Is it a disaster if Anand wins the candidates tournament?

Sort:
Pacifique
Irontiger wrote:
Pacifique wrote:
Scottrf wrote:

Ratings are a very good predictor of future results in a long series of games with a single opponent. They were in the last world championship and they will be in the next. Much better than people's subjective opinions about who is more experienced, home advantage or other factors. As I said before the last match and the women's championship.

Kramnik had to lose against Topalov in their 2006 match, according to yout statement - he had only 2743 against Topalov`s 2813.

Anand should not have to crush Kramnik with such a margin in their 2008 match too - he had only ELO 11 points more  than Kramnik (2783 vs 2772).

Do you need more examples, which will refute your rating fetishism?

That's anecdotal evidence. Of course having a superior rating does not mean a sure win - it only means a win is more probable. A die landing on 5 would not refute a claim that "obtaining 4 or less is more likely".


That being said, there is a good argument that in WC matches both players prepare opening novelties, play different lines and end up in position we are not used to see them in, and that affects their relative strength. But that's not the point you made.

Do you (and other rating fetishists) have any statistical evidence of World championship tournament & match ratings to back up your claims?

Irontiger
Pacifique wrote:
Irontiger wrote:
Pacifique wrote:
Scottrf wrote:

Ratings are a very good predictor of future results in a long series of games with a single opponent. They were in the last world championship and they will be in the next. Much better than people's subjective opinions about who is more experienced, home advantage or other factors. As I said before the last match and the women's championship.

Kramnik had to lose against Topalov in their 2006 match, according to yout statement - he had only 2743 against Topalov`s 2813.

Anand should not have to crush Kramnik with such a margin in their 2008 match too - he had only ELO 11 points more  than Kramnik (2783 vs 2772).

Do you need more examples, which will refute your rating fetishism?

That's anecdotal evidence. Of course having a superior rating does not mean a sure win - it only means a win is more probable. A die landing on 5 would not refute a claim that "obtaining 4 or less is more likely".


That being said, there is a good argument that in WC matches both players prepare opening novelties, play different lines and end up in position we are not used to see them in, and that affects their relative strength. But that's not the point you made.

Do you (and other rating fetishists) have any statistical evidence of World championship tournament & match ratings to back up your claims?

I won't insult you into asking whether you really know how Elo ratings work, but I wonder.

I guess we agree ratings are a fairly good indicator of past performance ? Well then, if it was a bad indicator of future performance, we should see everyone's rating going up and down as tournaments with different results than "expected" come. Go to the FIDE rating list and you will see that is not the case.

Pacifique

In Anand - Topalov 2010 match Anand`s rating was 2787 & Topalov`s - 2805.

Before the match, most experts considered Anand as better (despite of numerous brainless Topalov fans), but aknowledged that venue of the match (the capital of Bulgaria) gives additional chances for Topalov.

Anand won.

Scottrf

What expected score is there from players 18 elo apart? I'll give you a hint, it's statistically insignificant.

Anand and Carlsen are 100 elo apart. Try harder. 

Pacifique
Irontiger wrote:
Pacifique wrote:
Irontiger wrote:
Pacifique wrote:
Scottrf wrote:

Ratings are a very good predictor of future results in a long series of games with a single opponent. They were in the last world championship and they will be in the next. Much better than people's subjective opinions about who is more experienced, home advantage or other factors. As I said before the last match and the women's championship.

Kramnik had to lose against Topalov in their 2006 match, according to yout statement - he had only 2743 against Topalov`s 2813.

Anand should not have to crush Kramnik with such a margin in their 2008 match too - he had only ELO 11 points more  than Kramnik (2783 vs 2772).

Do you need more examples, which will refute your rating fetishism?

That's anecdotal evidence. Of course having a superior rating does not mean a sure win - it only means a win is more probable. A die landing on 5 would not refute a claim that "obtaining 4 or less is more likely".


That being said, there is a good argument that in WC matches both players prepare opening novelties, play different lines and end up in position we are not used to see them in, and that affects their relative strength. But that's not the point you made.

Do you (and other rating fetishists) have any statistical evidence of World championship tournament & match ratings to back up your claims?

I won't insult you into asking whether you really know how Elo ratings work, but I wonder.

I guess we agree ratings are a fairly good indicator of past performance ? Well then, if it was a bad indicator of future performance, we should see everyone's rating going up and down as tournaments with different results than "expected" come. Go to the FIDE rating list and you will see that is not the case.

Doooh. In fact ratings are tend to go up and down. In many OTB tournaments I`ve took part #1 rated player did not took the 1st place.

Look at chess-results.com for example, if you don`t believe me.

Irontiger
Pacifique wrote:

Doooh. In fact ratings are tend to go up and down. In many OTB tournaments I`ve took part #1 rated player did not took the 1st place.

I never said they don't go up and down. Of course they do. But they don't swing by 200 points on short timescales.

Again, just look at anyone's rating chart at fide.com. If you can find such a large swing, I predict it will be either because said player is still progressing (eg Carlsen five years ago) or had a long inactivity time.

