< Earlier Kibitzing · PAGE 26 OF 26 ·
|Jun-28-06|| ||Kangaroo: Ivanchuk - Shirov
<<1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O b5 6. Bb3 Bc5 7. a4 Rb8 8. c3 O-O 9. d4 Bb6 10. axb5 axb5 11. Na3 d6 12. Nxb5 Bg4 13. Bc2 d5 14. h3 Bxf3 15. Qxf3 exd4 16. Bg5 dxc3 17. Nxc3 Nd4 18. Qd3 h6 19. Bxf6 Qxf6 20. Nxd5 Qe5 21. Kh1 Nxc2 22. Qxc2 f5 23. Qc4 Kh7 24. Rae1 fxe4 25. Rxe4 Qxb2 26. Re2 Qa3 27. Qc2+ Kh8 28. Ne7 Rf6 29. Ng6+ Kg8 30. Ne7+ Kh8 31. Ng6+ Kg8 32. Rfe1 Qc5 33. Qa2+ Kh7 34. Ne5 Rd8 35. f4 Rxf4 36. Qb1+ Kg8 37. Nd3 Rxd3 38. Qxd3 Qf5 39. Qd1 Kh7 40. Re5 Qg6 41. Re6 Rf1+ 42. Rxf1 Qxe6 43. Qb1+ Kg8 44. Rc1 Qf7 45. Qe4 Qf6 46. Rd1 Kh8 47. Rb1 Kg8 48. g3 Qd4 49. Qe6+ Kh7 50. Qf5+ g6 51. Qf7+ Qg7 52. Rf1 Qxf7 53. Rxf7+ Kg8 54. Rd7 Ba5 55. Kg2 c5 56. Kf3 Bc3 57. g4 Bf6 58. Rc7 Bh4 59. Ke4 c4 60. Kd5 c3 61. Ke6 Bg5 62. Rxc3 Kg7 63. Rc7+ Kg8 64. Rb7 Bh4 65. Ke5 Bg5 66. Ke4 Bh4 67. Kf4 Kh8 68. Ke4 Kg8 69. Kd5 Bg5 70. Ke6 Bh4 71. Rc7 Bg5 72. Rd7 Bh4 73. Rd4 Bg5 74. Ra4 Kg7 75. Ra1 Kg8 76. Rh1 Bh4 77. Rf1 Kg7 78. Rf7+ Kg8 79. Rf6 Kg7 80. Rf3 Bg5>
|Jun-28-06|| ||Hesam7: Look up "Devoretsky's Endgame Manual" page 229. It states that the following diagram is drawn:|
click for larger view
|Jun-28-06|| ||DutchDunce: Congratulations to Rublevsky for winning the tournament. With his fine performance, he should be breaking the 2700 barrier in the upcoming rating list?|
|Jun-28-06|| ||you vs yourself: Huge win for Rublevsky! A great comeback after the disastrous olympiad.|
|Jun-28-06|| ||percyblakeney: <With his fine performance, he should be breaking the 2700 barrier in the upcoming rating list?>|
Since the next list comes already in a few days I doubt the Aerosvit results will be included, but a very good tournament for Rublevsky in any case.
|Jun-28-06|| ||aw1988: Wow, what a tournament!|
|Jun-28-06|| ||Xaurus: <he should be breaking the 2700 barrier in the upcoming rating list?>In the upcoming rating list, due out Saturday, Rublevsky probably drops a bit since he did so bad in the olympics. The results from Aerosvit won't count until the October rating list.|
|Jun-28-06|| ||Xaurus: To be more specific, Rublevsky lost 15 points from the olympiad.|
|Jun-28-06|| ||percyblakeney: Rating changes after Aerosvit:
|Jun-28-06|| ||suenteus po 147: Congratulations to Rublevsky for winning a top notch tournament against some of the world's elite!|
|Jun-28-06|| ||Ezzy: <DCP323 - But how can somebody so limited in his openings be a Super-GM, let alone a national champ? He (Rublevsky) makes it easy for his opponents to prepare against him,> Obviously not|
<DCP23 June 18th - Let's just wait and see if Rublevsky wins any major event or another Russian Championship from now on.> You didn't have to wait very long did you.
<DCP23 - first, if Rublevsky would ever be invited to Linares or Corus A, he would finish at the bottom of the list, and second -- there are many who are more deserving than he is to be called Russian Champion:> After all your negativity about Rublevsky, the least you could do is congratulate him on a great performance and admit that your predictions are pretty bad.
|Jun-29-06|| ||hitman84: <Harikrishna -12.8> OUCH!|
>Great! He deserves to be a super GM!
|Jun-29-06|| ||plang: "Rublevsky +21.1
If i add these up and divide by 12 I get +2.2. That means that the average performance in this tournament is +2.2. Shouldn't the average performance be 0.0? What causes this? Isn't this ratings inflation?
