| page 1 of 57; games 1-25 of 1,420
| page 1 of 57; games 1-25 of 1,420
< Earlier Kibitzing · PAGE 8 OF 8 ·
|Dec-31-17|| ||AylerKupp: <<nok> If Magnus really wants to make history, his best move is to play the Candidates as rating qualifier.>|
I don't think that's what you literally meant. I suspect that what you meant was the world champion participating in either a big tournament (think Curacao 1962, a quadruple round robin) or a series of knockout matches <of reasonable length>, longer than the 4-, 6-, or (maybe) 8-game matches held in the FIDE World Championship Knockout Tournaments of 1999, 2000, 2001, and 2004.
But it has nothing to do with Carlsen. It's FIDE that determines the format for determining who the WCC is going to be. While Carlsen or whoever the WCC is can certainly suggest what they would like and even threaten not to participate if they consider the format unacceptable, even Fischer could not force FIDE to conform to a specific format and set of rules or regulations.
And, BTW, I agree with what I think you meant. All that a WCC match or tournament determines, or tries to determine, is who the best chess player in the world is <at that time>. And just because a certain player is the best on any given year and is crowned WCC, it doesn't necessarily mean that they are the best two or three years from that time. So, like most major sports, I think that the reigning WCC should be included in the list of players competing for the WCC title.
Then it depends on what consideration, if any, the defending champion should be given. If the WCC title is to be decided in a big tournament or a series of knockout matches, then perhaps the defending champion would be automatically seeded into this event, just like the loser in the previous WCC match is currently. After all, all the other players must win or place in one of several events to qualify for this final tournament/set of knockout matches so automatically qualifying the defending WCC might be sufficient to consider that tradition has been upheld.
|Dec-31-17|| ||frogbert: <Even if he does poorly in those end of the year tournaments and his rating drops drastically, his average rating does not drop as much because he had the high rating at the beginning of the year.>|
I didn't suggest using a rating average, actually. I said (and meant) simply the highest rated player.
<without needing to introduce arbitrary thresholds for the maximum number of games allowed to be played before a player's rating is re-calculated. It seems to me that it would be a recipe for confusion if a player's (or several players') rating were to change in the middle of an event>
I certainly didn't mean to imply that ratings should be recalculated during events - rather that ratings should be «officially recalculated» *after* the event if a player had reached the specified number of games during said event.
Hence, a 9-round tournament wouldn't necessarily lead to recalculation, but if another 12 games were played in a different event, recalculation would be needed (with a threshold of 20).
Still, I'm a believer in the KISS principle, so recalculating after each event is probably the wiser choice.
Regarding the use of higher Ks like 15 or 20, it's mostly about how one wants to weight older versus more recent performances. But I agree - and I also wrote - that higher Ks lead to more volatility in the system. Maybe adapting something similar to Glicko would be a better solution. That adds a risk to «preserving one's rating by not playing», since your rating will change more quickly (same effect as using a higher K) when you resume playing after a period of inactivity.
|Jan-01-18|| ||Sokrates: Thanks for the explanations, <AylerKupp>. I have wondered why particularly Kramnik has been able to maintain his high ratings even though his results often have been on average in tournaments. It seems that the systems favours you, once you have achieved a high rating?|
|Jan-01-18|| ||rogge: Anand, the humorist :)
|Jan-01-18|| ||frogbert: <It seems that the systems favours you, once you have achieved a high rating?>|
Actually, I see nothing wrong in that, except for the ill-adviced choice of using rating average in the sense of averaging the 12 official, monthly ratings during a year (whether a player was active or not). In general, having achieved a high rating is a witness about consistent skill and previous accomplishments and should be honoured.
Anyway, there's a really simple fix to the averaging algorithm: 1) Weight each rating with the number of games played in the rating period, 2) only count periods where the player has actually played games (for January, if no games were played in December, use the number of games of the last active period - e.g. games played for the November list if the player played rated games in October). And of course, you need a minimum number of games played in the rating year of interest.
1) Previous accomplishments are somewhat honoured - which I think is right. (Unlike when using a yearly PR.)
