chessgames.com
Members · Prefs · Laboratory · Collections · Openings · Endgames · Sacrifices · History · Search Kibitzing · Kibitzer's Café · Chessforums · Tournament Index · Players · Kibitzing

🏆 Tata Steel Masters (2016)

Chessgames.com Chess Event Description
... [more]

Player: David Navara

 page 1 of 1; 13 games  PGN Download 
Game  ResultMoves YearEvent/LocaleOpening
1. Navara vs Carlsen ½-½312016Tata Steel MastersD37 Queen's Gambit Declined
2. Adams vs Navara  ½-½312016Tata Steel MastersB12 Caro-Kann Defense
3. Navara vs Giri ½-½422016Tata Steel MastersD97 Grunfeld, Russian
4. Yifan Hou vs Navara 1-0402016Tata Steel MastersB12 Caro-Kann Defense
5. Navara vs So ½-½462016Tata Steel MastersD38 Queen's Gambit Declined, Ragozin Variation
6. Ding Liren vs Navara ½-½412016Tata Steel MastersE60 King's Indian Defense
7. Navara vs Karjakin  ½-½352016Tata Steel MastersD45 Queen's Gambit Declined Semi-Slav
8. Navara vs Caruana 1-0552016Tata Steel MastersE32 Nimzo-Indian, Classical
9. Wei Yi vs Navara 1-0282016Tata Steel MastersC65 Ruy Lopez, Berlin Defense
10. Navara vs Mamedyarov ½-½312016Tata Steel MastersC76 Ruy Lopez, Modern Steinitz Defense, Fianchetto Variation
11. Van Wely vs Navara ½-½372016Tata Steel MastersA04 Reti Opening
12. Navara vs Tomashevsky  ½-½322016Tata Steel MastersE15 Queen's Indian
13. Eljanov vs Navara 1-0362016Tata Steel MastersE12 Queen's Indian
  REFINE SEARCH:   White wins (1-0) | Black wins (0-1) | Draws (1/2-1/2) | Navara wins | Navara loses  

Kibitzer's Corner
< Earlier Kibitzing  · PAGE 66 OF 67 ·  Later Kibitzing>
Feb-16-16
Premium Chessgames Member
  AylerKupp: <<Sally Simpson> For instance using my added 5 pts. a number I plucked from the air. (The Mannerheim/Schuller catch) Carlsen by my count won 12 games in classical chess with Black in 2015.> (part 1 of 2)

Going back to your Carlsen example. The ChessTempo database has Carlsen playing 125 games in 2015, 80 at classical time controls. His results at classical time control as White were 21 wins, 14 draws, and 5 losses; and 11 wins, 23 draws, and 6 losses as Black. And, again per the ChessTempo database, he was rated 2862 for his first game of 2015 on Jan-10-15, A Giri vs Carlsen, 2015, and was rated 2834 prior to his last game of 2015 on Dec-29-15, Carlsen vs Kramnik, 2015. That draw against Kramnik caused him to lose 1 rating point so his Elo rating change for 2015 was (2834 -1) 2862 = -29 points.

Now, if you want to <add> to Carlsen's rating the number of points he should have won by playing Black and adjusting for color, you would also have to <subtract> from Carlsen's rating the number of points he shouldn't have won by playing White and adjusting for color, because in that case the color-adjusted rating differential between Carlsen's rating and his opponent's rating would have been effectively less.

And don't forget draws. They would be similarly affected, although to a lesser extent, since a lower rated player with the White pieces should have greater drawing chances against a stronger rated player with the Black pieces than if the colors were reversed and vice versa.

Let's look at the actual numbers. Of the 80 classical time control games Carlsen played in 2015, he conveniently had White for 40 of those and Black for the other 40. When he had White he won 21 games, drew 14, and lost 5. When he had Black he won 11 games, drew 23, and lost 6. To reduce the effort (even I get tired of all this work sometimes), I used the Kosteniuk Elo calculator's feature of calculating Elo rating changes using average ratings. I calculated Carlsen's average Elo when he won, drew, or lost with the White pieces and when he won, drew, or lost with the Black pieces. Similarly, I calculated his opponent's average rating when they won, drew, or lost with both the White and the Black pieces. The ChessTempo database is particularly useful in this exercise because they are very meticulous in recording the ratings of both players. <chessgames.com>, take notice.

