Stats from the WCS 2019

Looking at just the round of 32, the winner of the match won 48 out of 52 matches played (92%). If one faction was overperforming like you think, the win rate would have been 48 out of 80, (60%) and the winner of the coin toss would have picked Soviets every time. In the round of 16, the winner of the match won 83% of the games, again not anywhere near 60%. The quarter finals and semi-finals had a 86% win rate by the match victor, again not anywhere near 60%.

We're not looking at the win:loss of individual players, we're looking at factions and sides. Players having an 85%+ win rate is irrelevant, since even a 100% win rate (3:0) translates into a 66% side win rate.

This is why I pointed out the "worst case possible", since that's what the side win:loss ratio can realistically be, assuming that the coin flips were actually random (i.e. no incredible coincidences of the better player ALWAYS playing allies first).

/edit

Like I said, this would be a lot easier with the complete data of every match - winner, and faction selection. Right now it's a pretty incomplete picture with a lot of extrapolation.

If that's the case, then the WCS rule-set was ignored.



It's also important to note that while the players had faction pick, they did not have 'side' pick until game 5; meaning players were always alternating between axis and allies every game (until game 5, where the leading player could chose to no longer alternate).

As EffenNewbie pointed out, "Game 5" only happened five times (1x R32, 2x R16, 2x finals); so that means 27 out of the 32 bracket games must have ended in a win-loss ratio of either 50:50, 66:33, or 33:66 axis:allies, where the player who was playing Axis/Allies was entirely random. In the case of a game 5, regardless of which side a player chose, the win:loss ratio must have been 60:40 for either axis or allies.

As a result of all this, the "worst possible case" (where the better player always played allies more) decreases, since we're now introducing 50:50 and 60:40 games into what was once exclusively a set of 66:33 games - meaning that Sov's win ratio against OST is now even higher than before over that "worst possible case".

For example, if 14 matches were 3:0, 13 matches were 3:1 and 5 matches were 3:2, and all matches went in favor of allies, the "worst possible case" goes down to 58.6:41.4, putting Sov 12.6% over what the maximum possible win:loss ratio is in a balanced game. There is simply no statistical way in which Sov did not over-perform vs ost; even when taking into account impossibly imbalanced matches and exceedingly unlikely coincidences in coin flips.

I'd be able to get a more accurate worst-case by knowing the outcomes (and faction picks) of every single match, but I haven't found that written down anywhere.

No. If Soviets, or any other faction were OP, game 5 would always happen and the winner would always be Soviet. This was so far from that result that I would have a hard time believing it had it not happened. The results look like something you would expect from a Warcraft II tournament, where all of the factions were actually the same but had different skins.

The only other thing that was abnormal was how well the seedings actually matched the players. There were practically no upsets, with the biggest upset being the #5 seed beating the #4 seed. The finals were between the 1, 2, 3 and 5 seeds. Whoever ranked them was either really good or really lucky.

https://docs.google.com/spreadsheets/d/14xEZNFVV08mLTxAxXHoepNUijXY-Elciyxx8GTrcCb4/edit#gid=1975264398

Not completely true. Well, if the games would have been played in a BO3 you would have a solid point. However, given that this is B05, it seems more tricky. I asked Sturmpanther about the how this actually was played out and if I understood correctly the way this worked was that the player that won the first two games had faction selection in the third match.

So, faction selection was not completely random, the better player typically would end up playing the faction that he would consider to be stronger.

Regardless of how the coin flips went, the four finalists got to choose factions roughly half of the time in six matches, versus players that left in the first round getting to choose factions in one match.

No. If Soviets, or any other faction were OP, game 5 would always happen and the winner would always be Soviet. This was so far from that result that I would have a hard time believing it had it not happened. The results look like something you would expect from a Warcraft II tournament, where all of the factions were actually the same but had different skins.

The only other thing that was abnormal was how well the seedings actually matched the players. There were practically no upsets, with the biggest upset being the #5 seed beating the #4 seed. The finals were between the 1, 2, 3 and 5 seeds. Whoever ranked them was either really good or really lucky.

https://docs.google.com/spreadsheets/d/14xEZNFVV08mLTxAxXHoepNUijXY-Elciyxx8GTrcCb4/edit#gid=1975264398

That would assume Soviets had a 100% win rate - they don't, and no one is claiming that. They had a 54.2% win rate against OKW and 71.2% against OST.

I'm just pointing out that the Soviet vs. OST win rate is so much higher than any other matchup that they are OP beyond any reasonable doubt. If their win rate against OST was 55%, or even 60%, there would be a discussion to have.

That would assume Soviets had a 100% win rate - they don't, and no one is claiming that. They had a 54.2% win rate against OKW and 71.2% against OST.

I'm just pointing out that the Soviet vs. OST win rate is so much higher than any other matchup that they are OP beyond any reasonable doubt. If their win rate against OST was 55%, or even 60%, there would be a discussion to have.

