* forgot to mention, only live games

So your ranking is based on most games played?

Go fuck yourself, bolshoit. Check out my GR and suck a dick.

Wow. You bash him no matter what he does, whether it is a joke or not.

i think i need to refine the model, like a global warming activist, until you are ranked below 100.

oh come on! if i were ranking based only on games played, i would have shown up somewhere!

Hopefully TGM sees my potential as well and puts me in the top 100 :)

What is your model?

@barnet unless you mean total games played over the user's entire career. but my code wouldn't have that information anyway.

Well, I'm reticent to bash a system that ranks me as #1. :P I just assumed that was because I've played a lot recently.

@zeus i don't know, i think it's something like GR, where if you lose to a stronger player it's not a huge disadvantage. i wrote the code a couple days ago in an ad-hoc fashion without a coherent plan so i can't entirely remember... it took me till now to release the results cause i didn't have the html files saved.

@barnett... yeah i'm not sure. i had a slightly modified version that ranked like this
1. xorxes 2. jc 3. EmmaGoldwoman 4. Frank 5. Scotty Baker 6. Skittles 7. Vaftrudner 8. BumChinRaptor 9. rokakoma 10. Poozer
etc
but then i noticed people in the list only had like 1 game in the list, so i just went with this other one. but for this top-10, i know what the model is, but i just don't know if people would like the result.

@Barn, its likely because you win every game you play as well.

Can you rank me, mr. shoi?

I like the result, less so the insult, but I had a very solid Jan/Feb so I would expect my ranking to be high regardless of methodology. But it would be interesting to see what actually went into the rankings.

Lando, if you're not in the top 25 of any gunboat rankings, it automatically makes them suspect :)

Lol, I don't think I have ever played this cat before.

It would be interesting to have a live game rating and for it to require a certain amount of games played to be in it. Just curious about the method of how you arrived at those results.

hmm.

@lando you may have helped find a bug in my html-parser. for some odd reason i have games with you in it where there are no winners (or drawers) recorded. this is what i have for you in terms of live wtagb jan/feb.


Winner(s) P-man, Barn3tt, MichiganMan,
Loser(s) pjmansfield99, Lando Calrissian, Dharmaton, Sargmacher,
WWinner(s) 2ndWhiteLine,
Loser(s) MiscellaneousBiscuit, Oliwoli, jmeyersd, MichiganMan, Lando Calrissian, The Czech,
Winner(s) NigeeBaby,
Loser(s) Fortress Door, ava2790, taos, myn, Mandala, Lando Calrissian,
Winner(s) Alexander the Awesome,
Loser(s) Dbivins, Lando Calrissian, Dharmaton, d3154, indianajones, mathieupoirier,
Winner(s) Dharmaton,
Loser(s) LordofDoom, Texastough, Stimpson, indianajones, Darth Masta, Lando Calrissian,
Winner(s)
Loser(s) mathieupoirier, Bob Genghiskhan, csmith98, Sargmacher, jc, indianajones, Lando Calrissian,
Winner(s)
Loser(s) Govitti, dtoubs33, b501, BALLS DEEP, Actaeon, jacobcfries, Lando Calrissian,
WWinner(s) Barn3tt, 2ndWhiteLine,
Loser(s) The Chinmeister, Indiana Monts, MichiganMan, Lando Calrissian, Sargmacher,
Winner(s) The Czech,
Loser(s) santosh, xpedior, Partysane, SplitDiplomat, Lando Calrissian, jmeyersd,
Winner(s) 2ndWhiteLine, MichiganMan, Dharmaton,
Loser(s) dr. octagonapus, drake2346, nnfolz, Lando Calrissian,

you know what? maybe the space in your name is doing it. those 2 games i'm guessing you did win.

or maybe you won more than that. i don't know. my code sucks. hahahahaha.

No idea, brew, sorry.

Is your system counting draws as wins? It looks like you only have wins/losses.

http://research.microsoft.com/en-us/projects/trueskill/details.aspx
This is interesting. It seems to be ready to handle multiple teams and draws. From what i have read Elo isn't really good for multiple teams. Kind of interesting i guess.

