This user has no status.
This user has no status.
🏆 Top 1% Commenter
Well-Known Member
Well-Known Member
Hello everyone!
I started to make an ELO ratings ladder for the WTT circuit players, to see what is the real strength of our favourite players.
However, I encountered a problem: since I have to enter the results manually, it takes a lot of time. There aren't many WTT Contender and CHampion events, but Feeder matches quite a few, with a high amount of players. I am considering just leaving feeder events out, because many players won't have enough matches to assess their strength, and I am not sure if it's worthwhile, as I would guess most of us is only interested in the top 100 players anyway. It would be nice to have the Bundesliga, CL, Pro A, CTTSL and T League results in as well, but I am not sure if I will have the energy to manage even the Feeder events.
However, if there is interest, I can be convinced to do them. I want some feedback on to where to draw the line, what to include?
-
Sidenote: yes, I am aware of ratingscentral, but while not including players with a high deviation might be more fair, it makes the rankings unusable. Also, it's blatantly incorrect and hard to read anything into it. I am not sure of the reasons as I have not studied the math and resources behind it, but for me it's not even fun.
Currently I started with an Elo based app for tracking the results, but I plan to switch to Glicko-2 (and in the future to some custom machine learning based solution - given enough time).
I started to make an ELO ratings ladder for the WTT circuit players, to see what is the real strength of our favourite players.
However, I encountered a problem: since I have to enter the results manually, it takes a lot of time. There aren't many WTT Contender and CHampion events, but Feeder matches quite a few, with a high amount of players. I am considering just leaving feeder events out, because many players won't have enough matches to assess their strength, and I am not sure if it's worthwhile, as I would guess most of us is only interested in the top 100 players anyway. It would be nice to have the Bundesliga, CL, Pro A, CTTSL and T League results in as well, but I am not sure if I will have the energy to manage even the Feeder events.
However, if there is interest, I can be convinced to do them. I want some feedback on to where to draw the line, what to include?
-
Sidenote: yes, I am aware of ratingscentral, but while not including players with a high deviation might be more fair, it makes the rankings unusable. Also, it's blatantly incorrect and hard to read anything into it. I am not sure of the reasons as I have not studied the math and resources behind it, but for me it's not even fun.
Currently I started with an Elo based app for tracking the results, but I plan to switch to Glicko-2 (and in the future to some custom machine learning based solution - given enough time).
