Forum
Elo (Skill) Rankings for Season 5
Created 2nd September 2009 @ 00:56
Add A Reply Pages: « Previous 1 2 3 Next »
nice idea but everybody has his own rating of teams :) because theres one big point this ELO thing or whatever its called cant rate and thats the personality of the team :p
I really could not be arsed to read the wall’s of text and looks straight at the table… i was thinking, isnt it flawed? You cant rank idk with teams from div1 using stats on games played last season, this may sound very dumb and theres probably a comeback but idk played div2 teams last season whereas div1 teams played harder opponents… LET THE FLAMING BEGIN!
I take it it’s just the ETF2L results taken into account?
I really could not be arsed to read the wall’s of text and looks straight at the table… i was thinking, isnt it flawed? You cant rank idk with teams from div1 using stats on games played last season, this may sound very dumb and theres probably a comeback but idk played div2 teams last season whereas div1 teams played harder opponents… LET THE FLAMING BEGIN!
Very true, and the biggest problem here is the initial ratings I gave to teams. Division 1 started with 2200, and division 2 started with 2000 points. I wanted a distance where good teams in division 2 could show they are division 1 worthy (ie, idk?), so if I increase the distance too much, you’d never see that.
I take it it’s just the ETF2L results taken into account?
At the moment, yes – only division 1 and 2a from the last season. But if there is interest I’m very curious about starting some sort of ladder based on this rating system when we have all the parameters sorted. The ladder would allow you to enter results from anywhere, such as importing ETF2L past history, Wireplay history, and enter friendlies. The idea behind it wouldn’t be a competitive ladder, it would be a comprehensive evaluations of teams and their performance.
so this point system shows us that div2 teams can compete with div1 top teams if they beat div2 teams?
so this point system shows us that div2 teams can compete with div1 top teams if they beat div2 teams?
There is no point to this system, don’t try to infer one.
It’s purely an interesting experiment in statistics, nothing more.
I promised a few 2b folk that I would compute the rankings for them too:
Standings:
1. Team Dignitas (league rating 2266, per-round rating 2130)
2. Crack Clan (league rating 2252, per-round rating 2195)
3. Team Coolermaster (league rating 2248, per-round rating 2414)
4. Team YoYoTech (league rating 2237, per-round rating 2218)
5. I Don’t Know? (league rating 2161, per-round rating 2288)
6. The Imperial (league rating 2155, per-round rating 2102)
7. Afterlife (league rating 2127, per-round rating 2106)
8. coolclan (league rating 2123, per-round rating 2154)
9. Fakkel Brigade (league rating 2120, per-round rating 2186)
10. rockit (league rating 2106, per-round rating 2240)
11. Team YoYoTech 2 (league rating 2038, per-round rating 2224)
12. LOL Team (league rating 2037, per-round rating 1982)
13. SHIFTplay (league rating 2034, per-round rating 2102)
14. Carpe diem (league rating 2029, per-round rating 2006)
15. dmg. (league rating 2025, per-round rating 1987)
16. eSrael (league rating 2025, per-round rating 2030)
17. animosity (league rating 2011, per-round rating 1975)
18. jAgarna (league rating 2002, per-round rating 2051)
19. eL Diabolo’s Brigade (league rating 1978, per-round rating 1931)
20. Team Fruit (league rating 1967, per-round rating 1941)
21. Dead Cows Strike Back (league rating 1965, per-round rating 1895)
22. Euphoria (league rating 1960, per-round rating 1895)
23. veAre.black (league rating 1932, per-round rating 1839)
24. The Devils Rejects (league rating 1918, per-round rating 1902)
25. Crosshair Gaming (league rating 1861, per-round rating 1831)
26. The Hampshire Heavies (league rating 1823, per-round rating 1777)
i never said anything about point i’m just saying it doesn’t work
How do you expect it to ‘work’, as cycles said it is purely interesting stats and a bit of tinkering, take it as it is really.
Continued working on this, finally added data about every season 5 game. This now has teams that dropped out, but still no default wins:
Divisions 1 – 6, all skill rated using a TF2 variant of Elo’s system (K-adjustments based on point difference). Whabaaaam! The outcome is different to what I posted before as I’ve up-ed the initial K-factor on this data to 60, because ETF2L is such a prestigious league (well, at least imo).
I’ll post something a bit more in depth about this tomorrow :)
Nice to see is very interesting imo u should do that at the end of the season
Again, very impressive. I like your idea to use the ELO system in a ETF2L ladder. Say, do you or can you learn (it’s quick and easy, I presume you wrote this exercise in c++), to code some php? :) *recruitment hat on*
Again, very impressive. I like your idea to use the ELO system in a ETF2L ladder. Say, do you or can you learn (it’s quick and easy, I presume you wrote this exercise in c++), to code some php? :) *recruitment hat on*
Written in Perl because it’s the best damn language ever invented, heh.
And you should talk to DeneusBeer, because he’s already recruited me ;) Well actually… I made him recruit me but yea… same thing. Would be nice to talk to all the admins together, maybe we could weave this into ETF2L properly :)
Add A Reply Pages: « Previous 1 2 3 Next »
content rss