Login

russian armor

Winrates indicate horrible balance

PAGES (8)down
14 Jun 2021, 08:11 AM
#61
avatar of Katitof

Posts: 17914 | Subs: 8

jump backJump back to quoted post14 Jun 2021, 08:07 AMVipper

The theory that there is a "massive" difference in fuel between 1vs1 and 4vs4 is simply not supported by numbers.

Where are these mythical numbers you speak of?
14 Jun 2021, 08:17 AM
#62
avatar of Vipper

Posts: 13496 | Subs: 1

jump backJump back to quoted post14 Jun 2021, 08:11 AMKatitof

Where are these mythical numbers you speak of?

Right here:
https://www.coh2.org/topic/61229/resource-income-per-game-mode/page/1#post_id609470

and there not "mythical", they are quite reliable and tell a different story than your theory.



"That said, the difference between 1v1 and 4v4 is smaller than I expected (around 200 fuel after 60 minutes). Also, it is noteworthy that before about 14 minutes the fuel incomes is slightly lower in the larger game modes. "
14 Jun 2021, 08:47 AM
#63
avatar of JohnSmith

Posts: 1273

jump backJump back to quoted post14 Jun 2021, 08:17 AMVipper

Right here:
https://www.coh2.org/topic/61229/resource-income-per-game-mode/page/1#post_id609470

and there not "mythical", they are quite reliable and tell a different story than your theory.



"That said, the difference between 1v1 and 4v4 is smaller than I expected (around 200 fuel after 60 minutes). Also, it is noteworthy that before about 14 minutes the fuel incomes is slightly lower in the larger game modes. "


As usual, you are misreading the stats, obviously not understanding it, taking a sentence out of the whole thread, and you are not connecting the data to the game. The data shows the average overall for ALL players for a whole match for each player for each game mode. In 4on4, even what OP describes as "slightly lower" fuel results in an N to the power of (player count) impact on the game-play due to the faster income. It is a massive difference in the end. This is easily noticeable by comparing the gameplay of 1on1 and 4on4; they have barely anything in common but the first few minutes. So even if the difference is in terms of numbers quite low, it is actually a big difference in the game flow. There is no basis for what consists of what you describe as "slightly lower"; it is just an observational description, not a fact related to the game (and you take one sentence out of a whole thread) that it is. A unit that arrives on the field even just a few seconds earlier changes the whole game. Make that times 8 (for all players) and you got a whole different game with every timing gone out of the window. It's great data, but it does suddenly said data and a few sentences becoming the ultimate fact that cannot be debated. There are loads of other things that don't make sense in the data if you actually read it.

There are still massive gameplay changes between the game modes. So what Katitof says is correct, and you're wrong again.

jump backJump back to quoted post14 Jun 2021, 08:00 AMKatitof

Yes.
Map dynamics/map control change completely mode to mode, influencing timings massively.
In 1s and 2s you have map constantly changing hands, making tech and units arrive later, in 3s and 4s its pretty static once you claim it and cache spam doesn't help either.

And units are balanced with assumption that player can use them to their full extend, that clearly doesn't help players who can't even spell units name properly.
14 Jun 2021, 08:47 AM
#64
avatar of redfox

Posts: 92

Is it technically possible for the balance modding team to limit certain units depending on the game mode?
14 Jun 2021, 09:21 AM
#65
avatar of Hannibal
Senior Moderator Badge

Posts: 3113 | Subs: 2

jump backJump back to quoted post14 Jun 2021, 08:47 AMredfox
Is it technically possible for the balance modding team to limit certain units depending on the game mode?

Not as far as I know.

jump backJump back to quoted post14 Jun 2021, 08:17 AMVipper

Right here:
https://www.coh2.org/topic/61229/resource-income-per-game-mode/page/1#post_id609470

and there not "mythical", they are quite reliable and tell a different story than your theory.



"That said, the difference between 1v1 and 4v4 is smaller than I expected (around 200 fuel after 60 minutes). Also, it is noteworthy that before about 14 minutes the fuel incomes is slightly lower in the larger game modes. "

Nothing against SiphonX, but the data shows weird oddities that do not make sense. You cannot use it as an argument.
MP earnings of below 100 for 3v3 and 4v4 within the first 3 minutes? 1v1 and 2v2 generate MUCH more FU and MUN income in the first three minutes, while 4v4 stays at the base income and 3v3 slightly elevated above the base income? This would mean that in larger game modes, there would be next to no territory captured in the beginning. No mode generates a CP until after minute 6? This makes me believe that the binning logic he uses has some inconsistencies.
Additionally the cumulative graphs say the FU/MUN income in small modes were higher early on (which does not make sense as stated above), while the per minute income in large modes is higher from the second bin onwards.