You seem to badly want this conclusion, but the data doesn't say that at all. You can't take the results of a bunch of 3-0 matches and conclude anything. The "effect" of the player is so much stronger than faction, map, rng, or any other factor.

You seem to badly want this conclusion, but the data doesn't say that at all. You can't take the results of a bunch of 3-0 matches and conclude anything. The "effect" of the player is so much stronger than faction, map, rng, or any other factor.

You absolutely can; I've shown this multiple times now. A single 3-0 match concludes nothing, a TON of 3-0 matches actually concludes a lot, provided most (or all) players play each faction an equal number of times. The effect of the individual player is mitigated almost entirely because of the number of games AND due to the random 'side' assignment.

Here's an example: a 12-match/5game series between 3 players (A, B and C), where A is very skilled, B is intermediate, and C is new. Let's also assume there's only 4 factions (Sov, USF, Ost, OKW) in this example, and players chose to alternate when they can (Sov -> USF, Ost -> OKW). For the sake of simplicity, let's say that only 2 factions are allowed per "match", otherwise calculating totals gets crazy, and we need a lot more games for the example to work (So Match 1 / Game 1-3 is always Sov vs. Ost, and Match 2 / Game 1-3 is always USF/Ost, etc.). I think this is reasonable, since the actual series had 84 games, not including the round-robin matches; we're using a total of 36 games (12 matches x 3:0 games).

Lastly, since these are horribly stacked games, let's assume any win is a clean sweep - 3:0. This pretty much describes my "worst possible case", except that each player has played each team an equal number of times.

Match #	Allies	Axis	Winner	Faction
Match 1	Player A	Player B	3:0 Player A - B	2:1 Sov/Ost
Match 2	Player A	Player C	3:0 Player A - C	2:1 USF/Ost
Match 3	Player B	Player C	3:0 Player B - C	2:1 Sov/OKW
Match 4	Player B	Player A	3:0 Player A - B	2:1 Ost/USF
Match 5	Player C	Player A	3:0 Player A - C	2:1 OKW/Sov
Match 6	Player C	Player B	3:0 Player B - C	2:1 OKW/USF
Match 7	Player B	Player A	3:0 Player A - B	2:1 Ost/Sov
Match 8	Player C	Player A	3:0 Player A - C	2:1 Ost/USF
Match 9	Player C	Player B	3:0 Player B - C	2:1 OKW/Sov
Match 10	Player A	Player B	3:0 Player A - B	2:1 USF/Ost
Match 11	Player A	Player C	3:0 Player A - C	2:1 Sov/OKW
Match 12	Player B	Player C	3:0 Player B - C	2:1 USF/OKW

From what I understand, this is pretty much what you've described as 'bad data' due to selection bias. The player's skill isn't matched well, so the games are always massively in favor of one player, and always 3:0.

Now let's look at the totals.

Sov vs Ost	Sov vs OKW	USF vs Ost	USF vs OKW
3:3	6:6	6:6	3:3

Sov Wins	USF Wins	Ost Wins	OKW Wins
9	9	9	9

Despite the players being horribly matched, the faction win/loss ratios are perfectly even. This is because for every 3:0 match between the uneven players, the factions actually win 2:1. When you average this out over enough matches (this is just 12), every 2:1 USF/OKW match ends up being balanced out by a 1:2 USF/OKW match - resulting in 50:50 win/loss ratios.

Conveniently, 50:50 is pretty much what we saw with Sov/OKW (54.2%), USF/Ost (54.5%), USF/OKW (53.3%), and even UKF/Ost (40%). Since we know that the vast majority of games were 3:0 (or close to that), it aligns my theory on even distribution. The better player always won, but over time, their wins with each faction/side balanced out.

So how do we get to 71.4% for Sov/Ost? Could the Sov player consistently be the better of the two? This is possible, but given the outcomes of the FOUR other match-ups, it's unlikely. I think it's much more likely that the 'worse' of the two players happened to win a few times due to game balance.

a TON of 3-0 matches actually concludes a lot

Soviets were picked roughly 40 times. OKW and OST picks are 2:1 in favour of OKW, so without knowing the exact number of SOV-OST matchups, I would assume it's somewhere around 13. That is not a ton of matches. With such a low amount of games, even one loss (which can be caused by anything from big mistakes, to really bad RNG, to bad player skill matchups, to simply losing a very close match by 1 VP) can skew the win rate by 5-10%. TLDR in my opinion this SOV vs OST winrate tells us absolutely nothing conclusive.

If that's the case, then the WCS rule-set was ignored.

Ah, thanks, I failed to find this. Seems like I misunderstood Sturmpanther. I actually could check if they alternated every time, but.... likely I won't. So let's assume they did.

The effect of the individual player is mitigated almost entirely because of the number of games AND due to the random 'side' assignment.

But there is no random faction assignment...

