How FS Can Scientifically (Objectively) Balance All Weapons

Warhammer 40,000: Darktide General Discussion
1

I’m new to these forums, didn’t mean to delete my previous (duplicate) post. So step by step:

1. Calculate all players’ WRV (Win Rate Variable) based on percentage wins with all weapons in the current patch. This is the control variable that all other data will be balanced to.

Note: WRV doesn’t matter on its own, we’re only interested in how it changes with different weapons. Data can also be restricted to certain difficulties/game modes if desired for targeted balance.

2. Calculate the difference from this WRV for each specific weapon that player has used (different Mk versions also count as different weapons for balance purposes). For example, let’s say player #2313 wins 6% less often when using the ‘Agripinaa Mk VII Combat Shotgun.’ This means their WRVD (Win Rate Variable Difference) for that shotgun is -6%.

3. Take step 1 and 2 for all players and aggregate the data for each weapon. For example, let’s say across all players, the ‘Agripinaa Mk VII Combat Shotgun’ has a WRVD of -5.4%. All players across all games win 5.4% less often when they take this weapon compared to their average. Objectively, that would mean this weapon is underperforming and needs buffs.

4. Buff(/nerf) positive/negative weapon variables as needed, ideally changing the variables to make each weapon more unique from the others. Repeat process every month or so for all weapons that have over, say, +/-2% WRVD from the average of all weapons.

Result: your favorite non-meta weapons are now viable without dragging down your team and generally everyone has a lot more weapon options to enjoy, rather than being stuck using just a small selection.

For a hypothetical extra lazy developer, every step of this process can be fully automated and rerun monthly or at any interval. Even step 4 with preset increments/decrements. Not that it should be, but just saying. For a hypothetical lazy developer.

If people are upset about nerfs to (objectively) overperforming weapons, get good? If people are really upset about nerfs, the process could only be used for buffs on (objectively) underperforming weapons to bring them to a baseline range… Anyway, just a thought because the weapons are not well balanced in this game, at all.

1 Like
2

Based

10 Likes
3

i’m not opposed to nerfs, not in the slightest.

problem with the win rate would be the enormous amount of people getting carried, no matter how abysmal their performance/loadout is.

in a pvp environment this might give a realistic approach cause everyone wants you dead and its your job to work against that.

here, even if its not my job, to see the match done i need to drag people along for better or worse.

if i’d take a summary of over 1000 matches i uploaded around 100 players were okay-ish and less than 20 on eye level…

almost each match (some cool fails i uploaded as well) was a “win”

8 Likes
4

immediately no. this would get shotguns nerfed and dueling swords buffed. this game is littered with very competent players running suboptimal builds and incompetent players running hyper optimized builds

15 Likes
5

That is a really good point

2 Likes
6

As noted in the OP, the win rate is not relevant by itself. We’re looking at the difference-in-win-rate across different weapons, averaged for all the games. Unless they are literally afk, there’s a difference between carrying a player who has decent weapons and carrying that same player who’s weapons cannot do anything. So it does not affect anything if, on the times they do win, it is because they are usually being ‘carried’ by more skilled players (with higher absolute WR). Only the difference between their different weapons is relevant.

Though again, this can be restricted in any way to eliminate data from easier difficulties and/or high-loss rate players, IF it is somehow distorting anything. Personally I would advocate excluding anything below Damnation since breakpoints are different so the weapons will work differently.

7

Purely statistical balancing will always be flawed just because of human nature. For Darktide’s case, weapon balancing in this way would firstly not account for players’ talent trees (an exe stance vet with dueling sword is going to be more vulnerable than a heavy sword VoC vet), won’t accurately represent weapons that aren’t picked often, and lastly as mentioned before it doesn’t account for player skill.

Additionally, pure statistical balancing will never be full proof because it doesn’t actual inform anyone of why a weapon is under preforming. Say for example they see that the Infantry Lasgun isn’t preforming well compared to other weapons. They might just bump the base damage, which while it might help doesn’t address the fact that it’s a super jumpy gun, has weird inverse damage fall off, the las pistol already does it’s job much better AND provides insane mobility, and in the current elite and boss dps rush meta it will always underpreform.

3 Likes
8

Human nature actually helps the inform the stats. But you do have a point about different classes. So IF there was a statistical outlier that was overperforming for one specific class, that would definitely be a special consideration.

Though in this case, human nature would already mean more players are playing that weapon with that class (and winning more) which would be reflected in the stats. So definitely if some individual weapons were extremely powerful on one spec, that would be a special balance consideration. This could only objectively be determined through stats.

9

Talents and blessings are a fine tuning over time.

I’d say first look at the gap against armour with certain weapons when specced into certain talents that enhance them.

Particular those that just take 3 strikes to down a crusher in few secs of very low TTK. Should be that crusher patrol are a fearsome pack that require focused engagement than someone solo them.

DS - 3 pokes and mobility to “almost” knife level.
Plasma - 3 shots. The only weapon to be pinpoint accurate even when hip firing, no ADS required.
Bully 3b - 3 light hits, by then bleed does the finish 2 secs later.

Then other things like cat 3 knife. Infinite stamina run using knife heavy move. Mobility like this means regardless of difficulty you can bypass to next stage of mission easily without effort.

As any melee weapon can be horde effective with right moveset people will lean towards more single target for worst case scenario of armour pack.

If there was good example of balanced single target - Rashad combat axe mk3. Not too hard, not too easy on crusher.