This data has clearly issues.
14 Jun 2021, 09:45 AM
#66
avatar of Vipper

Posts: 13496 | Subs: 1


Not as far as I know.


Nothing against SiphonX, but the data shows weird oddities that do not make sense. You cannot use it as an argument.
MP earnings of below 100 for 3v3 and 4v4 within the first 3 minutes? 1v1 and 2v2 generate MUCH more FU and MUN income in the first three minutes, while 4v4 stays at the base income and 3v3 slightly elevated above the base income? This would mean that in larger game modes, there would be next to no territory captured in the beginning. No mode generates a CP until after minute 6? This makes me believe that the binning logic he uses has some inconsistencies.
Additionally the cumulative graphs say the FU/MUN income in small modes were higher early on (which does not make sense as stated above), while the per minute income in large modes is higher from the second bin onwards.

This data has clearly issues.

You can argue the stat with SyphonX, they might or might not have issues.

Point here is the according to stats there not a "massive" fuel difference between 1vs1 and 4vs4 as claimed by a certain user especially early in the game.
14 Jun 2021, 09:48 AM
#67
avatar of JohnSmith

Posts: 1273

jump backJump back to quoted post14 Jun 2021, 09:45 AMVipper

You can argue the stat with SyphonX, they might or might not have issues.

Point here is the according to stats there not a "massive" fuel difference between 1vs1 and 4vs4 as claimed by a certain user especially early in the game.


Do not deflect. That's absolute garbage. If you are bringing someone else's data to the table as part of an argument, you should be ready to support it, and not drop it the moment someone else finds holes in your debate. Also, you're once again ignoring everything else that has just been said and trying to push across that you're right but you still being again completely wrong as usual. Just accept that Kat is right, there are massive differences between 1on1 and 4on4, and you are wrong and that even your argument makes no sense.
14 Jun 2021, 09:50 AM
#68
avatar of Hannibal
Senior Moderator Badge

Posts: 3113 | Subs: 2

jump backJump back to quoted post14 Jun 2021, 09:45 AMVipper

You can argue the stat with SyphonX, they might or might not have issues.

Point here is the according to stats there not a "massive" fuel difference between 1vs1 and 4vs4 as claimed by a certain user especially early in the game.

What?
Absolute nonsense. If there is reason to believe that the data might be wrong, you can't say "don't argue with me" and still try to push your statement.

This data probably has severe issues. You cannot use it as a backup for your claim. It is not about who to argue with, it is about drawing conclusion based on proper data. Period.
14 Jun 2021, 10:00 AM
#69
avatar of Sander93

Posts: 3166 | Subs: 6

This data probably has severe issues. You cannot use it as a backup for your claim. It is not about who to argue with, it is about drawing conclusion based on proper data. Period.

Even though this went off topic, ultimately this conclusion leads us back to what is at hand.


Ahh...so I use all all games then people complain Allies players are simply noobs and don't know how to play and if I choose Top 200 the sample size is too small.

I guess that proves the game is totally balanced then. How could I ever doubt the balance team?

No one claimed that the game is totally balanced or that you shouldn't doubt the balance team. The point that was made was only that you need to be careful drawing conclusions from these stats, because they likely do not paint a complete picture. The non top 200 stats are likely more heavily influenced by some external factors such as map pool and (lack of) player skill, while top 200 stats with 200-400 games are too small a sample size to be reliable and obviously vary too wildly (60/40 wr in 3v3 vs 40/60 wr in 4v4) to be anything close to conclusive. Which really is a shame because as they'd likely be the best representation of balance.

Are Axis a bit too powerful in teamgames at the moment? I'd say probably yes.
Is it as bad as these raw stats seem to suggest? I don't think so.
14 Jun 2021, 10:12 AM
#70
avatar of Hannibal
Senior Moderator Badge

Posts: 3113 | Subs: 2

No one claimed that the game is totally balanced or that you shouldn't doubt the balance team. The point that was made was only that you need to be careful drawing conclusions from these stats, because they likely do not paint a complete picture. The non top 200 stats are likely more heavily influenced by some external factors such as map pool and (lack of) player skill, while top 200 stats are too small a sample size to be reliable and obviously vary too wildly (60/40 wr in 3v3 vs 40/60 wr in 4v4) to be anything close to conclusive. Which really is a shame.

