Forum
My first ban....THIS SUCKS
|
fiero600 wrote
at 1:37 AM, Monday February 25, 2008 EST
ok so anyone that knows me...knows that i dont like Pat Whalen...he's as close to a PGE as you can be without actually suiciding on him..cuz thats not how fiero rolls. anyways... today i went to sit at a table and get the message "You are not allowed to sit with Pat Whalen, for the remainder of the month."
Ironically....its 230AM and there's only 1 high stakes tables playing at a time...so i cant play unless i can beat him to red... how stupid...how can i be banned from someone i dont even like? |
|
Ryan wrote
at 1:03 PM, Monday February 25, 2008 EST "ignoring games where one does poorly but the 'partner' does well is a mistake"
It doesn't ignore these games. |
|
MadHat_Sam wrote
at 1:23 PM, Monday February 25, 2008 EST Oh, that was what I gathered from the table chat. What does it ignore then? Games where both do poorly?
|
|
|
Ryan wrote
at 1:42 PM, Monday February 25, 2008 EST yeah
|
|
|
fiero600 wrote
at 1:49 PM, Monday February 25, 2008 EST wait wait.... so ppl can be banned from eachother for doing well togther in only "5%" of the games?
|
|
|
MadHat_Sam wrote
at 1:53 PM, Monday February 25, 2008 EST Ignoring games where both do poorly is a even greater mistake. The whole sum of games really needs to be looked at Ryan. A true RC would account for all the games that two players played together not just the ones where one person does well. Cherry picking the stats of wins doesn't really show the true picture of what type of correlation there is between player X and Y.
|
|
kdicefreak wrote
at 2:32 PM, Monday February 25, 2008 EST good job ryan. if i understand you correctly, you are ignoring the correlation between -1 and 0, right? in probability terms, you are looking at probabiliity > 0, ignoring the negative portion of the curve. i think it's a great idea.
i hope the RC keeps on improving and eventually prevent all those behaviours that takes the fun away from the game. |
|
|
MadHat_Sam wrote
at 2:59 PM, Monday February 25, 2008 EST I could see some reasoning in wanting to avoid the ability for people to swing the RC by tanking a bunch of low rated games but you do also have the point values gained and lost and could have corrected for this.
The RC catches some good correlations don't get me wrong, but the fact that one can get spiked for 3-4 games with 1-2 finishes for player X-Y but ignore like 10-20 games where each of them finished 5-7th, that seems like ignoring a large portion of statistically relevant data. |
|
|
Ryan wrote
at 3:12 PM, Monday February 25, 2008 EST It's not 3-4 games but you must have played more than 10 games where either of you have done well. It looks at all of these games and determines if the distance between the two are abnormally high.
|
|
|
MadHat_Sam wrote
at 3:36 PM, Monday February 25, 2008 EST Ryan, I kinda get what you are saying.
Would you mind posting the data that spiked me and yellowfin. I am really curious because fin is far from a player I would have expected to spike. I think with the low limit of 10 games there will be some silly RC hits early in the month when fewer players are playing the higher tables. |
|
Pat Whalen wrote
at 12:43 AM, Tuesday February 26, 2008 EST It actually does make sense to disregard the games in which both players finish badly because it is the same as including them. It shows a correlation in their play, and for PGAers who can't even play the game without a teamate (Imp and Prove), when one loses, the other has a better chance of losing as well
|