2) It's basically the games played throughout the year that count - and a positive trend is rewarded, a negative one is punished
3) One maintains a certain «king of the hill»-notion, because you actually have to climb up the ladder and consistently prove yourself: two extraordinary events in a row wouldn't be enough.
In fact, a bit of the same reasoning (1 & 3) goes for a suggestion of naming the #1-rated player at the end of each year the World Champion. :)
9 games in December of Y-1
Jan 2800 - 9 games
Feb 2800 0
Mar 2800 0
Apr 2800 0
May 2804 2
Jun 2804 0
Jul 2804 0
Aug 2790 7
Sep 2780 14
Oct 2780 0
Nov 2780 0
Dec 2773 13
Weighted avg: (2800x9 + 2804x2 + 2790x7 + 2780x14 + 2773x13) / (9 + 36) = 2784,6
Compare to FIDE's «model» where the avg becomes: (2800x4 + 2804x3 + 2790 + 2780x3 + 2773) / 12 = 2792,9
|Jan-01-18|| ||moronovich: <rogge: Anand, the humorist :)
Lol! What a picture.
More tiger than madras.
Happy New Year <rogge> :)
|Jan-01-18|| ||rogge: Happy New Year, may the Force be with you :)|
|Jan-01-18|| ||frogbert: Happy new year to both of you, <rogge> and <moronovich>!|
|Jan-01-18|| ||rogge: U2, <frog> :)|
|Jan-01-18|| ||LameJokes55: Happy new year everyone!
This refers to comparison of rating systems (Classical and blitz).
Blitz follows k20 pattern while classical that of k10. This could be a small reason for fluctuation. However, I believe, there are two more reasons.
First one is result-orientedness. Classical as a genre is far more draw-oriented. Once a player reaches his peak rating. He tends to stagnate around this rating; give or take few rating points. Blitz, by contrast, is far more result-oriented.
The second one is luck. In classical, higher-rated and better player, on an average, wins more often. Thus, maintains his rating. Chess is game of mind. Classical genre affords time for the mind to think.
In blitz, there is very little time to think. Luck factor plays a greater role. Better player could bleed rating points and inferior one could gain a windfall.
|Jan-01-18|| ||moronovich: Happy New Year to you too <frogbert> !|
|Jan-01-18|| ||AylerKupp: If the issue is the concern about extreme fluctuations in the Rapid and Blitz ratings (and what constitutes "extreme" is, of course, subjective), perhaps the simplest correction is to reduce the K-factor for Rapid and Blitz below 10, say, 5. Given the unavoidable higher fluctuation of ratings at faster time controls (the shorter the time control, the greater the likelihood of mistakes) using K=20 in Rapid and Blitz is the equivalent of fighting fire with gasoline.|
If K=5 had been used in Grischuk's case, the 79 point decline in his Blitz rating in Nov-2017 would have been 1/4 of that, or 20 points. And his 155 point gain in his Blitz rating in Dec-2017 would have been less than 1/4 of that, 1/4 (~ 39 points) because of the K-factor reduction and less than the ~ 39 point gain because his rating would have been higher to begin with, and therefore the difference between his ratings and those rated higher than he was would have been less.
After all, the K-factor is semi-arbitrary. It is selected to give greater weight to recent performances and lesser weight to earlier performances. It is chosen to try to maintain each rating change to ~ 1.5 sigma, but this 1.5 is also semi-arbitrarily chosen.
Changes to the K-factor for Classic time control games from K=10 to K=20 was extensively discussed in 2009. In the end, FIDE decided to leave it alone.
