Login

russian armor

coh2chart and Its Worth

23 Jul 2016, 10:44 AM
#2
avatar of FichtenMoped
Editor in Chief Badge
Patrion 310

Posts: 4627 | Subs: 3

Nice breakdown!
23 Jul 2016, 11:06 AM
#3
avatar of pigsoup
Patrion 14

Posts: 4299 | Subs: 2

:) Siphon X. strikes again!
23 Jul 2016, 18:34 PM
#4
avatar of Jaedrik

Posts: 446 | Subs: 2

"On top of the factors we mentioned above, players just might need more time to adapt. In order to truly gauge the state of the balance, we need to go deeper."

To the contrary, balance does not exist outside of particular metagames. A state of balance only exists as the observed state of a particular metagame at a particular time. There is nothing outside of the past and present in regards to balance.

I admit, thought experiments can reliably prove that something may be ultimately balanced, but we humans are prone to failure and not taking everything into due account.

These distinctions ought to be made, otherwise people will more often use this article as a bludgeon for their poor thought experiments rather than humbly accepting criticism and working to improve one's theory on the future metagame.
26 Jul 2016, 00:38 AM
#5
avatar of Mistah_S

Posts: 851 | Subs: 1

USF win rate on the rise.
HMMMMMMM I WONDER WHY THAT IS!?
6 Aug 2016, 12:00 PM
#6
avatar of Rappy

Posts: 526

Great work Siphon.

I think a better balance chart would involve match up between levels, not between ranks as rank is only relative to the number of people who play each faction, whereas level is not. This is how I believe the matchup tries to match people, not on rank.

8 Aug 2016, 08:39 AM
#7
avatar of Siphon X.
Senior Editor Badge

Posts: 1131 | Subs: 2

jump backJump back to quoted post6 Aug 2016, 12:00 PMRappy

Great work Siphon.

I think a better balance chart would involve match up between levels, not between ranks [...]. This is how I believe the matchup tries to match people, not on rank.


Thanks!

Well, actually its neither, matches are supposedly made between players of a similar (hidden) ELO value.

This works out (at least for lower levels) so that roughly people of similar levels are matched, but it depends on the faction matchup as you can check here. For example, OH players seemed to be predominantly matched with USF players that are about 1.5 levels lower than them.

However, two details about levels, relevant in this context are:
- For ranks below 201, levels and ranks are more or less the same thing: A rank 200 OH player and a rank 200 UKF player both have rank 16.
- The problem at the ranks around 200 is that depending on the number of players for the specific ladder, certain levels are missing.

In the post I linked above I got around that by looking at top % (quasi levels) as well as top ranks. But in the end, the thing you'd really need is a significant number of samples between players on the same skill level; which of the four values that I give there reflects that best is up for debate.

The main point I'm trying to make in this post is that any statistics from automatch (regardless if you do it like in coh2charts, or like I did in beforementioned post) will be heavily influenced by matchmaking. In fact, the matchmaking is tasked to nullify any differences in balance and what you are left with are scrambled remnants of balance, combined with other matchmaking shortcomings. So, I'm not saying those are meaningless, it's just not the hard truth about balance that people often use it as in discussions.

Plus what Momo said in the other thread...
16 Aug 2016, 08:18 AM
#8
avatar of pigsoup
Patrion 14

Posts: 4299 | Subs: 2

to people who know stats, would allies having basically 50% more players due to having 1 more faction always skew the winrate for allies? as said in the article, if one side has more players, that side would not have enough counterparts on the opposite side.
16 Aug 2016, 09:04 AM
#9
avatar of Siphon X.
Senior Editor Badge

Posts: 1131 | Subs: 2

jump backJump back to quoted post16 Aug 2016, 08:18 AMpigsoup
to people who know stats, would allies having basically 50% more players due to having 1 more faction always skew the winrate for allies? as said in the article, if one side has more players, that side would not have enough counterparts on the opposite side.


I'd say no. Thing is, in order to play a game you have to have an allied player and an axis player. Like, you can't have more games on allied side than on axis side in total. If there really are more players on allied side, it means they have to wait longer for an axis player to become available so effectively the fewer axis players play more games in the same time.

Now, what may happen is that people in certain skill ranges play more or less games for a certain side. Assume VonIvan is the only player on axis side, and I and I2 would be the only player on allied side. In that case, Von plays twice and wins both and I and I2 would loose. So axis would have a 100% win rate.

In coh2charts this would look like axis has two games in the 1-250 bracket and a 100% win rate, and allies would have 2 games in the 500+ bracket and a 0% win rate.

So, generally, if more good players play on one side, they would eventually be matched with worse players from the other side, which in turn would result in a higher win rate for the first side. The correlation of excess players on the allied side in 1-250 bracket vs. winrate that is mentioned in the post could be an indication of that.

But again this should be independent of the number of factions. If all UKF games would be played by guys like me, axis winrates in the 1-250 bracket would go up. If UKF would be played only by Hans, axis winrates would go down.

Now, if you think about team games: Not sure. The general mechanic will still be the same, like when one side has better players, it will result in higher winrates, obviously. However, in team games it is less straight forward as the matchmaking has to create teams of equal skill, which I figure is very tricky.

Less players means that the pool of players that matchmaking can pick from is smaller. This might be both good and bad, depending on how good the assumptions behind the matchmaking algorithm are....
16 Aug 2016, 09:55 AM
#10
avatar of Siphon X.
Senior Editor Badge

Posts: 1131 | Subs: 2

I'll rephrase this, because after rereading my response I thought it wasn't overly clear:

Obviously, the relevant thing is not the number of players but the numbers of games played. And each game requires the same number of players on the axis and the allied side, so there can't be more games played as axis than as allies.

Now, if there are more players for one side (maybe because it has more factions), a number of players on the other side have to play more games.

If we look at coh2charts, there are more games played in the 1-250 1v1 bracket on allied side than on axis side. However, this conversely means that in the other brackets (plus non-ranked players) there need to be more axis players.

Yes, the fact that there are more high ranked games played on allied side will likely bias win rates in favor of the allies. But the issue is less that there are more games played on allied side in total (because that's not possible) but rather that there are more games played by skilled players on allied side.
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
  • Soviets flag T.R. TheSeekingGuy
  • Soviets flag T.R. Opportunity Cost
  • Oberkommando West flag VonAsten
  • Oberkommando West flag Jove
uploaded by Seeking

Board Info

92 users are online: 6 members and 86 guests
Esxile, javad, XARDAS, A_E, Vipper, Sander93
109 posts in the last 24h
1025 posts in the last week
4747 posts in the last month
Registered members: 19172
Welcome our newest member, paitowarna99
Most online: 1221 users on 25 Feb 2020, 12:03 PM