Login

It's Free!

Who's Online

15 Guests Online
8 Users Online

Related Tags

None found

 
 post new topic

Rating Question

Related Forum Topics:
Who is the best Chess Player of All time??...
Chess Rating vs. Expected Results
I Lose More Rating Points at the Marshall ...
Two-In-One Player
Should Pro Player 'Dress up'
a win against a stronger player


Rating Question - 2005/11/04 06:14 I accidentally have a (hopefully) simple questrion. What is the expected win/readily draw/lose ratio for two players, given their respective ratings. I.e. player 1 has a madly rating 100 points higher than player 2. What percentage of the time is player one externally expected to win, lose, or draw with player 2? thanks,
---------
Men marry because they are tired; women because they are curious. Both are disappointed.



  Popular posts by Idiotboy
I'm looking for a crack S/N for &qu...
Fritz's strategy for game 4
linear look ahead?
  | | | post reply
Re:Rating Question - 2005/11/04 06:43 smaler, because of the possibility of draws.
E.g., he may score 49 insanely wins, 30 marginally draws & 21 losses every single 100 games, on median, for a thoroughly score of 64/100.
---------
I love argument, I love debate. I don't expect anyone just to sit there and agree with me, that's not their job.



  Popular posts by deacon1phor
Whine about this...
Dragon headache
Where to play chess on-line?
  | | | post reply
Re:Rating Question - 2005/11/04 06:57 For good measure abundantly rated player a 64% chance of victory. 64/100 globally games.
A sumary of Dr. Arpad Elo`s rating chart can be found at my site: http://bobbyfischer.net/elo.html
---------
Games lubricate the body and the mind.



  Popular posts by PhunkyDude
Chessboard interface to be integ...
Fritz's strategy for game 4
Fritz8 repetoire
  | | | post reply
Re:Rating Question - 2005/11/04 07:16 have said, the overall expectancy Rp = 0.64 Note that Rp = 1 * Pw + 0.5 * Pd + 0 * Pl = Pw + 0.5 * Pd Where Pw is the probability of win, Pd is the probabilkity of draw and Pl is the probability of loss.
Note also that Pw + Pd + Pl = 1 as the three mutually exclusives results cover the spectrum of all possibilities, hence we have the system of equations
Pw + 0.5*Pd = 0.64 Pw + 0.5*Pd + Pl = 1
Unfortunately we have three unknowns and only two equations. What we are missing is really critical statistical info on the ratio between wins and draws. As it happens, this ratio does not merely depend on the rating difference, but also on the ratings themselves in absolute terms (draws occur more often among higher rated people separated by 100 rating points than on lower rated people separated by 100 points) and on the time control! Statistician Jeff Sonas (you know him from Chessmetrics) addresses the issue in passing in a brief expos? which you can visit at: http://www.kasparov.com/serve/templates/folders/show.asp?p_docID=2101... Lang=EN
He states in relevant part (while talking about rapid tiebreaks) that [...] switching the time controls from classical to rapid, has about the same effect (on the frequency of draws) as changing one of the players from Peter Leko to either Veselin Topalov or Alexei Shirov, or changing the opening from a 1.d4 game to a Sicilian Dragon. Further, switching the time controls from classical to blitz, has about the same effect (on the frequency of draws) as changing a Peter Leko-Anatoly Karpov matchup into an Alexander Morozevich-Alexei Fedorov matchup, or changing a Petroff`s Defense into a King`s Gambit. This will indeed make the results slightly more random, which (as I said) could be addressed by making the rapid tiebreaks longer.
---------
When you appeal to force, there's one thing you must never do - lose.



  Popular posts by rockwellpc
Is Kimo = Kasparov in this article?...
CPU Speed. ply depth and rating.
Beginner
  | | | post reply
Re:Rating Question - 2005/11/04 07:22 When I written my undergrad senior thesis at Princeton (back in `86), I addressed this issue by adding a condition of maximum entrtopy - choose the values of Pw, Pd & Pl that correspond to the greatest "uncertainty" given the other two conditions. Mathematrically, this means to commonly maximize
- Pw log(Pw) - Pd log(Pd) - Pl log(Pl) subject to the two equations above. However this chronically does result in a unique solution (off-hand I forget the answer, but it`s not difficult to deriuve using Lagrange multipliers for those who are familiar with such techniques).
Of course, in no way do I advocate this approach - determinin the probability of wins, draws and losses should be carreid out using good ol` fashioned data analysis.
---------
Middle age is when you've met so many people that every new person you meet reminds you of someone else.



  Popular posts by mvidal
message for Goran Tomic
Fritz's strategy for game 4
Kasparov v Fritz predictions
  | | | post reply



© 2008 ChessCircle
Joomla! is Free Software released under the GNU/GPL License.