In order to determine whether K=5 is right number or whether another value of the K-factor would be more desirable, it would be worthwhile to calculate the average of the maximum and minimum rating fluctuations at Classic time controls for a period of one or two years (or possibly longer). Then the K-factor for Rapid and Blitz games (they need not be the same) for the same time period could be tweaked until the average of the maximum and minimum rating fluctuations at these time controls is similar to that for games at Classic time controls. Since the ratings fluctuation at Classic time controls have been considered reasonable in the sense that there doesn't seem to have been much, if any, concern about it, whatever the rating fluctuation was for Classic time controls should be considered acceptable for Rapid and Blitz time controls.
|Jan-01-18|| ||AylerKupp: <<frogbert> Anyway, there's a really simple fix to the averaging algorithm: 1) Weight each rating with the number of games played in the rating period, 2) only count periods where the player has actually played games (for January, if no games were played in December, use the number of games of the last active period - e.g. games played for the November list if the player played rated games in October).>|
An interesting idea, one which I had not considered. Since I already had a spreadsheet used to calculate the rating averages with all the needed data (rating for each month, number of games played), I modified it to reflect your suggestion, with one change to make the calculations easier which I think is minor and which I don't think would change the results significantly.
First, I counted only the ratings for those months where the player was active, without weighting the ratings by the number of games played. But I didn't see the reason if the player was not active in December of the previous year for using for January the number of games for the player's last active period; I simply considered the player not active for the previous December and therefore did not consider the player's rating from the January list in the calculations.
I then used the weighted ratings as you suggested for each month where the player was active (again not including the rating in January if the player was not active the previous December).
My spreadsheet covered the top 30 ranked players but, realistically, only the top 10 have much of a chance to qualify for the Candidates Tournament on the basis or rating. So here are the results for the top 10 players; (Average FIDE monthly ratings, <resulting rank>), (Unweighted average FIDE monthly rating when active, <resulting rank), and (Weighted average FIDE monthly rating when active, <resulting rank) along with , of course, the players' names:
(2832.42 <1>) (2829.00 <1>) (2829.89 <1>) Carlsen, Magnus
(2808.58 <2>) (2809.17 <2>) (2809.75 <2>) Caruana, Fabiano
(2806.42 <3>) (2809.17 <2>) (2809.16 <3>) So, Wesley
(2804.08 <4>) (2805.60 <4>) (2806.73 <4>) Kramnik, Vladimir
(2796.00 <5>) (2795.30 <5>) (2796.16 <5>) Vachier-Lagrave, Maxime
(2792.67 <6>) (2794.36 <6>) (2795.67 <6>) Aronian, Levon
(2785.92 <7>) (2789.14 <7>) (2793.56 <7>) Mamedyarov, Shakhriyar
(2785.58 <8>) (2784.11 <9>) (2785.48 <9>) Nakamura, Hikaru
(2785.25 <9>) (2786.40 <8>) (2786.94 <8>) Anand, Viswanathan
(2775.75 <10>) (2776.40 <10>) (2778.30 <10>) Karjakin, Sergey
As you can see, there is not much variation in their average rankings regardless how they are calculated and, most importantly, almost none in the rankings. Compared to the rankings using the average FIDE rating, using the unweighted average FIDE rating when active resulted in Caruana and So both being ranked #2 instead of #2 and #3, and the rankings of Nakamura and Anand were reversed. And using the weighted average FIDE rating when active gave the same results, except that using the weighted average caused Caruana to be ranked #2 and So #3, just like using the average FIDE ratings for each month.
So I don't think that only considering the months when the player was active, whether weighted by the number of games played or not, would yield significantly different rankings. Or at least it didn't in 2017.
|Jan-01-18|| ||AylerKupp: <frogbert> With respect to your other opinions:|
1. I don't attach much importance to honoring previous accomplishments. Whatever a player accomplished prior to the previous WCC cycle is nice, but I think that only his/her accomplishments since that time are relevant to the next WCC cycle.
2. I agree that a positive trend should be rewarded and a negative (or maybe non-negative) trend should be punished. But I don't think that your suggestion does that. Maybe by weighting Jan's results by 1/12, Feb's results by 2/12, etc. that could be accomplished, but I'm not sure how to calculate "fair" weights; i.e. not "too little" weight to early months and not "too much" weight in later months. At least philosophically.
3. I think that the only ladder climbing that is of any relevance is that done in the time period since the previous WCC cycle. What happened prior to that may be interesting, but not relevant.
