questionable matchmaker behaviour
-
For the record I was not in a premade.
-
@blackyps yeah my bad i forgot to. i just assumed the 1400 wasn't in one because i've never seen him before lol
-
me and blast were not in a premade
-
What’s the game id?
-
@BlackYps BTW now that I finally have access to the server logs again I think we can find out what the parties were for a particular match with a bit of digging. In this case it looks like everyone was searching separately, but I think you already identified the issue.
-
-
@Askaholic what issue have I identified? If Maiar and Lyuker were not premade then something funky is going on
-
@BlackYps Oh I assumed your PR to cap the new bonus was related. I’ll send you the info on this match though in case it helps with something else.
-
That PR was actually not related to this specific match
-
gonna steal the topic kappa
Nobody was teamed up with the 1400, but it choose the 1400 to match in the game even though i was searching for longer (he matched in a previous match too that failed to launch, unless that somehow saves the total queue time).
Didn't ask everyone but i know everyone well enough to say that i'm 99% sure the only party that existed was macro/rezy together.
Which means a team of something like me (2600) + eco (2100) + jagged (1700) + Xayo(1900) = 8.3k vs 2000+2200+2300+1800= 8.3k could've been made.
There's roughly 10 games with an average rating of +-500 from 1400, while there are 2 with an average rating of +-800 for me and this was one of them. Which means i have to wait until both the games are over, which has already taken 20min and counting, until i can have a next potential match, while in the meantime there's already been 4+ launches that are within 500 average rating of 1400.
Not sure if i missed something
-
game id pls, so we can search the logs
-
#18899750
-
M Morax referenced this topic on
-
Just going to tie these two threads together here:
https://forum.faforever.com/topic/5405/better-tmm-map-pool-selection-algorithm
-
M Morax referenced this topic on