10

Completely irrelevant to the method I described. The only number being aggregated would be the performance difference in different weapons for the SAME player. Though anecdotally I’ve never seen a player with a shotgun perform well…

11

now thats actually irrelevant.
even assuming fatshark adds a system for getting such stats it would take a massive amount of games to account for the difference that happen in each game
even then not all players play all the weapons i for example play like 2-3 melees per class

12

Spreadsheet balancing is the cardinal sin of game design and frankly I think you should feel ashamed for suggesting it.

Ubisoft tried the win-rate thing with R6 and it has resulted in a constantly changing meta nobody is happy with. Instead of analyzing WHY players are complaining, and what about, they just look at numbers and change random values to try and find the magic ideal balance, which will never exist.

9 Likes
13

For melee I agree with that generally. It does seem like a requirement for “meta” weapons are that they can reasonably deal with carapace. Maybe they could think of more interesting ways to balance that would make weapons more different rather than more same, but it might just be that simple for carapace. Otherwise crushers will just bulldoze you.

Anyway I wasn’t speculating on how they would balance so much as how to determine what needs balance and how to verify the balance was effective.

14

i’d say not by much, given a thousand scoreboards in a huge amount of cases i ended up with more kills/damage than everyone else combined.

bit of a stretch but i’d call myself a rifle guy, decent to good shot with what i could get my hands on so far. decades in between and far from any regular practise.

but i literally suck with pistols, cant make head or tail of it, but unless i magdump into the target, well lets say its ugly. most expensive,best, fancy wont change that. maybe a red dot :rofl:
…

that outta the way back to topic.
weapons are tools, much as talents or your character as a whole.
lets be honest, most people suck at utilising these tools properly, no need to sugarcoat it.

so practical example, lets say i carry a group of 3:

besides picking them up, i dont “care” for their participation (dog/net outta the equation for the time be), i go as fast/efficient as possible and never hold back.
that over the span of 30 minutes results in me netting the vast majority of… well about everything, usually up to collecting the most plasteel/diamantine…

end result everyone else double digit kill while mine’s in the triples.

would there be a difference in the kind of weapon when each 3 are unable to keep up/chime in in the first place ?

not killing with a bolter is as little as not killing with a pistol/flamer/shotgun etc.

its not the tool that did a bad job its the dude behind the keyboard.

since i bring my best kit and the others are slower in reaction time, by the time they act stuffs already dead.

perfect world, we take the 40 best darktide players there are, give them 10 teams and 10 matches with every gun on every map.

THEN i’m on board with the idea.

i rule myself out cause i only play ogryn but i’m sure theres a good amount of jack of all trades that play every character equally and are willing to contribute to the cause :+1:

15

Yea that sounds like Ubisoft. I didn’t suggest changing random values though. As I alluded to, weapon changes should be made with purpose towards a specific theme so you actually feel like you’re using a unique weapon rather than a nearly duplicate reskin.

But my post was more how to measure how their changes are working or what needs to be changed in the first place. Because it’s pretty clear over multiple games FS doesn’t know how to even start balancing weapons.

It’s not even that they balance them in a soulless/random way like Ubisoft but they don’t even try to balance them in the first place… They make a lot of weapons which are nice on paper but no one uses 80% of them because they aren’t good enough to work at a fun difficulty.

1 Like
16

ANY competent developer already has these stats. They have public servers so it would be extreme incompetence if they don’t already have the stats. Do you believe they just ‘balancing’ from vibes?.. Maybe actually; I hope not. And sure, only look at numbers for weapons that have been played for, say, 5+ total games (exclude the first 5 because learning).

For weapons not even being played 5 games, that is already telling them what they need to know…

Okay? As long as you play more than 1 ranged or melee weapon there is comparative data. If you only play 1 weapon then there’s no meaningful information.

17

bruh fatshark cant even get version control right. theres leaks and returning bugs every patch.

i don’t think you understand the amount of data they have to process.

for example i played mostly ogryn this month cycling between pickaxe and shield. from what i remember i only lost one game.
so win rate is almost useless
what else can be used? damage maybe?
ok lets see maybe i changed builds but kept the weapons, maybe in some games the ratio of chaff to elites was different, maybe i got really cracked up teammates, maybe i didn’t.

you can not build a system to balance weapons with the insane amounts of variables in each game, player, patch, etc
fatshark can not do this. if they could they would be working with big data
frankly im not even sure this gone be done with statistics.

18

You should first ask yourself whether you need to get good in games balancing

What you said is not feasible for a multitude of reasons

  1. bad players tend to take meta weapons; pro players look for suboptimal alternative builds;
  2. not all weapons are equally easy to use: does a weapon get fewer wins because it is weak or because it is difficult to use?
  3. not all weapons play the same role. Not all weapons are designed to get you wins, survive to a solo clutch; many weapons are strictly dependent on others;
  4. not all weapons are used equally, this can alter statistics;
  5. victory or defeat is given by an incredibly large number of factors… is the result of a combination of the four weapons, of how these four weapons synergize with each other, from the skill of the players, from their build, from the RNG of the level itself; excluding the first four points, I do not deny that despite all these random factors, it is still possible to get relevant statistics… but the matches analyzed should be a disproportionate number. More than you can imagine. Not only many many resources would be needed… but considered the number of players and so the number of matches, it would be something that would take an enormous amount of time
3 Likes
19

This is immediately an issue as it doesn’t account for player skill, difficulty, teammates, stat rolls, blessings, perk synergies, or Talents.

2 Likes
20

we’d need a “jury” of really proficient players that’d hunker down until everyone agrees based on testing “everything/everyone/everywhere”

i’d say 40 people to make 10 teams and what… half a year on a pizza and energy diet… who’d say no to that :rofl:

1 Like