Are Axis a bit too powerful in teamgames at the moment? I'd say probably yes.
Is it as bad as these raw stats seem to suggest? I don't think so.

Just to add to this, everyone trying to find the one stat that proves his own argument right and others wrong should consider that the number of games is literally so low that win rates easily vary by 10% per month and up to 20% if you switch between modes (3v3 and 4v4).

So if someone really thinks that this 5% difference in win rate were real proof, by lord I urge you to publish your method in some high level statistics journal because I guarantee you that there are literal millions of people out there waiting for a proper and reliable statistical test to draw good conclusions from such type of data.
14 Jun 2021, 10:39 AM
#71
avatar of Vipper

Posts: 13496 | Subs: 1


What?
Absolute nonsense. If there is reason to believe that the data might be wrong, you can't say "don't argue with me" and still try to push your statement.

The claim that there is an issue with stat is speculation in your part and he describes where he got his stat and how he used.


This data probably has severe issues. You cannot use it as a backup for your claim. It is not about who to argue with, it is about drawing conclusion based on proper data. Period.

There is also another analysis (as far as i can remember) where he compared times for tech between 1vs1 and 4vs4 and the conclusion is the same. There is simply not a massive difference.

Actually the only noticeable difference is CP and corresponding units.
14 Jun 2021, 10:44 AM
#72
avatar of Esxile

Posts: 3602 | Subs: 1



So if someone really thinks that this 5% difference in win rate were real proof, by lord I urge you to publish your method in some high level statistics journal because I guarantee you that there are literal millions of people out there waiting for a proper and reliable statistical test to draw good conclusions from such type of data.


That's not how standard deviation works. If we had factions being on the top 5% on certain modes and on bottom on other modes, I'll agree with you. But after 3 months seeing the same factions on top on each mode shows different things.

Few game are decided by early game in teamgame, what decides the game are late game units able to crutch on the opponent and we've seen the balance team working on downgrading Allied late game options while keeping Axis one intact.
It wouldn't surprise me to see the gap increasing once the new patch kick on.
14 Jun 2021, 11:47 AM
#73
avatar of Protos Angelus

Posts: 1515


Even though this went off topic, ultimately this conclusion leads us back to what is at hand.



No one claimed that the game is totally balanced or that you shouldn't doubt the balance team. The point that was made was only that you need to be careful drawing conclusions from these stats, because they likely do not paint a complete picture. The non top 200 stats are likely more heavily influenced by some external factors such as map pool and (lack of) player skill, while top 200 stats with 200-400 games are too small a sample size to be reliable and obviously vary too wildly (60/40 wr in 3v3 vs 40/60 wr in 4v4) to be anything close to conclusive. Which really is a shame because as they'd likely be the best representation of balance.

Are Axis a bit too powerful in teamgames at the moment? I'd say probably yes.
Is it as bad as these raw stats seem to suggest? I don't think so.


But then every tournament statistics would be invalid as they have an inherent low sample size. There is a difference between sample sizes and trends. Sure, the sizes are not great, but the trend is there, no matter the size. Month after month the axis have 5%+ win advantage which is huge. The only acceptable difference would be 50 +/- 2%. So [48,52]%. And the constant BS about the "pick only top 200" games for the analysis. I've seen sh** players in top 20 that spammed mortar pits and complained how axis is OP when they went down to lefh or stuka or werfer.
Rank means sh**. It's more of a general guideline than an actual rule. If I've managed to go from rank 20 to 300 and back to 50 in a span of 3 days, then you know that the rankings in complete random teams are BS at best. Complete and utter garbage on average.
While I do agree that the stats paint a worse picture than it is, you're not really doing anything to fix the balance problem for 3v3+ modes. A couple of gimmicky buffs for brits and USF (Pesrhing), coupled with hard nerfs (calliope late game arty tank that can only take one shot, while being slow AF) really make me wonder the sanity of your team.

I mean, would you call the following a HUGE BUFF?
Every unit in the brit arsenal, both stock and doctrinal is getting a 1% buff in one area.
On paper, you'd be able to say "we've buffed every unit and brits should now be OP", whereas in reality you've done nothing.