I got the following Elo rating changes for Carlsen under the following conditions:

Carlsen has White, Carlsen wins +73.42
Carlsen has White, Carlsen draws -23.20
Carlsen has White, Carlsen loses -33.73

Carlsen has Black, Carlsen wins +36.22
Carlsen has Black, Carlsen draws -29.19
Carlsen has Black, Carlsen loses -40.55

Total Elo rating change -17.03

Feb-16-16
Premium Chessgames Member
  AylerKupp: <<Sally Simpson> For instance using my added 5 pts. a number I plucked from the air. (The Mannerheim/Schuller catch) Carlsen by my count won 12 games in classical chess with Black in 2015.> (part 2 of 2)

This is fairly different than the actual -29 Elo rating change when each game was rated individually rather than using averages and I'll attribute the difference to the inaccuracies of averaging or some other non-obvious non-linearities but I am not really sure. But note that the ratio is -29 / -17.03 = 1.7. Speaking of pulling numbers out of the air or from parts of our anatomy.

Then I did the color rating adjustment as I indicated above, <subtracting> 36 points from the rating of the player with the White pieces and <adding> 36 points to the rating of the player with the Black pieces. I got the following Elo rating changes for Carlsen:

Carlsen has White, Carlsen wins +93.76
Carlsen has White, Carlsen draws -9.82
Carlsen has White, Carlsen loses -28.99

Carlsen has Black, Carlsen wins +5134
Carlsen has Black, Carlsen draws -49.90
Carlsen has Black, Carlsen loses -45.59

Total Elo rating change +10.80 x 1.7 = +18

So, if we had adjusted for color, instead of <losing> 29 rating points in 2015, Carlsen would have <gained> 18 rating points.

But, as <frogbert> indicated, over a large number of games ("large" is the operative word) the number of Whites and Blacks that a player has will even out, so the effect of adjusting the result for color should not be significant.

Therefore I think that a rating adjustment for color will only be significant when dealing with a small number of games. And remember that Dr. Elo was only interested in calculating ratings for a large number of games, so I wouldn't consider it a "flaw" any more than using a continuous probability distribution when calculating rating changes based on a discrete number of games. Still, since the "calculation" is trivial, it might be worthwhile to do. But we need to check it first to see what changes it causes as a function of the number of games considered.

If you or anyone else is interested in looking at ChessTempo's database of 2015 Carlsen game and the spreadsheet calculation for color corrections, you can download them from http://www.mediafire.com/view/a6504....

Feb-16-16
Premium Chessgames Member
  AylerKupp: <<plang> An obvious example of rating inflation is when more points are added than are subtracted.>

Yes, that's a good example that I had not thought of. I was hoping for, but not really expecting, <tuttifrutty>'s definition of rating inflation. The informal definition that I use, imprecise and subjective as it is, uses the concept of "intrinsic strength" to determine if inflation is taking place.

We all have a concept of "intrinsic strength". How strong is player A really? If player A's intrinsic strength has not changed and his rating goes up, then the increase in his rating (if we're using rating as a measure of playing strength) is not really warranted, so rating inflation is taking place. If player A's intrinsic strength is increasing, then if an increase in his rating is also taking place, that increase is justified and rating inflation is not taking place.

The problem, of course, is that we don't have a method for measuring and calculating intrinsic (or absolute) playing strength, all we have are methods for calculating relative playing strength, so my informal definition is pretty useless.

Feb-16-16  Sally Simpson: Hi AylerKupp,

So it was 11 wins not 12. My bad, it's sorting out which games were and were not blitz games. I counted one in.

You have made things far too complicated.

I only want 5pts extra for a Black win.

No bonus for a draw.

