Season 2: Winners Cup Week 2 Top 32 Recap

The top 32 is done and so it is time to use  random statistics to try and extrapolate uncertain 'facts'.

One fact I am particularly keen to investigate is whether the best players are actually the best players. There was a discussion on discord recently where several high rated players stated that everyone is actually rubbish and that the ratings don't matter as not enough games have been played yet. Let us see if there is any truth to these wild, unsubstantiated accusations.

First let us look at the simple results

Tag	ID	Rating	Pod Place	Clan	Vs	Tag	ID	Rating	Pod Place	Clan
MindsDesire	155	1965	1st	Dragon		Erik	131	1854	2nd	Crane
Bayushi Yuuji	24	1792	1st	Scorpion		Daidoji Micah	169	1853	2nd	Crane
Kingsley	98	1978	1st	Dragon		Cody	107	1657	playoff	Crab
Heinzel	89	1707	playoff	Scorpion		Asahina Neph	161	1611	2nd	Crane
Jokingly	166	1774	1st	Crab		Kiramode	117	1801	1st	Lion
Shosuro Sakuya	118	1677	playoffs	Scorpion		Fordream	209	1769	1st	Scorpion
Action_johnny	249	1591	playoffs	Crab		SirLargeness	127	1740	2nd	Scorpion
Sahini Sashimi	194	1776	1st	Dragon		Calimsha	171	1635	2nd	Scorpion
Jeremy	206	1700	1st	Crane		Wydwen	218	1541	2nd	Scorpion
Bill.jr	48	1710	playoffs	Lion		Sotomatic	193	1704	2nd	Phoenix
Tsuke	160	1681	1st	Lion		Sruman	144	1765	playoffs	Lion
Shenful	162	1874	2nd	Scorpion		BryonKrane	181	1881	1st	Scorpion
Mirumoto Kai	202	1876	2nd	Dragon		Shosurpo Kiseki	4	1434	2nd	Scorpion
Chutereve	52	1814	1st	Crane		Togashi Komhl	180	1780	1st	Dragon
Shosuro Wild	158	1626	1st	Scorpion		Isawa Eujin	139	1841	2nd	Phoenix
Goblin Guide	219	1695	1st	Lion		VinceVelorn	116	1608	1st	Crab

So what fabulous facts can we garner. Let us initially avoid the thorny question of skill and look at how well the clans are doing (apologies to Unicorn but as they had no players in the top 64 (who played) so we are just comparing the remaining six clans)

The average is a consistent 16.666%. So Scorpion was dominant the 64 and increased that dominance in the 32 but fell back slightly in the 16. The next most successful clan is Dragon which started below average but then climbed up in the 32 and then drastically improved in the 16. Of the others Crab has  kept their participation percentage steady. Crane started above average but has dropped steadily since. Lion has had a horrible time, though this is partially due to two mirror matches (with few players to begin with) occurring in both top 64 and top 32. Phoenix was below average in the 64. Dropped badly in the 32 but due to its two players both winning has managed to make some of that loss back again.

Now to the thorny question of quality. If the supposition that the ratings have not had enough games to be valid then we would expect fairly frequent beatings of high ranked players by low ranked players or at least a lot of volatility in results. For these purposes let us assume that a heavily 'favoured' player would be one with a rating one hundred points higher than the opposition. Now this does not mean that the favoured player 'will' win just that if you were betting their chances of winning would be better (any given Sunday and all that). If you look again at the chart I have highlighted in green heavily favoured players and in red their opponents. Cream means the game is broadly average. So I expect the green players to win and the red players to lose. Cream players could go either way. What then are the results?


85.71% of heavily favoured players beat their opponents
thus only 14.28% of unfavoured players  (just the one, Wydwen) broke the odds and won.

Let's look at it another way. Players in the top cut came in via three methods. 1st and 2nd place in a pod auto qualified. 3rd and 4th went through a playoffs (one wasn't actually needed due to drops but in principle they would have). Now this is broadly indicated by the ratings as only the top pod players get a dominant rating by winning a lot whereas the playoff players by definition will have lost some games so will (usually) have lower ratings.

In the top 32 we have

1st placed pod players:  46.875%
2nd placed pod players: 34.375%
playoff pod players: 18.75%

after the top 32 these percentages change to

1st placed pod players: 50%
2nd placed pod players: 37.5%
playoff pod players: 6.25%

So as with the ratings you can only say that those players who came 1st and 2nd in the pods continued and increased their dominance in the play offs at the extreme expense of the play off players. You could also say that as the competition continues the 1st and 2nd players are steadily increasing their dominance.

Therefore at this point the ratings are indicating, with a high degree of accuracy, who the better players currently are and as such have confirmed their validity. The 'ratings don't count and we are all crap' supposition has been proved incorrect.

None of this says that a lower ranked player cannot or will not beat a higher ranked player just that it is unlikely. For the remainder of the winner's cup the prospective bill.jr for people who love beaing the odds this season is Wydwen.

Finally regardless as to the last match result (which has not played yet but is even in ratings and involve two Scorpions so does not affect the calculations above). The next round will have five games with perceived dominant players and three that are broadly equal. The gap between the dominant and their opponents is dropping though and when we reach the top 16 nerves, stress and luck can all cause larger issues so it will be interesting to see if the dominant players continue in the same vein. Ignoring the average games and looking just at 1st/2nd/playoffs we have four 1st vs 2nd games, two 1st vs 1st, one 1st vs playoff  and finally one 2nd vs 2nd.

I look forward to some quality games, hopefully all streamed...