But I don't really blame your team for it. It takes a lot of brain processing power to do a complete overhaul and unless you're getting payed for this, I too wouldn't really be bothered to do a good job. Couple that with the absolute garbage map design in 3v3+ map pool, it's no wonder that the axis has win rate larger than f****ng 5%.
That and your inability to give any sort of late game power to USF that does not crutch on 2x jackson to keep the KTs and tigers in check (well, until the Jagd/ele arrives and the axis just does a hail Mary push with 3+ heavy tanks and elite infantry supported by stukas and werfers and you straight up lose unless you have doctrinal mines on USF). I mean, you don't even have to take the winrates here. There was a post in Trending on the front page, that showed the number of games each faction played in each game mode, both random and AT. Brits and USF were dead last in each mode except 1v1 for USF (which is expected as 1v1 is a completely different balance problem, and USF is the best faction probably for 1v1). Or just sum up the number from the current winrate page.
I'd say that's the biggest proof. The player's unwillingness to play USF/brits in teamgames is there for a reason.
14 Jun 2021, 11:49 AM
#74
avatar of Katitof

Posts: 17914 | Subs: 8



But then every tournament statistics would be invalid as they have an inherent low sample size.

They are.

Not once tournament stats were used as an actual metric of balance.
At best, it was indicator of a trend.
14 Jun 2021, 11:52 AM
#75
avatar of aerafield

Posts: 3029 | Subs: 3

jump backJump back to quoted post13 Jun 2021, 20:26 PMLeodot


Let me guess:
1) 99% axis play-rate
2) 4-digit ranks


Oh look, 99% axis playrate rank 3000 4v4 player

14 Jun 2021, 12:06 PM
#76
avatar of Hannibal
Senior Moderator Badge

Posts: 3113 | Subs: 2

jump backJump back to quoted post14 Jun 2021, 10:44 AMEsxile

That's not how standard deviation works. If we had factions being on the top 5% on certain modes and on bottom on other modes, I'll agree with you. But after 3 months seeing the same factions on top on each mode shows different things.

Fact is that the win rates can jump vastly between different months and even between related game modes such as 3v3 and 4v4. I agree that standard deviation is no indicator of balance being good or bad at all. But people here claim that some factions are clearly UP because of their win rate being lower than another faction's by a couple of percent, while next month or in the next game mode the difference is much smaller or even reversed. This is a strong hint that the data we have is not robust in the first place.
My point is that we have to consider much, much more than just the win rates that the site gives us. We cannot solve the balance issue on the current "hard" data alone. Like it or not, we have to make a ton of "educated guesses", and that's exactly what the balance team is supposed to do. The win rate data can help, but I get the impression that too many people assume it should be the main or even the only metric for balance. Given the quality of the data, this is quite naive.
Especially if we consider that the data is highly interdependent in team games. One bad faction will drag their Allied faction(s) down in the win rate. Some of the win differences can literally be a good team suddenly having two free weekends. Again, those win rates are a hint that something might be off, but definitely overblown in the discussions here.

jump backJump back to quoted post14 Jun 2021, 10:44 AMEsxile

Few game are decided by early game in teamgame, what decides the game are late game units able to crutch on the opponent and we've seen the balance team working on downgrading Allied late game options while keeping Axis one intact.
It wouldn't surprise me to see the gap increasing once the new patch kick on.

I think this has mostly to do with USF and especially UKF having so many holes in the line up compared to Axis factions that picking doctrines is not a nice addition, but essential to the faction. Allies have 2/3 of their factions hampered by this, while both Axis factions unlock all "basic" tools in a normal teching path (OKW having some quirks but still).

Overall the patch still goes into the right direction in my eyes, but as with every patch we have to check which changes do not work. Overall the patch should not be seen as a general faction balance patch, since fixing the factions via commanders will create more issues once core units get changes later down the line. Previous patches have fallen short on more fundamental changes to mostly UKF, but I'd attribute this more to Relic than the balance team.

Off note, I am quite happy though that team games are finally discussed for general balance. It actually shows us how far the game has come that the biggest issues have moved from easier to analyze 1v1s to chaotic large modes.
14 Jun 2021, 13:29 PM
#77
avatar of Esxile

Posts: 3602 | Subs: 1


Fact is that the win rates can jump vastly between different months and even between related game modes such as 3v3 and 4v4. I agree that standard deviation is no indicator of balance being good or bad at all. But people here claim that some factions are clearly UP because of their win rate being lower than another faction's by a couple of percent, while next month or in the next game mode the difference is much smaller or even reversed. This is a strong hint that the data we have is not robust in the first place.
My point is that we have to consider much, much more than just the win rates that the site gives us. We cannot solve the balance issue on the current "hard" data alone. Like it or not, we have to make a ton of "educated guesses", and that's exactly what the balance team is supposed to do. The win rate data can help, but I get the impression that too many people assume it should be the main or even the only metric for balance. Given the quality of the data, this is quite naive.
Especially if we consider that the data is highly interdependent in team games. One bad faction will drag their Allied faction(s) down in the win rate. Some of the win differences can literally be a good team suddenly having two free weekends. Again, those win rates are a hint that something might be off, but definitely overblown in the discussions here.


