EPTHistory of tournament results
| Forum Index > General Forum |
|
xebra
United States225 Posts
| ||
|
Locked
United States4182 Posts
but honestly, its unlikely you'll reach something as decent even as KeSPa, which is a fairly decent ranking for current players or you can look at that all-time ranking they use if that is what you want also using amateur things would not only be impossibly hard to compile due to many factors, including disorganization and informality but also not very good for a critical rating was that 5 edits for me? about average ~_~ | ||
|
EvilTeletubby
Baltimore, USA22247 Posts
| ||
GrandInquisitor
New York City13113 Posts
| ||
|
xebra
United States225 Posts
| ||
|
MoltkeWarding
5195 Posts
| ||
|
decafchicken
United States19921 Posts
That will get you to april 08 2005 | ||
|
nortydog
Australia3067 Posts
the editor's note at the end makes me sad  | ||
|
xebra
United States225 Posts
| ||
|
SoleSteeler
Canada5396 Posts
 who wrote it? | ||
|
Locked
United States4182 Posts
On November 26 2005 20:12 xebra wrote: In fact, the chart in the wikipedia article is mine, and some people (including myself) are dissatisfied with it, though it is interesting, nonetheless. As for the KeSPA ratings, they are not good, mathematically speaking. For example, they ascribe different weights to results at different events, which is not good. Events shouldn't even be rated in the first place, games should be, and rating points should be gained or lost based on your performance, your current rating, and your opponent's current rating. That doesn't mean KeSPA ratings are utterly terrible at predicting match outcomes, but using the Sonas system we can actually devise a rating system that is provably better at predicting match outcomes than any other, which would therefore make it the best. I want to develop such a ratings chart, but I need a lot more data than just who got first and second place. I need a record of every pro game ever played. If Doggi lost to Clon in the second round of the challenge league of the first OSL in 2004, I want to know about it. but the thing is..... events SHOULD be rated? imagine if i have this friend, who, with $10,000 on the line, i would beat with about an 80% concistency, however we were only matched in a bo5 of that calibur once and i won it 3-1. now lets say there was some amateur clan match where i found i was matched against that same friend and we were just having fun so in several games i dt rushed pvz, zealot rushed pvt, and then tried a 4pool, summarily losing 0-3 does that mean with a total of 3-4 official matches between us he should be ranked higher than me? or maybe i'm misunderstanding your idea of weighting "games but not events"? | ||
|
QuietIdiot
7004 Posts
Rank = R Points = P S = Score P / R = S Points are attained/lost by the win/lose you have on a person with a certain rank. For example: in Pgt, a C- losing to a D- would lose lots of points but in turn, the D- player would receive alot. Thus moving him up the ladder and moving his opponent down. Average out there rank from their points and comes out an interval that can be used to be averaged to the nearest whole number, etc. O my god, that doesn't make any sense at all, o whatever, the chess system is fair enough imo, and I'm IN THE steve fan club now Weeeeeeeeeeeeeeeeeeee (SO tired so sleepy ) | ||
FakeSteve[TPR]
Valhalla18444 Posts
| ||
|
FakeSteve[TPR]
Valhalla18444 Posts
| ||
|
draeger
United States3256 Posts
After deleting all of the repeats (boxer vs yellow on Boxer's page and yellow vs boxer on Yellow's page,etc), I had a comprehensive list of every game played in progaming history as recent as the beginning of this last liquibet season and back to the ancient days of pro-gaming. The final idea was to have a program that would predict liquibet matches for me. I could say "Boxer vs Yellow" and it would run a statistical equation of my choosing to determine who would win based on past record, past race record, record of each player on the given map, and record of each race on the given map, etc. It could also take into account streaks and whatnot. It never got finished though. While msot of the games are there, the only thing I actually implemented was a way to search for racial record on a given map and a method for adding new games to the database. It would've been something like this: http://www.deported.net/progaming/ (only 1 link actually works) | ||
|
GrandInquisitor
New York City13113 Posts
On November 26 2005 20:33 MoltkeWarding wrote: http://en.wikipedia.org/wiki/Teamliquid Wikipedia articles like that make me sad. It demeans what Wikipedia is supposed to be about, because an article written on us by us is so hopelessly POV and non-neutral it will never be encyclopedic. | ||
Liquid`Jinro
Sweden33719 Posts
;p And yeah, http://www.pgr21.com used to have a very up-to-date ranking which took into account every single game ever played +_+ Unfortunately it has not been updated since february this year, and they don't have the matchup rankings/vs race rankings. | ||
|
AnGuRuSO
Kazachstan323 Posts
![[image loading]](http://img418.imageshack.us/my.php?image=xebra3kj.jpg) Yea that's really mature... | ||
|
xebra
United States225 Posts
Draeger, is your data just like (Boxer:Yellow, 123:100), or is it actually a chronological list of all games played, by whom, and on which map? If it is the latter, that's just what I'm looking for. Can you send it to me? | ||
|
draeger
United States3256 Posts
CREATE TABLE `translations` ( `player_id` mediumint(9) NOT NULL auto_increment, `english` varchar(20) NOT NULL default '', `korean` tinyblob NOT NULL, PRIMARY KEY (`player_id`) ) ENGINE=MyISAM DEFAULT CHARSET=latin1 AUTO_INCREMENT=713 ; CREATE TABLE `maps` ( `name` varchar(25) NOT NULL default '' ) ENGINE=MyISAM DEFAULT CHARSET=latin1; CREATE TABLE `games` ( `game_number` int(11) NOT NULL default '0', `date` varchar(20) NOT NULL default '', `league` varchar(30) character set euckr NOT NULL default '', `winner_id` int(11) NOT NULL default '0', `winner_race` varchar(7) NOT NULL default '', `loser_id` int(11) NOT NULL default '0', `loser_race` varchar(7) NOT NULL default '', `map` varchar(30) NOT NULL default '', PRIMARY KEY (`game_number`) ) ENGINE=MyISAM DEFAULT CHARSET=latin1; It's not the best database layout, but it's the best I could do with the data I received. If you think it would be helpful, let me know what format you would like it in (tsv, csv, mysql dump, etc) and I can email it to you. To answe your question, it keeps track of every game played. You should know this already because I linked you to the page where you could search matchup ratios on various maps. It listed every game that it counted. So for instance, if I do a quick query on the games table: select * from games limit 1; It'll return a row which contains the following data: <games> <game_number>13423</game_number> <date>2005-02-18</date> <league>IOPS 04~05<br>4° A?¶ 1°æ±?</league> <winner_id>14</winner_id> <winner_race>Zerg</winner_race> <loser_id>1</loser_id> <loser_race>Terran</loser_race> <map>Neo Guillotine</map> </games> The ids can be joined with the players table to determine the name. | ||
| ||
