Game count...

[KyberPrizrak]

Injustice?

[amstan]

????

[Tortortor]

the questions is: why some bots have 10 games in the past and at the same time some bots have only 4 games. They have similar rank, so they should have same "next_game_priority". Isn't it?

[erdman]

And some of us are stuck way down in the rankings because we have only played four total games, three of which are 1-v-1 games against starter-ish bots, with the 4th match being 3-handed. Hard to make much progress in the rankings by beating a single weak bot.

[amstan]

Sigma is not the same as game count. You can sometimes be more unsure of a rank after a game happens.

Example:
You won 2 games in total, but you play a third game and you end up last. At that point you would like to play more games just to be sure you're actually supposed to be on top.

[fourmidable]

From quickly looking through the "lastest games" list, I obverse a strange behavior. It appears most players selected in the latest games already have a large number of games played (8, 10 or even 16), while most of the contest participants are still at 4 games played. I'm wondering what makes the pairing algorithm prefer these players. Is it because they now have a lower sigma (from having played more games), and thus are seen as a better benchmark for the seed player?

[Zaph]

[amstan]

[amstan]

As i suspected. The discrepancy of the game amounts is due to the trueskill matchmaking.
They'll all even eachother out after cutoffs.

Janzert wants at least 10 games for everyone.

[bluegaspode]