I think this has mostly to do with USF and especially UKF having so many holes in the line up compared to Axis factions that picking doctrines is not a nice addition, but essential to the faction. Allies have 2/3 of their factions hampered by this, while both Axis factions unlock all "basic" tools in a normal teching path (OKW having some quirks but still).

Overall the patch still goes into the right direction in my eyes, but as with every patch we have to check which changes do not work. Overall the patch should not be seen as a general faction balance patch, since fixing the factions via commanders will create more issues once core units get changes later down the line. Previous patches have fallen short on more fundamental changes to mostly UKF, but I'd attribute this more to Relic than the balance team.

Off note, I am quite happy though that team games are finally discussed for general balance. It actually shows us how far the game has come that the biggest issues have moved from easier to analyze 1v1s to chaotic large modes.


The patch is going to increase the late game issue because its not a balance patch, more of a players wishes release, The few changes that are really about balance making through it will not be enough. As I said previously this team isn't a balance team but a modding team adding up what they wish and covering it under a "balance" blanket.

Balance has been wreck some patches ago already, the in-game economy is a joke with the profusion of manpower avalaible unless you seriously screw-up early on and how the game allow player to have huge army composition with the constant reduction of Popcap for anything. Axis faction having everything stock wouldn't be a problem if, as some year ago, a player couldn't build standard army composition as of today. Nobody would build as standard 2 HMG and 2 Atgun in every game if every faction popcap were seriously taken in consideration for what its suppose to be: to generate gaps, to not allow players to be able to response to every threat as standard play, to force players to build around their strategy and counter their opponent´s one.
Today as a result of not having taken in account the economy side of the game and only focus on units stats (as if the game was just an excel) 2/3 of the balance aspects on which the game should rely doesn't exist anymore.

The way the balance is today is a dead-end. Not because you can't balance it but because many important aspects of it has been relegated into no-existance. And all of that just for having given the keys to modders and not people with a real vision about game design as a whole.

At the end I agree with you, those 5% deviation in favor for Axis faction will always remain for the good, because otherwise that would indicate that Allied factions are overperforming on their individual unit statistics.

On a personal note, because some here were talking about it, tournament on 1vs1 or 2vs2 are getting boring to watch, those actual TOP players leading the column now are boring to watch, they're specialist on planning, slow building their army composition, no taking any risk before having it all in place, they perfectly feet this meta of lack of lacks, no urgent decision making to counter their opponent, just sit build and wait that support so well Axis faction design on teamgame random.
14 Jun 2021, 13:49 PM
#78
avatar of LMAO

Posts: 163

delete sandbags and the game will be suddenly more interesting
14 Jun 2021, 14:43 PM
#79
avatar of blvckdream

Posts: 2458 | Subs: 1




What you said would make sense if there was some actual deviation around the 50% mark. According to the data I have presented there is no such thing. Allies are always losing with the only exception being the month of May in 3v3 Top 200.

I agree that having a 45-55% WR would be normal if both sides had times where they have 50+ WR%. But this is not the case, Allies deviate around the 43-44% WR mark and Axis deviate around the 55% WR mark. The data is from 3 full months for 2v2/3v3/4v4. More than 200.000 games taken into account.
14 Jun 2021, 16:17 PM
#80
avatar of Stark

Posts: 626 | Subs: 1

I just wanna bring up that most valid and fair matchmaking is 1v1. All of us can name hundred games with your enemy team was rank 1000+ as me being top30. On higher modes it's even worse.

And yes, rank means something. If you are rank below 200 it means it's very very possible your loose game because of your mistakes or you haven't used tools that faction gives you.
PAGES (8)down
1 user is browsing this thread: 1 guest

Ladders Top 10

  • #
    Steam Alias
    W
    L
    %
    Streak
Data provided by Relic Relic Entertainment

Replay highlight

VS
  • U.S. Forces flag cblanco ★
  • The British Forces flag 보드카 중대
  • Oberkommando West flag VonManteuffel
  • Ostheer flag Heartless Jäger
uploaded by XXxxHeartlessxxXX

Board Info

530 users are online: 530 guests
0 post in the last 24h
3 posts in the last week
23 posts in the last month
Registered members: 48733
Welcome our newest member, service
Most online: 2043 users on 29 Oct 2023, 01:04 AM