English Chess Forum

David Sedgwick wrote: ↑

Wed Jul 10, 2019 10:27 am

NickFaulks wrote: ↑

Wed Jul 10, 2019 7:45 am

If I may be forgiven for introducing some facts into the discussion, in the 2018/19 winter season Rausis played 26 games against 2350+ opposition. His results were

v 2350-2450 _ +8 =5 -0 for a performance of 2650

v 2451-2670 _ +3 =10 -0 for a performance of 2640

I suggest that he may not be expected to litter the landscape with free rating points in the way some people believe.

So Rausis is 2600 - 2650 strength. His rating is close to 2700.

I'm not sure where that 2600 number comes from.

There is no dispute that Rausis has recently gained around 20 points from games which do not statistically justify that outcome. The point I was making, quite clearly I think, was that claims that he avoids playing against serious opposition are not well informed.

Paul Cooksey wrote: ↑

Wed Jul 10, 2019 8:00 am

I'm not convinced. If Rausis was a strong player, surely he would have beaten Mamedyarov with black in the Bundesliga instead of conceding that draw.

Well, getting a good position against Shak in the first place is arguably quite compelling evidence that you are *pretty* strong

NickFaulks wrote: ↑

Wed Jul 10, 2019 11:54 am

I'm not sure where that 2600 number comes from.

https://ratings.fide.com/id.phtml?event=11600098.

I was giving Rausis the benefit of the doubt that his 2600+ ratings in 2017 and 2018 reflected a genuine increase in his playing strength compared with previously. I am disinclined to be so charitable about his most recent spike.

Roger >Women. All received an extra 100 points, except Zsuzsa Polgar.
That was political though as an anti Polgar measure. <

Incorrect. Women were underrated at the time. I was not then a member of the QC, otherwise I would have suggested commonsense. Decide on the number of points augmentation, say 120. Then, anybody who played only men would receive no extra points, anybody who played only women 120. It might have represented one day's work. In between, proportionately.
Instead they decided on 100 for everybody, except Szusza. Originally Sheila Jackson was offered I think 25. But Robert Bellin raised such a stink in Malta that she got the full 100. The reasoning had been that she, too, mainly played men. The political fact was that Szuza was no longer the highest rated woman in the world, it was the Women's World Champion.

So, when the next rating list came out, I investigated, expecting there to be a decline from the giddy heights of 100 extra points. I discounted Judit and a Russian woman who had gone up enormously. The average of the rest of them went UP. So much for the idea the women donated rating points to the men.

Stewart Reuben wrote: ↑

Thu Jul 11, 2019 1:26 am

The reasoning had been that she, too, mainly played men.

With the possible exception of exclusively Olympiad players, that would have applied to all the British women. There weren't exclusively female FIDE rated events. Once the British Ladies Championship was absorbed into the main Championship, I don't recall any exclusively female events at all in the UK.

Having a minimum rating of 2200 for 95% of players and one of 1900 for 5% would create distortions regardless. Players just about a minimum are liable to have flattering ratings by the simple effect that only their better performances get rated. I don't think we saw what the female players under 2200 were really worth until the male minimum was extended down to 2000.

David Sedgwick wrote: ↑

Wed Jul 10, 2019 2:05 pm

NickFaulks wrote: ↑

Wed Jul 10, 2019 11:54 am

I'm not sure where that 2600 number comes from.

https://ratings.fide.com/id.phtml?event=11600098.

I was giving Rausis the benefit of the doubt that his 2600+ ratings in 2017 and 2018 reflected a genuine increase in his playing strength compared with previously. I am disinclined to be so charitable about his most recent spike.

Rausis was caught cheating by Laurent Freyd at a tournament in Strasbourg yesterday. It involved a phone in a toilet. There has been triumphant posting on Facebook about the result by Emil Sutovsky and Yuri Garrett. It seems as though Rausis had been on the radar for this for a while, without ever being caught.

Alex Holowczak wrote: ↑

Fri Jul 12, 2019 7:44 am

Rausis was caught cheating by Laurent Freyd at a tournament in Strasbourg yesterday. It involved a phone in a toilet. There has been triumphant posting on Facebook about the result by Emil Sutovsky and Yuri Garrett. It seems as though Rausis had been on the radar for this for a while, without ever being caught.

Chess24.com story.
Sutovsky Facebook post.

