain stats that are not just the first few stats on the list (strength/speed/agility). The stat gains are displayed in the order the array was randomised. It proooooobably doesn't really mean anything :>

@aleator

ahhh ok got it, i'm glad someone with a better understanding of code could explain that out for me xD my original interpretation of what you meant was having an array of like, 100 integers stored per wolf and each time a wolf leveled up, those integers would randomly be pulled out and applied to the wolf's current stats, but the way you just described it makes 1000x more sense lmao

(edit: also, thank you for explaining the reasoning behind scrying stone resource consumption vs. data storage, i knew those things were different but didn't know how much of an impact each had on the server in comparison to each other, but i'm assuming resource consumption strains it more than data does, right? and yes, deleting images definitely makes sense! but i assumed all stored wolf data beyond images also got deleted too, beyond the bare essentials that would still be relevant to living wolves, like lineage data, which is why i was thinking that the devs were probably trying to minimize as much stored data as possible for limited storage reasons)

in that case, i do think your theory is pretty likely as well! i also figured that the random order in which stat gains were presented upon level up had ~something~ to do with how the stats were being rolled, but for the life of me i couldn't even make a guess as to how or why. so your explanation definitely makes sense, and works pretty well with your theory too!

ok, so i will definitely keep that in mind as i level my gen 2 wolves! they're hunters so hopefully i'll even manage to get them to level 20, but if their final stat distributions end up being super similar (or even the same :o) then it would definitely support your theory pretty well!

and also yes, i have nbws that lean way more to having a "weak stat" and nbws that have more averaged non-favored stats, which is why i'm not entirely sure if it's just three average stats with varying stat gains thanks to rng or if there really is a weak stat.

I sort of fell into breeding specialists and plan to keep it up, especially after watching them develop. I have three main tracks which fell out of what the NBW parents trained in. 

Selene 

Once my biomes were fully scouted, my starter female continued to scout the rainforest until she reached level 20. Her stats are level 20 are: 217/157/126/103/112. I bred her to my lead male a couple times, who seems to favor speed/agility, but had lower stats than her at the time. I've kept three so far (I will probably use her last heat for my lead's last litter). One is from my first litter (Cesium), so my user log doesn't go far back enough for all her leveling and she's only level 8 since I use her as a pup sitter. 

The second is a chaser (LC). At level 15, her % gains have been: 15/27/25/18/16

The third is a rainforest scout (Pandeia). At level 19, her % gains have been: 32/29/15/11/12. She hunted for her first 5 levels (chaser I think) 

I also have three young adults from her and heritage-less male with high strength. They all have higher strength (and speed) stats, of course. Two haven't been trained, and I'd be willing to give one to someone who wants to experiment. The third has been training as a chaser until I move her to scouting. 

All have had puppies of course, but most are at the lower level. LC has a daughter at level 14 who is a chaser. Her stat gain % is 32/26/15/13/14. Her sire had higher strength/speed relevant to the rest of his stats (scout). 

Raida

Raida, an older NBW, started as a stalker and became a scout so I could open up the glacier. Her stats are level 18 were 100/100/94/225/172. She has one daughter at level 15 that I don’t own. Her highest stats are wisdom/smarts and she was a scout, but don’t know where she mostly worked. 

I have two of hers from the same litter that I kept, from a sire with higher relative wisdom/smarts.

1. Raika (glacier scout). Stat % gains at level 20: 13/16/13/27/31

Silicon is unbred (can’t decide if I want to sex change him). Raika has had a couple litters, but none above level 7. Her oldest, Ricia, was specifically bred from a relatively higher wisdom/smarts sire. She has been leveling through stalking and some scouting. Her % gains at level 7 are 11/15/9/37/28. 

Raika’s newest litter is sired by one of the highest wisdom studs I could find. Almost all were born with 130+ wisdom and higher relative strength. I plan to keep one or two to continue my wisdom line, but I would be interested in seeing one of them be specifically trained as either a chaser or finisher (by me or someone else). 

Others

Vanadis — breed from one of my NBW chasers and a sire with higher strength/speed/agility, she is a chaser. Stat % gains at level 15 are 15/31/29/13/13

Younger wolves that I’m training are usually from specialized mothers and sires chosen to reflect their best stats. I’m training a finisher who is by a high strength sire out of a finisher NBW. A couple pups from Pandeia (mentioned above, high strength/low wisdom), and another pup from a high strength sire out of a finisher mother (not NBW). That pup’s strength stat is almost twice as much as all her others. I have a couple young pups from a finisher mother and high strength male that I will be raising up as well. 

I’m happy to contribute as much as possible and I’d love to collaborate with other people breeding specialists, especially those avoiding more common lines. :) My goal is full set of 500+ stat hunters so I can move to the glacier without worrying as much about food. Looks like I need more lines for chasers! Haha 

This has been a huge revelation for my sister and I! We figured it out randomly when we were looking at stats on some studs, and started trying to put a spreadsheet together, but it took me almost an hour to just do one wolf. The cruncher has been an absolute life saver!!! It has added so much more depth to our stat breeding projects and I am hoping it will push us towards breeding some more uncommon studs.

We're still compiling all our data and trying to figure out how it carries down through the lines. I'll make us a nice spreadsheet with all our data compiled to share once I get some time to put it together. I do have all my suspected boosts for my wolves listed on my cave page though. We have found that you can usually tell pretty clearly by level 6, but some might need a little more time to be clear.

Has anyone found any that only have one stat boosted? We have one that is not showing an obvious gain in a second stat, but it could just be weird RNG because I don't think she is too high leveled yet.

If both parents have one of the same stats boosted are the pups definitely going to have that one or could they inherit both of the other boosts? For example if you breed a (strength smarts) with (strength wisdom) will it always have strength or could it inherit smarts and wisdom instead? Curious to see how some of our planned litters turn out!

According to about my whole cave, Sunfall definitely passed down a lot of strength boosts! Strength seems to be by far the most common as we are going through and trying to estimate the boosts on studs to use. 

Currently the 2 highest stats of a wolf continue to be the favored growing stats. By level 20, a wolf will see around +60 in their 3 unfavored stats, and around +120 in their 2 favored stats. Stat gains aren't related to the role they're maintaining, just whatever the 2 highest stats they have are. That's why breeding for specific stats is preferred for specific hunting roles :D

From bugbox:

Comment by Abbey posted 2021-02-05 22:15:57 The stat gain is based on the stats they have - it will try to prioritise whatever the highest two stats the wolf currently has. It is not based on role.