How about: Wins = 4, Survive=1, Draw=2, Loss=0. . _ (Discounting the D. 'S)
Total all (W.+S.+D.+L.) x 25% ==> 00.00% to 100.00% rating.

i found the problem in my code. it wasn't the space in the name, it was generosity. hahahahaha. the "donated" icon messed up the parsing. so i fixed that small defect and i ran it again, i'll post the results when i go to the other computer and copy them over.

@dharmaton, i can compute those results too, do you want them? the only issue i see with that is there is no penalty for losses, so somebody who survived 3 games and lost 10 will have the same score as someone who played one game and drew it. drew may not be correct english.

or sorry i mean a higher score, 3 vs 2.

ok, this should be better. the old ones shafted all users who donated.

1. Barn3tt 2. Dharmaton 3. The Czech 4. SplitDiplomat 5. Sargmacher 6. NoFog 7. jc 8. 2ndWhiteLine 9. Lando Calrissian 10. LakersFan 11. chris-d-9 12. nnfolz 13. jmeyersd 14. SocDem 15. DILK 16. MichiganMan 17. Filliam H Muffman 18. xorxes 19. 5 minutes 20. EmmaGoldwoman 21. Dudlajz 22. pjmansfield99 23. Vaftrudner 24. Scotty Baker 25. rokakoma 26. Poozer 27. BumChinRaptor 28. HITLER69 29. President Eden 30. ericisawesome 31. Alexander the Awesome 32. moskowitz 33. ericfogg 34. cspieker 35. Frank 36. Indiana Monts 37. stranger 38. Sydney City 39. caesar101dog 40. mapleleaf 41. carne 42. frenchtourist 43. Ramse 44. hugu37 45. santosh 46. pbuerre 47. captain 48. diplomancer83 49. Arminius 50. skoczek86

@bolshoi: yes please. The 'floor' is at zero, instead of minus one, makes for easier & direct calculation.

...oh,... : -1 for Resigns.

@dharmaton one problem though is i only have two groups for each game. winners, which is either the winner or the list of people who drew. and losers which includes survivors resigners and defeated

so you want to count a 3-way draw the same as a 7-way draw? cause i don't have any minimum phases requirement or anything like that.

Count the Draws on a curve... ;) ... So your telling me: all you can count is a win/loss ratio, ie: x%. for Rank.

actually i guess i can count anything if you want to. i'm just not sure how much better a survived is than a defeated or resigned. but yeah, if you give me exactly the ranking you want, i can calculate it, even if if considers resigns worse than defeateds for some reason. i'll work on it tomorrow though, just pm me exactly what you want. i'll take a look tomorrow.

I don't think there's much difference between a survive and and a defeat. Of course, I almost always play wta.

bolshoi you are.... good for something?

I'd like to see your finished product, with a clear explanation of how your metric is calculated (much like TGM does). It's a noble undertaking, but I'm still not sure how you intend to improve on GR.

You should definitely have a minimum phase requirement to throw out draws due to CDs in the first few turns. I don't anticipate it making a big difference in the rankings, but it ensures it only includes "quality" games, as opposed to games with a CD where a few stubborn folks insist in drawing with six or seven rather than canceling.

ok i published the results so that everybody can see for themselves what i'm using to do the calculations. remember these are only LWTAGBCV. and only from jan/feb.

https://docs.google.com/document/d/1RWkE1fGax6HC7vFO-74JX7j24s-BUTyLgtb-vGdhdtA

#28 is my lucky number...

coincidence or galactic alignment !!???!?

blank document, bolshoi?

hm?? i guess that /edit is necessary in the link for some reason.

https://docs.google.com/document/d/1RWkE1fGax6HC7vFO-74JX7j24s-BUTyLgtb-vGdhdtA/edit

hahaha. yeah. constant win due to constant errors. that is bolshoi.

Hey, I have an objection. Your initial post lists purported winners and LOSERS. I'd like to say that just because I lost, that doesn't mean I'm a LOSER. :-)

Those who won and those who didn't.... that is usually my word choice.

OK, so you have win/draw vs the rest, but how are you going to rate people based on that?

Thats what I'm not really understanding.

you don't understand cause i haven't said, but it basically tries to find the best players and pull them out one by one from the list of players. but like i said, i'm open to any rating system people want. i might calculate solo/loss ratio or something like that. the calculation i have now isn't really that simple.