|Jun-29-06|| ||Xaurus: <plang>Must be rounding errors somewhere. :)|
|Jun-29-06|| ||plang: That is too big for a rounding error. I am still trying to understand ratings inflation. In another thread there is an attempt to compare Steinitzs rating with 21st century players. Unless you understand ratings inflation and account for it this type of analysis isn't very useful.|
|Jun-29-06|| ||percyblakeney: Those are the rating changes given by http://www.echecs.com/actualite/une...|
|Jun-29-06|| ||plang: These are the same as listed earlier in the thread. I still don't understand why they are biased upwards. There was a net gain of 26 rating points in this tournament. That does not make sense to me.|
|Jul-02-06|| ||refutor: <plang> if you add 21.1+9.5+14.4+3.9+3.9+0.6+0.5 you get 53.9|
if you add 6.1+5.6+17.7+9.5+12.8 you get 51.7
so if you subtract them you get a net gain of 2.2 (if you divide that by twelve the average performance in the tournament is 0.18333) this makes sense, as the biggest rating jump was by the top guy, meaning that he was rated lower and seriously outperformed the expected rating. look up the rating formula, it is not necessarily zero sum
|Jul-03-06|| ||plang: Ok, I'm confused. I thought in each game the same number of points are subtracted from the loser as gained from the winner. If the winner is gaining more that would, of course, be inflationary.|
|Jul-04-06|| ||refutor: not necessarily...the idea is that players like rublevsky are improving faster than the rating system can handle. if that didn't happen, the rating system would actually be *deflationary*. let me give you an example|
a young up and coming player (let's call him magnus c.) goes from 2100-2500 in a period of one year. in a one:one scenario this means that he takes 400 points from his peers. that means (let's say) 8 people lose 50 points not through any loss in their strength, but because "magnus c." increased so rapidly and they were just unfortunate to be playing in the same tournaments as him. think about it.
let's say "alexei s." and "vladimir k." were both approximately even strength and approximately rated 2700. "alexei s." was one of the people who lost 50 points to young "magnus c." putting his rating to 2650. "alexei s." and "vladimir k." play at a tournament and draw a bunch of games ;) this too would drop "vladimir k." (a 2700 rating strength) a number of points even though he had never played "magnus c." but merely someone who had played him on his reign of terror ;)
|Jul-04-06|| ||AdrianP: There is a ratings calculator here:
Just doing a bit of fiddling around it seems to me that the formula is symmetric - e.g. if A plays B and A wins his rating goes up by exactly the same amount as B's goes down (there is also development coefficient which would probably produce asymmetric results, but I imagine this coefficient would be the same for all the participants in this tournament). So it remains a mystery to me, as to <plang>, why the amount of rating points won is not the same as those lost.
I had a look here http://www.chessbase.com/newsdetail... (Wijk Aan Zee 2006 results) and, to my calculator, the performance ratings almost exactly balance (there is 2 point difference over say 500 points -ve/+ve) which is likely to be a rounding error?
I don't fully understand your (<refutor>) explanation but that doesn't mean you're not right...!
Of course, this does not mean that there is not rating inflation - having more rated players would be enough to give rating inflation, I would have thought.
|Jul-04-06|| ||plang: I think what you are saying is that when a player plays someone whose rating doesn't reflect his actual strength (such as a rapidly improving junior) that distortions are created. Are you saying that there is something in the ratings formula to account for that? It seems like this phenomena could work the other way as well; say with an older player whose skills are deteriorating. As a matter of fact, as new talented players enter the system other older players whose skills are decreasing are leaving it. Maybe not in equal numbers but I don't know whether any distortion is being caused. It still seems to me that if more points are added to a rating for a win than subtracted for a loss that that will lead to an increase in overall ratings (inflation). On the July, 1985 ratings list there were only two players rated higher than 2640; Karpov at 2720 and Kasparov at 2700. On the July, 2006 list there are 66 players rated higher than 2640 and 13 rated higher than 2720. While there is an influx of new players (from China and India in particular) I do not believe that this explains the increase in ratings.|
|Jul-07-06|| ||harifan: <hitman84: <Harikrishna -12.8> OUCH!> Yup, it was a painful tournament for Hari. I have no doubts in my mind that he will be bounce back.|
|Jul-07-06|| ||themadhair: Go to (http://en.wikipedia.org/wiki/ELO_ra...) and view the section on mathematical details - that should help explain the descrepency.|
As for the FIDE online calculator my advice is avoid it and calculate the rating changes by hand. Consider the rating perfromance calculator (http://www.fide.com/ratings/calcula...) Consider these results it gives - Suppose you played opposition with an average rating of 2000 and score 3/4. FIDE calculator says your rating performance is 2025. Now suppose you played the same opponents in a longer tournament and scored 72/96. You should have the same rating performance (fraction is still 3/4) but the FIDE calculator gives a rating perofrmance of 2600.
The moral of the story? Don't rely on the FIDE calculators.
|Sep-24-06|| ||samsal27: <I have no doubts in my mind that he will be bounce back.> He did bounce back with victories in the Max Gyorgy Memorial and the World 960 Championships in Mainz. Way to go Hari!|
< Earlier Kibitzing · PAGE 26 OF 26 ·