jaguilar/rank

rank

This project creates skill ratings for players and teams of players. It is still in a fairly rudimentary state.

The output below is the rating of the various teams playing in the spring split of the LCS for the first four weeks of play. The model of skill is a gaussian, where the mu represents the expected performance and the sigma represents the deviation in performance. In practical terms, this means that the mu represents "how good" the algorithm thinks the team is, and the sigma represents how uncertain the algorithm is about its guess (and/or how consistent the team is -- these two are more or less the same value).

The format of the following output is:

TEAM_NAME(SIGMA~MU)

After four weeks, the algorithm believes Team Curse and Team Dignitas are way better than the other teams.

> python rank_data.py 

clg(25.0~8.33) col(25.0~8.33) crs(25.0~8.33) dig(25.0~8.33) ggu(25.0~8.33) mrn(25.0~8.33) tsm(25.0~8.33) vul(25.0~8.33)

clg(26.0~8.21) col(25.0~8.33) crs(28.5~8.21) dig(24.0~8.21) ggu(23.0~8.25) mrn(25.0~8.33) tsm(26.2~8.21) vul(22.2~8.25)

clg(26.0~8.21) col(24.1~8.29) crs(29.8~8.15) dig(27.6~8.09) ggu(23.3~8.18) mrn(22.5~8.25) tsm(26.6~8.09) vul(19.8~8.15)

Upset: vul(19.8~8.15) beats clg(26.0~8.21) at odds of 0.24
clg(24.4~8.1) col(22.6~8.17) crs(30.4~8.13) dig(28.5~8.05) ggu(20.2~8.07) mrn(22.5~8.16) tsm(26.6~8.09) vul(24.5~8.04)

Upset: clg(24.6~8.0) beats crs(29.7~8.05) at odds of 0.27
Upset: ggu(18.3~7.98) beats vul(22.8~7.9) at odds of 0.3
clg(26.9~7.94) col(19.5~8.03) crs(29.1~7.98) dig(31.8~7.93) ggu(20.1~7.95) mrn(25.4~8.01) tsm(26.0~7.92) vul(21.1~7.86)

Crazy result: col(19.5~8.03) beats dig(31.8~7.93) at odds of 0.084
Upset: vul(21.4~7.79) beats dig(29.9~7.91) at odds of 0.16
Upset: ggu(19.2~7.91) beats clg(26.9~7.86) at odds of 0.18
clg(24.9~7.84) col(22.2~7.95) crs(31.3~7.88) dig(27.8~7.89) ggu(21.2~7.88) mrn(22.3~7.92) tsm(26.6~7.83) vul(23.4~7.78)

clg(22.2~7.74) col(19.8~7.86) crs(31.3~7.88) dig(29.6~7.81) ggu(21.2~7.88) mrn(22.3~7.83) tsm(28.7~7.74) vul(24.6~7.74)

Crazy result: clg(21.7~7.71) beats crs(32.3~7.84) at odds of 0.11
clg(24.5~7.67) col(19.5~7.75) crs(30.5~7.81) dig(30.3~7.72) ggu(23.1~7.78) mrn(21.8~7.73) tsm(28.7~7.64) vul(21.2~7.65)

Upset: col(19.5~7.75) beats clg(24.5~7.67) at odds of 0.27
Upset: vul(21.2~7.65) beats tsm(28.7~7.64) at odds of 0.18
clg(22.8~7.64) col(21.7~7.64) crs(29.7~7.74) dig(30.3~7.72) ggu(23.9~7.72) mrn(21.7~7.65) tsm(28.4~7.59) vul(21.2~7.58)

Upset: clg(22.8~7.64) beats dig(30.3~7.72) at odds of 0.18
Upset: mrn(20.5~7.6) beats clg(24.7~7.62) at odds of 0.3
clg(23.1~7.59) col(20.7~7.6) crs(29.0~7.67) dig(26.8~7.66) ggu(27.0~7.62) mrn(22.1~7.57) tsm(29.8~7.56) vul(21.2~7.58)

Upset: col(20.7~7.6) beats dig(26.8~7.66) at odds of 0.23
Upset: col(22.5~7.57) beats ggu(27.0~7.62) at odds of 0.29
Upset: vul(22.5~7.54) beats ggu(26.9~7.56) at odds of 0.29
Upset: dig(23.5~7.6) beats crs(27.9~7.61) at odds of 0.29
Upset: mrn(20.0~7.49) beats clg(25.2~7.5) at odds of 0.26
clg(22.5~7.43) col(23.6~7.46) crs(23.3~7.51) dig(25.5~7.51) ggu(25.7~7.47) mrn(23.3~7.43) tsm(32.2~7.45) vul(23.5~7.41)

clg(21.5~7.4) col(23.6~7.46) crs(21.6~7.44) dig(24.4~7.47) ggu(27.0~7.37) mrn(23.3~7.43) tsm(33.2~7.4) vul(25.1~7.32)