alu/tra-superscript
1
0
Fork 0
mirror of https://github.com/titanscouting/tra-superscript.git synced 2025-09-26 07:10:18 +00:00
Code Issues Packages Projects Releases Wiki Activity

experimental trueskill support

Browse Source
This commit is contained in:
Dev Singh
2022-03-28 10:13:42 -05:00
parent 9752fd323b
commit 3fe2922e97
4 changed files with 110 additions and 28 deletions
Download Patch File Download Diff File Expand all files Collapse all files

40
competition/module.py
View File

@@ -1,8 +1,10 @@
import abc
import data as d
import signal
import trueskill
import numpy as np
from tra_analysis import Analysis as an
from ts_predict import win_probability
class Module(metaclass = abc.ABCMeta):
@@ -180,9 +182,9 @@ class Metric (Module):
red = {}
blu = {}
for match in matches:
red = d.load_metric(self.apikey, self.competition, match, "red", self.config["tests"])
blu = d.load_metric(self.apikey, self.competition, match, "blue", self.config["tests"])
@@ -198,6 +200,9 @@ class Metric (Module):
gl2_red_vol_total = 0
gl2_blu_vol_total = 0
ts_red_team = {}
ts_blu_team = {}
for team in red:
elo_red_total += red[team]["elo"]["score"]
@@ -206,6 +211,10 @@ class Metric (Module):
gl2_red_rd_total += red[team]["gl2"]["rd"]
gl2_red_vol_total += red[team]["gl2"]["vol"]
tmu = red[team]["ts"]["mu"] or 25
tsigma = red[team]["ts"]["sigma"] or 8.333
ts_red_team[team] = trueskill.Rating(mu=tmu, sigma=tsigma)
for team in blu:
elo_blu_total += blu[team]["elo"]["score"]
@@ -214,25 +223,43 @@ class Metric (Module):
gl2_blu_rd_total += blu[team]["gl2"]["rd"]
gl2_blu_vol_total += blu[team]["gl2"]["vol"]
tmu = blu[team]["ts"]["mu"] or 25
tsigma = blu[team]["ts"]["sigma"] or 8.333
ts_blu_team[team] = trueskill.Rating(mu=tmu, sigma=tsigma)
red_elo = {"score": elo_red_total / len(red)}
blu_elo = {"score": elo_blu_total / len(blu)}
red_gl2 = {"score": gl2_red_score_total / len(red), "rd": gl2_red_rd_total / len(red), "vol": gl2_red_vol_total / len(red)}
blu_gl2 = {"score": gl2_blu_score_total / len(blu), "rd": gl2_blu_rd_total / len(blu), "vol": gl2_blu_vol_total / len(blu)}
print('here')
if match["winner"] == "red":
observations = {"red": 1, "blu": 0}
ts_obs = [1,0]
elif match["winner"] == "blue":
observations = {"red": 0, "blu": 1}
ts_obs = [0,1]
else:
observations = {"red": 0.5, "blu": 0.5}
ts_obs = [0,0]
ts_red = list(ts_red_team.values())
ts_blu = list(ts_blu_team.values())
new_red, new_blu = trueskill.rate([ts_red, ts_blu], ranks=ts_obs)
new_red_ts = {}
new_blu_ts = {}
for key, value in zip(ts_red_team.keys(), new_red):
new_red_ts[key] = value
for key, value in zip(ts_blu_team.keys(), new_blu):
new_blu_ts[key] = value
print("red" if win_probability(new_red, new_blu) > 0.5 else "blue", match["winner"])
# now there are new trueskll ratings for each team based on the win/loss from DB
red_elo_delta = an.Metric().elo(red_elo["score"], blu_elo["score"], observations["red"], elo_N, elo_K) - red_elo["score"]
blu_elo_delta = an.Metric().elo(blu_elo["score"], red_elo["score"], observations["blu"], elo_N, elo_K) - blu_elo["score"]
@@ -250,6 +277,9 @@ class Metric (Module):
red[team]["gl2"]["rd"] = red[team]["gl2"]["rd"] + red_gl2_delta["rd"]
red[team]["gl2"]["vol"] = red[team]["gl2"]["vol"] + red_gl2_delta["vol"]
red[team]["ts"]["mu"] = new_red_ts[team].mu
red[team]["ts"]["sigma"] = new_red_ts[team].sigma
for team in blu:
blu[team]["elo"]["score"] = blu[team]["elo"]["score"] + blu_elo_delta
@@ -258,6 +288,10 @@ class Metric (Module):
blu[team]["gl2"]["rd"] = blu[team]["gl2"]["rd"] + blu_gl2_delta["rd"]
blu[team]["gl2"]["vol"] = blu[team]["gl2"]["vol"] + blu_gl2_delta["vol"]
blu[team]["ts"]["mu"] = new_blu_ts[team].mu
blu[team]["ts"]["sigma"] = new_blu_ts[team].sigma
temp_vector = {}
temp_vector.update(red)
temp_vector.update(blu)