If I check out the games I have for the four semifinalists (and seems I have everything there except VonAsten vs. Stalingrad):

VonAsten played exclusively as OKW and USF.
Luvnest almost exclusively went SOV and OKW. He had 3 games as OH, none of these vs. SOV. He also was the only one playing as SOV against OH.
Noggano played mostly OKW and USF; he has 2 games as SOV, and 2 games as OH vs. SOV (winning 1).
Jove played 6/5/5 games as SOV, OKW and USF and only 2 as OH (one against SOV, one against USF, winning both).

Obviously you can argue that the players that mained OH didn't get to the semifinals because OH is shut down hard by SOV, but that's conjecture.

And then number of games is still limited...

I'm not saying the numbers are meaningless, they certainly can indicate a trend. However, I doubt that the data base is sufficient to prove something without the shadow of a doubt given the many factors that play a role.

We won't.

WWWHHHHYYYYY NOOO....

Why have units uncounterable?

60TD Ez mode here. 45:32 mark
Tightrope mentioned shreks, but at 60 range, how hard to make the pgren run and gun the distance.

Why have units uncounterable?

Units are counterable.

You being incapable of that doesn't change it.

60TD Ez mode here. 45:32 mark
Tightrope mentioned shreks, but at 60 range, how hard to make the pgren run and gun the distance.

Where are axis anti tank guns?
Where is axis infantry?
Do you see these massive empty unprotected flanks on left and right of the map?
Where was SCAS?

You know when units are uncounterable?
When you don't build counters to them, or have counters on other side of the map, as evidenced by that vid you gave as "proof".

Units are counterable.

You being incapable of that doesn't change it.



Where are axis anti tank guns?
Where is axis infantry?
Do you see these massive empty unprotected flanks on left and right of the map?
Where was SCAS?

You know when units are uncounterable?
When you don't build counters to them, or have counters on other side of the map, as evidenced by that vid you gave as "proof".

Where are Atg?
Costs constrain, KV8, IS2, Incendiary..and..much easier to get decrew by Allies infantry with their dps fire on move...katsuya... pak howie...Pak slow ass reposition vs incoming tanks...
Sorry i doubt paks will the day since pak are easy targets.

The main show is how easy for 60Td to just walk in and clear Axis 'heavy armor' tanks these days.

Low risk high rewards.

This the 2v2 i keep experience with 60Td.
So frustrating you cant fight back cos you are as so short.

Where are Atg?
Costs constrain, KV8, IS2, Incendiary..and..much easier to get decrew by Allies infantry with their dps fire on move...


The main show is how easy for 60Td to just walk in and clear Axis 'heavy armor' tanks these days.

Low risk high rewards.

So frustrating you cant fight back cos you are as so short.

And axis had frag bombs to stun all these TDs and SCAS to easily destroy them.

You're making excuses for a perfect example of what happens if you don't bring ATGs or don't even attempt to flank mass TDs.

Sorry, relic can't patch bad army compositions just because few handicapped players are incapable of getting X when opponent gets Y.

And axis had frag bombs to stun all these TDs and SCAS to easily destroy them.

You're making excuses for a perfect example of what happens if you don't bring ATGs or don't even attempt to flank mass TDs.

Sorry, relic can't patch bad army compositions just because few handicapped players are incapable of getting X when opponent gets Y.

It is no excuse. It is about having units that have clear oppressing advantages.

Besides, relic has pivoted to make Wehr support teams much easier to kill/decrew.

You speak as if ATg and SCAS are freely available to pop out. Or Wehr have cheaper units to throw out and flank a 60Td assualt

It is no excuse. It is about having units that have clear oppressing advantages.

Where were units that have clear, oppressing advantage against TDs again? You know, ATGs and AT infantry?

Besides, relic has pivoted to make Wehr support teams much easier to kill/decrew.

Please, explain to us all how its easier to decrew wehr support teams then USF or UKF or OKW support teams, we all want to know.

You speak as if ATg and SCAS are freely available to pop out. Or Wehr have cheaper units to throw out and flank a 60Td assualt

Ironic.

Where are Atg?
Costs constrain, KV8, IS2, Incendiary..and..much easier to get decrew by Allies infantry with their dps fire on move...katsuya... pak howie...Pak slow ass reposition vs incoming tanks...
Sorry i doubt paks will the day since pak are easy targets.

You speak as if KV-8, incendiary, katy, pak howie are freely available to pop out. Or allies have cheaper units to throw out and decrew 60 range ATGs under fire of LMG grens, PGs, HMG42, Pwerfer, brummbar and Tiger.

You absolutely can; I've shown this multiple times now. ..

No, you're still not understanding. When the matches are almost all 3-0, wining faction rate is determined solely by which faction was picked. Imagine for a second that instead of 64 players, there was just 2 players, Player B who won all of the matches and Player A who lost all of the matches (matches were all 3-0). If Player B plays more than half of his/her games as Soviets, you would conclude that Soviets are OP because the win rate was more than 50%

You keep trying to assert that players choice is random and follows a normal distribution, but it isn't and doesn't.

I'm not saying the numbers are meaningless, they certainly can indicate a trend. However, I doubt that the data base is sufficient to prove something without the shadow of a doubt given the many factors that play a role.

+1