Aetheldred

Please find the part where GM Yermolinsky talks about Anand. If this guy can't make a proper prediction...

http://www.chessclub.com/article/fide-candidates-2014

Pacifique
Scottrf wrote:

What expected score is there from players 18 elo apart? I'll give you a hint, it's statistically insignificant.

Anand and Carlsen are 100 elo apart. Try harder. 

Didn`t you promise to stop discuss with me Mr.Liar? Look at the top 20 FIDE rating list. According to that list, players like Vachier-Lagrave and Dominguez are in the same league as Anand, Karjakin & Svidler.

Pacifique
Irontiger wrote:
Pacifique wrote:

Doooh. In fact ratings are tend to go up and down. In many OTB tournaments I`ve took part #1 rated player did not took the 1st place.

I never said they don't go up and down. Of course they do. But they don't swing by 200 points on short timescales.

Again, just look at anyone's rating chart at fide.com. If you can find such a large swing, I predict it will be either because said player is still progressing (eg Carlsen five years ago) or had a long inactivity time.

Do we have 200 points difference in Candidates tournament or World championship match? Your argument is beside the point.

Scottrf
Pacifique wrote:
Scottrf wrote:

What expected score is there from players 18 elo apart? I'll give you a hint, it's statistically insignificant.

Anand and Carlsen are 100 elo apart. Try harder. 

Didn`t you promise to stop discuss with me Mr.Liar?


No I didn't. You shouldn't call people liars if you're as bad at reading (and writing) as you are.

Pacifique
Scottrf wrote:
Pacifique wrote:
Scottrf wrote:

What expected score is there from players 18 elo apart? I'll give you a hint, it's statistically insignificant.

Anand and Carlsen are 100 elo apart. Try harder. 

Didn`t you promise to stop discuss with me Mr.Liar?


No I didn't. You shouldn't call people liars if you're as bad at reading (and writing) as you are.

So you have no arguments again.

blitzjoker

Going back a few hours, I think the 27 straight blacks would occur once in 255 years in a wheel running once a minute for 24 hours a day.

So that guy was quite unlucky really.

According to mathspig, the longest recorded run on an even chance is 26, but who knows.  The Martingale is an ok system if you stop at about 8 and don't mind losing a bit occasionally.  Sometimes you can get a round of drinks in anyway.

Polar_Bear
Irontiger wrote:

So what ? Ratings are a hoax ?

Yes, sort of. According to inventor Arpad Elo ratings shouldn't be used to estimate actual strength of players.

blitzjoker

What is all this ratings stuff?  Presumably most of us agree that a 2850 is more likely to beat a 2750 than the other way round, but that sometimes he won't?  Or is someone arguing that's not the case? I got lost somewhere along the way, but it doesn't seem very complicated.

Irontiger
blitzjoker wrote:

What is all this ratings stuff?  Presumably most of us agree that a 2850 is more likely to beat a 2750 than the other way round, but that sometimes he won't?  Or is someone arguing that's not the case?

I don't even know what other people think here.

Since you get trashed when you open your mouth here, I will try to keep it shut now.

jesterville

ELO was never meant to a "predictive" tool for Chess game outcomes. It was meant to discribe the relative playing strength among players. That being said, one can reasonably expect that where there is a great imbalance between the two players...that the stronger player will most often win/not loose. But where the diffence in the competing ELOs are only marginal...then the use of ELO as any "predictive" model, is as good as flipping a coin.

The end result of some sports are win/lost, such as Basketball, Tennis, Hockey etc. But chess also has a third outcome-draw , and thus more difficult to predict. 

Polar_Bear

Ratings (together with recent progress) could be used to estimate strength and/or predict near future with caution if they weren't skewed. Unfortunately, this is not the case. Players don't play in vacuum, they try to optimize their ratings too often. Almost everyone adjusts his style and immediate goals to opponent's rating and actual standings.

Shivsky
Aetheldred wrote:

Please find the part where GM Yermolinsky talks about Anand. If this guy can't make a proper prediction...

http://www.chessclub.com/article/fide-candidates-2014

Quite amusing to read right now. This excerpt in particular:

 I don’t even mention Anand in this context, because Anand will not win in Khanty. Vishy is finished and done for, stick a fork into him. It’s not the fear of Magnus – it’s the fear of playing chess that did Anand in, and it happened some 3-4 years ago.

As much as we'd like to think otherwise, Anand knows Anand better than we do. 

Elubas
Shivsky wrote:

Here's a game I've come up with:

Who can make the most angry Anand fan(anatic) forum post?

Here's how you score:

Play the racist card:  +5 Points
Grammar/Spelling/Punctuation mistakes: 1 Point each
Abbreviating multi-syllable words with just one letter:  +10 Points.
Fantasizing that Anand was simply caught "unaware" by Carlsen in Chennai: +20 points
Confusing Anand's truly "nice guy nature" with "his practical chances of beating Carlsen" + 25 points
Irony Bonus:  Showing none of the class, humility and good nature of the player you're rooting for by being a consummate jackass in your post.  +100 points.
 

 

Awesome post.

johnmusacha

I'm interested in knowing from what nation NM Pacifique got his NM title.  He is the only NM I've seen on this site of nine millions with an "international" designation.

Judging by his English skills, I don't think he is a US or Canadian NM.  Just wondering.