Community data collection project. Lets test if K factor impacts LP gains/losses.

Kai Guy·6/20/2019, 6:27:49 PM·1 votes·1,259 views

Lets try to work backwards here to understand Riots MMR system better. For those who have not looked into Elo systems. K factor is part of the adjustment formula. Rn=Ro+K(W-We). It is used to represent system uncertainty and the reason it is used is so accounts with low # games move on the ladder more rapidly. In the context of league, This gets smurfs/newbies out of the starting MMR range quicker with proper implementation.
There are 2 common ways for the K variable to change. 1. number of games. 2. Position on ladder. (higher MMR has lower K) I expect riot uses both.

What I would like to get help with is checking for if it impacts LP gains. To check this what I need is screen shots of LP gains and losses at the end of a match and the OP.GG for what that match's average team MMR and your current MMR. If you can provide multiple that's better.

Its likely impossible to reverse engineer their formulas but I think its feasible to find out if K impacts LP with enough data points to check for a trend. Going to look for if there is a soft cap for LP in higher MMR and if low # games accounts get higher value changes compared to high # game accounts.

0 Comments