White losses points as per the current system and no more. White does not lose an extra 5 pts. This would have White's playing too timidly. The object of the exercise is to encourage Black to have more ambition and not be happy with a draw.

So keep everything as it is. (warts and all) but if a player with the Black pieces wins he gets an extra 5pts.

The flaw is grades will go through the roof very quickly.

I was just suggesting a way to award a Black win through ELO without White suffering any extra penalty grading wise.

Of course if FIDE did accept my idea then through time some happy chappie will reach eventually 10000. The ELO database is programmed to hold only 4 digits and it will either crash and wipe out everything or start going backwards.

Feb-16-16  Bobwhoosta: Now that we've got that all wrapped up: Who's the greatest player of all time?
Feb-16-16  Sally Simpson: HI Bob,

Check the ELO grading list.

Feb-16-16  BOSTER: Marasmus is getting stronger.
Feb-16-16  tuttifrutty: There is no definition of elo rating inflation therefore there is none, it's all but thin air, a mumbo jumbo calculations, it's a product of malfunctioning brain and I don't want to know what the heck you all are smoking.
Feb-16-16
Premium Chessgames Member
  AylerKupp: <<Sally Simpson> You have made things far too complicated.>

Thank you, I'll take that as a compliment. I take perverse pride in making even the simplest things complicated. :-)

But I think that Dr. Elo was trying to develop a system that was based on sound mathematical principles. And symmetry suggests that if Black is to be given a bonus for their performance in the face of White's advantage, then White should be given an equivalent demerit to compensate for its inherent advantage.

As far as difficulty is concerned, all I'm suggesting is that White's rating for the purpose of calculating their change in rating following a game to be reduced by 36 point to reflect the historically accurate measure of White's advantage, and that Black's rating be increased by 36 point for the same reason. After these "effective ratings" are calculated (one subtraction and one addition), everything is done the same way as it is being done today. One addition and one subtraction, everything else staying the same, doesn't seem all that complicated to me.

And your 5 points to be added to Black's Elo in case of a win, is somewhat arbitrary and asymmetrical, and doesn't address draws. In the case of a draw shouldn't Black's Elo be increased by 2.5 points (or 2 or 3, depending on how you feel about draws)? That only seems fair.

But you're right about what might happen if a (very) happy chappie's rating reaches 10,000. Kind of the Y2K issue of 16+ years ago. Maybe we should start referring to this as the R10K problem. Luckily, I think that we have an adequate amount of time to address the "problem".

Feb-16-16
Premium Chessgames Member
  AylerKupp: <<Bobwhoosta> Now that we've got that all wrapped up: Who's the greatest player of all time?>

Like the story of a businessman searching for an accountant and asking each of their accountants applicants "How much is 2 + 2?" The accountant who was hired responded: "How much do you want it to be?"

So the answer to your question is: "Who do you want the greatest player of all time to be?"

Feb-16-16
Premium Chessgames Member
  AylerKupp: <tuttifrutty> I haven't been able to find a generally agreed upon definition of elo rating inflation either . But you were quite adamant (Tata Steel (2016) (kibitz #1661)) that there was no rating inflation. How can you categorically deny that there is no rating inflation unless you have at least a personal concept of what rating inflation is? After all, just because there isn't a definition of "something" doesn't mean that this "something" doesn't exist.

I told you what my flawed concept of "rating inflation" is. Why won't you tell us yours? It's personal after all, I'm certainly not going to challenge it.

Feb-16-16  frogbert: <plang: An obvious example of rating inflation is when more points are added than are subtracted.>

This is what I define as systemic rating inflation, plang. Or deflation, if the opposite is happening. And we need to consider the entire pool of players, although it's possible to say something about what's happening to certain sub sets of the pool as well.

Like I've posted about already years ago now, the highest rated players are on average losing rating points - quite contrary to what people seem to think. From one list to the next, usually we find that the sub set of all players rated 2700+ - and similarly the set of all players rated 2600+ (which includes the first sub set) - are losing rating points on average, or as a group.

