Battle 1062714 detail - Zero-K

Loading...

Map: Cobalt Dream v1.0

Cobalt Dream v1.0

Title:	MM 4752: 1v1, Rank Neutron Star
Host:	Nobody
Game version:	Zero-K v1.9.2.0
Engine version:	104.0.1-1544-ge1f249f
Battle ID:	1062714
Started:	3 years ago
Duration:	10 minutes
Players:	2
Bots:	False
Mission:	False
Rating:	Competitive

Watch Replay Now
Manual download

Team 1
Chance of victory: 97.3%

rank

Admin

Bakuhatsu

Team 2
Chance of victory: 2.7%

rank

mad_kraft

Spectators

rank

Flipstip

rank

Fumica

Filter: Player:

sort
Legomenon 3 years ago (edited 3 years ago) I'm surprised mad_kraft is getting matched with Bakuhatsu here with a 650~ WHR gap, after mad_kraft has lost several games to very strong opponents previously today ( bloa and Flipstip in the previous two games). My understanding based on https://zero-k.info/Forum/Post/230325#230325 is that the matchmaker would skew towards giving easier games after several defeats in a row, is my understanding wrong or is the maximum WHR range so wide that this is an acceptable pairing despite the streak modifier? +0 / -0
DeinFreund 3 years ago I can't check the exact numbers here, but it also depends on the previous enemies of the opponent. Maybe Bakuhatsu got some strong opponents too? +0 / -0
Legomenon 3 years ago (edited 3 years ago) Bakuhatsu didn't play any MM in the last couple of days (and those games were wins against weaker opponents), so should have been a blank state for this game. Bakuhatsu did get a bunch of losses against Kingstad/GoogleFrog in casual games on a test version (and therefore not counting for WHR) earlier in the UTC day, which should not impact MM of course. +0 / -0
Steel_Blue 3 years ago madcraft sometimes finds me in matchmaking. +0 / -0
Bakuhatsu 3 years ago mad_kraft must be secretly op! +0 / -0
esainane 3 years ago At game start, mad_kraft had 1978 elo, and Bakuhatsu had 2613 elo. Replicating rating curve compression with last known parameters (which may have changed since) - https://www.desmos.com/calculator/xc5bg9zuki - the lowest matchable elo at 500 search width for a 2613 elo player should be 1983 elo, before any win/loss streak adjustments. It seems reasonable to assume win/loss streak adjustments are capable of moving boundaries by at least 5 elo. So this seems to be within expected parameters of the current system, and only barely outside expected parameters of the old system. DeinFreund are these values still accurate? +0 / -0