And I favor treating the defending champion like any other contender for the title. He could try to qualify for the Candidates Tournament by participate in the qualifying tournaments as Carlsen did in the World Cup in 2017 or attempt to qualify by <performance> rating in the years since the previous WCC cycle like Caruana and So did in 2017 by average rating.
These, of course, are my opinions and nothing more.
|Jan-01-18|| ||frogbert: <I agree that a positive trend should be rewarded and a negative (or maybe non-negative) trend should be punished. But I don't think that your suggestion does that>|
It does, when you have a positive trend, play more games and defend your previous/current level. I can provide a couple examples tomorrow, if you like. :)
|Jan-01-18|| ||tuttifrutty: <These, of course, are my opinions and nothing more.>|
I agree, they are just mumbo jumbo which mean zilch...much like frogbert's long post is nothing but just that.
How many times do I have to penetrate and drill your pea size brains to tell you that elo is just a measurement of past performances (history), and the difference in the ratings between two players serves as a predictor of the outcome of a match.??? Geez...you boys need a brain surgeon to fix some loose wirings...
|Jan-02-18|| ||Rolfo: Happy New Year frogbert|
|Jan-02-18|| ||Sokrates: Happy New Year to all of you - even those on my ignore list! :-) I look forward to great chess in 2018 - in spite of FIDE.|
|Jan-02-18|| ||Rolfo: Happy New Year to all :)|
|Jan-02-18|| ||AylerKupp: <<tuttifrutty> How many times do I have to penetrate and drill your pea size brains to tell you that elo is just a measurement of past performances (history), and the difference in the ratings between two players serves as a predictor of the outcome of a match>|
You don't have to tell us that; we all know that and everyone else that reads our posts understands that we know that. It is, after all, pretty obvious. It's so obvious that it's implicit in all our discussions except that your zero-size brain seems incapable of understanding what we're taking about so you refer to the discussions as mumbo jumbo.
So, since you don't realize that we all know that, you keep repeating it over and over again. I suppose that you think that you need practice with your typing. In that case, keep up the good work.
And a Happy New Year to you. I look forward to more of your silly posts throughout the year and all the chuckles that they elicit from everyone.
|Jan-02-18|| ||john barleycorn: <tuttifrutty> is just confused that his hero W. So did not visit his CG player page for season wishes to his "truest" fans.|
|Jan-02-18|| ||tuttifrutty: <And a Happy New Year to you.>|
And Happy New Year to all including frogbert and those who got out of the hot kitchen...:-)
|Jan-02-18|| ||beatgiant: <tuttifrutty>
How would we summarize your beliefs in a nutshell? What I gathered so far is:
1. You think Carlsen is weaker than what we'd believe based on ratings.
2. You think So is stronger than what we'd believe based on ratings.
3. You think that playing in closed events with nobody but the strongest opponents makes it easier to maintain a high rating, compared to playing in open events filled with lower-ranked players. 3a. This is because you believe the high-ranked players are overrated and the lower-ranked players are underrated.
4. You think that beating a certain opponent for the first time is a better indicator of future performance than ratings.
5. You don't place much stock in the math behind predictive modeling.
Did I get all that right?
|Jan-06-18|| ||siamesedream: <<Magnus Carlsen's blog>|
While it is easier to perform consistently well in Rapid with more time available and less erratic play, I’ve had such a good year in Blitz that aiming for anything else than victory didn’t make sense initially.
Round one was weird, really weird. The game was interesting, and low on time my opponent Inarkiev made an illegal move (checking me while his king was in check). On reflex I moved my king out of check, reaching a legal position (where only his king was in check and it was his turn to move), instead of claiming a win by pointing out his illegal move. Next he stopped the clock and claimed a win arguing that I had no legal moves as his king was in check! Somehow the arbiter agreed to this preposterous idea and round 2 was delayed more than 30 minutes as I tried (and fortunately managed) to convince the chief arbiter that the game should continue in the final and fully legal position both based on the rules, and the 2014 Dubai World Blitz precedent where my trainer Peter Heine Nielsen had a similar experience!
Highly agitated I was thoroughly outplayed by Sjugirov in round 2. Three wins in row followed by a disastrous finish scoring 50% in the rest of the games, where I blundered badly in every game, left me trailing leader Karjakin by two points in 20th place.