Now, is that reasonable? Well, it is. It's just another example of what's called regression towards the mean. In order to have gotten as highly rated as they are, several of the 2700s have technically overperformed; they have achieved results that they can't maintain in the long run.

Still, being able to demonstrate systemic inflation in itself isn't enough to claim the kind of inflation people usually talk about. But since ratings aren't designed to map chess skills to rating numbers consistently over time, I don't care too much about that other discussion, which mostly rests on a misunderstanding of Elo-based rating systems.

Feb-17-16  Keyser Soze: <AylerKupp: <tuttifrutty> I haven't been able to find a generally agreed upon definition of elo rating inflation either . But you were quite adaman>

For him Elo rating mesures the volume of activity (feeding frenzy) of pampered goldfishes inside of an Aquarium. Therefore it doesn't apply for the Barracudas.

Feb-17-16
Premium Chessgames Member
  perfidious: <plang: An obvious example of rating inflation is when more points are added than are subtracted. In the US there are some examples of rating floors. I believe a Life Masters rating can't fall below 2200 so if you defeat him your rating goes up but his doesn't go down.>

Correct--if still active, I would be one such player.

Feb-17-16  Absentee: <AylerKupp: I told you what my flawed concept of "rating inflation" is. Why won't you tell us yours? It's personal after all, I'm certainly not going to challenge it.>

You're talking to a tolengoy sock.

Feb-17-16  EvigOptimist: AylerKupp - You must have made an error in your spreadsheet, because with your calculations Magnus won more points with black AND white for those games that he won.

Carlsen has White, Carlsen wins +73.42
Carlsen has Black, Carlsen wins +36.22

compared to

Carlsen has White, Carlsen wins +93.76
Carlsen has Black, Carlsen wins +51,34

In the spreadsheet, it looks like you have subtracted points for white instead of adding it. Its easier to win with white, so you have to add points to reflect that.

I do agree that in a perfect system you should do such adjustments, but over a year I guess it will be under 5 point difference, maybe less, so I dont think its worth changing. And for lower rated players, the difference between white and black are less than for the top players.

Feb-17-16  Sally Simpson: Hi AylerKupp,

Again no bonus for a draw. That would only drive both players into a deeper shell and probably increase the number of draws.

The bonus only applies to a Black win.

It's not unfair because as has been stated the number of Whites and Black a player has over a year is close to 50-50.

Just wanted a wee cherry to dangle in front of the Black player. Make them sharpening up their repertoire. Give them the Larsen, Fischer, Kasparov attitude.

Everyone says there are too many draws, it's a possible method to combat it.

But of course grades will go bonkers.

If some lad wins 5 Black's (a 25 pt bonus.) and loses a game as White his opponent will feed off those extra 25 pts plus their own 5 pt bonus.

"..and everyone is getting fat 'cept Mama Cass."

Feb-17-16
Premium Chessgames Member
  perfidious: <Sally S...."..and everyone is getting fat 'cept Mama Cass.">

<Creeque Alley> is a droll account of the group's early trials and tribulations.

Feb-17-16  Everyone: <"..and everyone is getting fat 'cept Mama Cass."> No, yo mama is so fat ....
Feb-17-16
Premium Chessgames Member
  AylerKupp: <<Keyser Soze> For him Elo rating measures the volume of activity (feeding frenzy) of pampered goldfishes inside of an Aquarium. Therefore it doesn't apply for the Barracudas.>

Thank you. I had considered most of the items on your list but I had forgotten about the barracudas.

Feb-17-16
Premium Chessgames Member
  AylerKupp: <<Absentee> You're talking to a tolengoy sock.>

Oh, I'm quite aware of that, they're quite easy to spot. I was baiting him since he (once again) dug a hole for himself and I was trying to see how much deeper he could get into it. I don't expect to hear from him again but, if he does respond, I look forward to being amused by another banality.

I suggested to <chessgames.com> that, in an effort to reduce the number of sock puppets, that they have a middle membership account between Regular members and Premium members. A Regular member account would be free and able to read what was posted but not post. A Posting member account, for a very nominal yearly fee (say US $ 5.00), would be able to post in addition to read. And Premium member accounts would have the same privileges as they have today.

