lingqingmeng
commented
10 years ago In Starcraft's dynamic placement matching service, a winning streak (5 or so games in a row) results in a broadening "frontier". This frontier doesn't actually improve the K factor for ELO scores, its specifically implemented so people get matched with levels that the game deems appropriate based on your ELO acceleration -> (d(dELO/dt) )/ dt. If you (the player) keep beating these exponentially-increasing difficulty matches, your ELO will naturally start growing faster obviously since your large "frontier" is matching you up with super hard levels, sooner than a person would if he's not on a massive winning streak.
On Tue, Apr 22, 2014 at 1:05 AM, Patrick Severin notifications@github.comwrote:
I don't care how it's done, but I think Baylor has a point: it takes too long to get to your appropriate ELO. Either a high initial K factor, placement matches, or something else that would move ELO faster for skilled players is a must, seeing how we'll probably only get a bit of time out of each player. @AOEChamp https://github.com/AOEChamp and @FourierTransformer https://github.com/FourierTransformer may have ideas about this.
— Reply to this email directly or view it on GitHubhttps://github.com/4950/DirtyGame/issues/365 .
Quillwerth
I don't care how it's done, but I think Baylor has a point: it takes too long to get to your appropriate ELO. Either a high initial K factor, placement matches, or something else that would move ELO faster for skilled players is a must, seeing how we'll probably only get a bit of time out of each player. @AOEChamp and @FourierTransformer may have ideas about this.