In the morning on the last day I wasn’t as confident as usual, and more or less accepted that 1st might be out of reach. Against Grishuk with black in the first round I immediately noticed that my brain worked better. He played a topical Queens Gambit line and I decided to deviate from theory after his Qa4. He missed the best continuations, and I was slightly better with my rook and queen on the e-file. After Qc5+ he had to play Qd4 with compensation for the pawn. Moving his king instead he was simply a pawn down with a worse position, and the rest was a matter of technique. 0-1. Against Harikrishna I had a pleasant advantage and continued to put pressure on him throughout the game. I made a few inaccuracies and he defended well, and it was only short on time he blundered and lost. Quite a good game by both I think.
Next as black against Mamedyarov I managed to equalize and tried to tempt him into playing for a win in the queen and rook ending. He did, and when I counterattacked he blundered badly with Qd5 allowing Re1+!
Karjakin drew his first three games and with very few victories on the top boards I was back at the top board in round 4. He allowed an opening that has worked well for me in the past. My kingside attack consists mainly of natural moves while his defense is more difficult. He went wrong with Qf7 and his position collapsed quickly. I was suddenly sole leader!
I drew with white against Vachier-Lagrave in round 5 and managed to win the next four games by putting pressure on my opponents both on the board and on the clock to secure clear first with a round to go. When Aronian offered a draw with white in the last round I saw no reason to play on and try to interfere with the distribution of the other medals. It was great to feel once again that winning can be such an all-encompassing emotional experience. This is probably the Rapid or Blitz victory I’ve enjoyed the most!
Karjakin took 2nd and Anand 3rd after late game blunders by their opponents in the high-stake last round games.
We returned to Norway for New Years eve, and the Blitz victory is clearly a highlight of 2017 and a great prelude to the 2018 tournaments.
I’d like to wish all of you at Arctic Securities, your families, and your customers and contacts, a Happy New Year! I look forward to our continued (for the 9th year!) cooperation in 2018!
I’ll revert with a short summary of 2017 closer to my next classical event Tata Steel Chess starting on January 13th, but you might want to follow the final in the online chess.com Blitz Battle Championship Wednesday night. As last year, I face Nakamura in the final.
Magnus Carlsen, Oslo, January 2nd 2018 >
|Jan-06-18|| ||tuttifrutty: <Highly agitated I was thoroughly outplayed by Sjugirov in round 2.>|
Hmmm...what would be his excuse when he lost to the same player during olympic 2010 in 25 moves???? Could it be that Sjugirov is another Nepo and Bu??? You tell me...
<I immediately noticed that my brain worked better.>
Only if Grischuk didn't forget to take his daily dose of omega 3, it would have been a different story.:-)
< Earlier Kibitzing · PAGE 8 OF 8 ·
Spot an error? Please suggest your correction and help us eliminate database mistakes!
NOTE: You need to pick a username and password to post a reply.
Getting your account takes less than a minute, totally anonymous,
and 100% free--plus, it
entitles you to features otherwise unavailable.
Pick your username now and join the chessgames community!
If you already have an account, you should
Please observe our posting guidelines:
- No obscene, racist, sexist, or profane language.
- No spamming, advertising, or duplicating posts.
- No personal attacks against other members.
- Nothing in violation of United States law.
- No posting personal information of members.
See something that violates our rules? Blow the whistle and inform an administrator.
NOTE: Keep all discussion on the topic of this page.
This forum is for this specific tournament and nothing else. If you want to discuss chess in general, or
this site, you might try the Kibitzer's Café.
posted by Chessgames members do not necessarily represent the views of Chessgames.com, its employees, or sponsors.|
your profile |
Premium Membership |
Kibitzer's Café |
Biographer's Bistro |
new kibitzing |
Tournament Index |
Player Directory |
Notable Games |
World Chess Championships |
Opening Explorer |
Guess the Move |
Game Collections |
ChessBookie Game |
Chessgames Challenge |
privacy notice |
Copyright 2001-2018, Chessgames Services LLC