The story is here: https://chess24.com/en/read/news/gm-igo ... t-cheating

I've also been following for a while and players have been extremely skeptical of his results. As has been pointed out it's tough for even a GM to consistently beat these 2200/2300 players.

Roger de Coverly wrote: ↑

Thu Jul 11, 2019 9:05 am

Having a minimum rating of 2200 for 95% of players and one of 1900 for 5% would create distortions regardless. Players just about a minimum are liable to have flattering ratings by the simple effect that only their better performances get rated.

If "about" is a typo and you meant "above" then you are precisely 180 degrees out. If you are just above the minimum then all your games against stronger opposition, against whom you expect a negative score, are counted and none of their games against weaker opposition.

Brian Towers wrote: ↑

Fri Jul 12, 2019 10:31 am

If you are just above the minimum then all your games against stronger opposition, against whom you expect a negative score, are counted and none of their games against weaker opposition.

I disagree, at least for the period where the cut off was first 2200 and then 2000. Suppose your playing style is that you can draw with everyone within 200 points of yourself in both directions. Also assume you meet a field of your strength, assuming it to be 2000. Then if you meet players with ratings between 1800 and 2200 with your own rating of 2000, your performance is 2000. Now discard all results against players below that. These are the games that would lose you points, so your rating is higher.

For the several years where the rating cutoff was 2000, 4NCL players below 180 would have fIDE ratings of enough over 2000 to keep them. That was not least because accidents against players below 2000 were ignored.

Let's not forget that it's about a 1 in 2 chance that you draw against a player rated 200 points above you and also a 1 in 2 chance that you fail to beat someone 200 points below.

Roger, what you've written has obvious errors on almost every line. I'm too lazy to "Fisk" what you've written. I'll just point out two things.
1) Earlier in the thread you appear to accept that playing predominantly much weaker opposition is a surefire way to increase your rating over time.
2) There are far more players with playing strength between 1800 and 2000 than there are between 2000 and 2200.

Roger de Coverly wrote: ↑

Fri Jul 12, 2019 11:00 am

.

Let's not forget that it's about a 1 in 2 chance that you draw against a player rated 200 points above you and also a 1 in 2 chance that you fail to beat someone 200 points below.

Not sure this is quite right

Brian Towers wrote: ↑

Fri Jul 12, 2019 10:31 am

Roger de Coverly wrote: ↑

Thu Jul 11, 2019 9:05 am

Having a minimum rating of 2200 for 95% of players and one of 1900 for 5% would create distortions regardless. Players just about a minimum are liable to have flattering ratings by the simple effect that only their better performances get rated.

If "about" is a typo and you meant "above" then you are precisely 180 degrees out. If you are just above the minimum then all your games against stronger opposition, against whom you expect a negative score, are counted and none of their games against weaker opposition.

If Roger had said "Players just below a minimum are liable to have flattering ratings by the simple effect that only their better performances get rated", he would have been right.

Suppose the rating cutoff is 2000 and the player's true strength is 1950. If they perform at, or below, true strength they don't get a rating because they haven't met the minimum standard to get one. They only get a rating when they perform well, at a level over 2000. (Obviously, they're liable to lose it fairly soon when they have an average result which brings their rating down below 2000.)

JustinHorton wrote: ↑

Fri Jul 12, 2019 11:34 am

Roger de Coverly wrote: ↑

Fri Jul 12, 2019 11:00 am

Let's not forget that it's about a 1 in 2 chance that you draw against a player rated 200 points above you and also a 1 in 2 chance that you fail to beat someone 200 points below.

Not sure this is quite right

I think Roger has just phrased it awkwardly without, I'm sure, any intention to mislead. What he means is that your expected score against a player rated 200 points higher than you is about 0.25 (see table 8.1a here), i.e. in 2 games you are expected to lose 1 and draw 1.

Ian Thompson wrote: ↑

Fri Jul 12, 2019 11:40 am

If Roger had said "Players just below a minimum are liable to have flattering ratings by the simple effect that only their better performances get rated", he would have been right.

It applies just as much to players slightly above the rating cutoff. At 2050,with a 2000 cutoff the draw with the player rated 2250 counts but the draw with the player rated 1850 doesn't. Drawing with a player rated 2010 who is 1950 in strength doesn't cost so much either.