This would not, of course, discourage determined and well-heeled sock puppets but it would at least give pause to those who indiscriminately create accounts for the purpose of trolling. Unfortunately (I think) <chessgames.com> did not agree with the concept. Oh well.

Feb-17-16
Premium Chessgames Member
  AylerKupp: <Sally Simpson> Well, I understand your bias against draws but to be fair, achieving a draw with the Black pieces against a higher rated opponent is an achievement and should be rewarded just like achieving a win with the Black pieces against a higher rated opponent, except obviously not as much. And if there is a 300 point rating differential between opponents and the much higher rated opponent has the Black pieces, then defeating the much lower rated opponent is not that much of an achievement, and I don't think that it deserves a bonus. But then again, De Gustibus Non Disputandum Est.

Perhaps draws could be discouraged if the 3-1-0 scoring system was more widely accepted. Do you know of any databases that compare the percentage of draws in tournaments with the 3-1-0 scoring system with the percentage of tournaments with the 1--0 scoring system? I don't remember any database where the scoring system in use for a tournament was one of the database fields. It would be interesting to know the difference, if any, between the two.

Feb-17-16  frogbert: <AylerKupp> You present your theory about rating inflation as a hypthesis. What I wonder is what it takes to falsify your hypothesis. And then I think of your specific hypothesis of what has driven the suggested inflation.

Since you're using a definition of inflation <that can not be proven> - simply because there's no agreed upon way to measure "intrinsic strength" - don't you think that you also have some bootstrapping issue? In short you have a hypothesis to explain some phenomenon that you can't show exists, given your definition of inflation. To me this would've been a slight problem... :)

I guess you see why I've taken the approach I did.

PS! A final comment: testing whether Elo-based ratings do or do not map "intrinsic playing strength" appears a slightly strange endeavour, since Elo-based systems make no claim to being capable of making such a mapping. Elo-systems measure relative success within a given pool of players.

Feb-17-16  frogbert: If you feel like moving this discussion to your forum, that's perfectly fine, of course. :)
Feb-17-16  morfishine: <frogbert: If you feel like moving this discussion to your forum, that's perfectly fine, of course. :)> Nice discussion, but <AlyerKupp> rarely digresses or adheres to suggestions like this where elongated or massive or continuously humongous posts which are off-topic or out-of-place or simply in the wrong forum and needs to be elsewhere and clearly needs to be placed in there appropriate location

Good luck with that one

*****

Jump to page #    (enter # from 1 to 67)
search thread:   
< Earlier Kibitzing  · PAGE 66 OF 67 ·  Later Kibitzing>

NOTE: Create an account today to post replies and access other powerful features which are available only to registered users. Becoming a member is free, anonymous, and takes less than 1 minute! If you already have a username, then simply login login under your username now to join the discussion.

Please observe our posting guidelines:

  1. No obscene, racist, sexist, or profane language.
  2. No spamming, advertising, duplicate, or gibberish posts.
  3. No vitriolic or systematic personal attacks against other members.
  4. Nothing in violation of United States law.
  5. No cyberstalking or malicious posting of negative or private information (doxing/doxxing) of members.
  6. No trolling.
  7. The use of "sock puppet" accounts to circumvent disciplinary action taken by moderators, create a false impression of consensus or support, or stage conversations, is prohibited.

Please try to maintain a semblance of civility at all times.

Blow the Whistle

See something that violates our rules? Blow the whistle and inform a moderator.


NOTE: Please keep all discussion on-topic. This forum is for this specific tournament only. To discuss chess or this site in general, visit the Kibitzer's Café.

Messages posted by Chessgames members do not necessarily represent the views of Chessgames.com, its employees, or sponsors.
All moderator actions taken are ultimately at the sole discretion of the administration.

Spot an error? Please suggest your correction and help us eliminate database mistakes!

Copyright 2001-2021, Chessgames Services LLC