Open sprunk opened 1 year ago
Battle: https://zero-k.info/Battles/Detail/1597034
Nightwatch message: [02:10] Nightwatch 10 players balanced Normal to 2 teams ( 1=29%) : 2=71%)). 252 combinations checked, spent 0ms of CPU time
The average current rating of the players was a := (2902+1507+1504+1474+1258)/5 = 1729
for team 1 (1 according to playerlist) (the one with MSPR), and b := (1776+1688+1579+1576+1248)/5 = 1573.4
for team 2. With d := 10^((a-b)/400)
, the chance to win for team 1 is d/(1+d) = 71%
if I'm not mistaken.
^ looks like this line is incorrect and calculates the loss chance instead? There's also some weird 2 / #teams
factor involved. Perhaps check what happens in FFA since it's likely to be incorrect in a more obvious way there.
Apparently "team 1" and "team 2" refer the other way around in
gamestart message vs the playerlist. Ask!predict
dunno#1049
on discord for moreEDIT: it seems to only be the game start message, predict uses different code.