the funniest thing is seeing redditors backwards-reasoning their way into thinking online matchmaking systems have all these crazy metrics when literally all they do is just check an ELO ranking. guess it's easier to invent a bunch of mystical statistical reasoning that you have no evidence for than accept you are bad at a video game

"over a long enough timespan my win/loss rate always converges to 50%!!" yeah lol that's literally what the ELO ranking algorithm is designed to do you fucking clod

eventually your ranking will stabilize because ELO is literally statistically designed so that your ranking will converge at a number where you will win and lose 50% of the time, if you win more the number goes up and if you lose more the number goes down, this isn't magic


this post is peak redditor brain, a bunch of pseudo scientific nonsense dressed up like the OP knows a single thing he's talking about

> I just noticed I played around 1.5k games in those 9 months (holy fuck), so my sample size is actually considerable.

your "sample size" is actually tiny because literally tens of thousands of matches are played per day you moron lol

> if you don't have a BSc in CS you can still read the abstraction and conclusion

fucking redditors lamao

the redditor mind loves to believe that they believe only in extracting conclusions from hard data but will also assume that something like a "true matchmaking ranking" exists separately from the one that is measured by the system

vro theyre just measuring how much you win and lose against people with certain numbers lmao just get better at the game dumbass

i'm all for critiquing how we extract meaning from statistics and measurements but like, this system is literally just measuring how much you win and lose, not much more to say than that lol

> Given the problem described in the graph, it's almost certain the algorithm matching players is some type of PTAS (Polynomial-time approximation scheme).


@dankwraith i'm laughing my ass off at this part in the conclusion where they go "it does create more games where you have little-to-no control over the outcome." oh word? you played an entire 50-60 minute game of dota and had no control over how it ends?

@kara actually its pure luck that dendi got 7k mmr, it could have been any of us,

@dankwraith nerds will be presented with a chaotic system of many independent actors and think it's a ploy to get at them

@dankwraith I particularly love "I've identified win streaks happen when I win hard games and lose streaks happen when I play badly even against opponents who are playing worse than I normally do". Sure sounds like you've devised some corkboard and string system whose outputs are indistinguishable from one that tries to create even games.

@dankwraith "how do you improve win rate? Well, by getting better at the game, and encouraging your team to play well. I am very smart."

@dankwraith hi there! cs student who looked at a statistics book once,

@esvrld @dankwraith that's the secret, i got the MMR vaccine as a child and now i'm bad at video games

@SapphicGiraffic @esvrld @dankwraith Ohshit! *Immediately regrets getting an MMR booster a couple years ago*

Sign in to participate in the conversation is a place for friends