ChessBase Magazine 177: Arsenal for the aspiring player

by Nagesh Havanur
5/26/2017 – An issue packed with rich material. Four Great events, Women’s World Championship, Tata Steel (Carlsen overtaken by Wesley So) Gibralter and FIDE Grand Prix. 2284 OTB games, of which 149 are annotated. Commentators include Wesley So, Levon Aronian and Mihail Marin among others. Eleven opening surveys, ranging from the Sicilian to the Slav, to surprise your opponent! Exercises in opening traps, middle game tactics and endgame technique. DVD lessons by Rainer Knaak, Rogozenco, Simon Williams and Karsten Müller plus a live commentary by Danny King. CBM 177 offers both an armory and ongoing course for ambitious players. Review by Prof. Nagesh Havanur.

ChessBase 14 Download ChessBase 14 Download

Everyone uses ChessBase, from the World Champion to the amateur next door. Start your personal success story with ChessBase 14 and enjoy your chess even more!


Along with the ChessBase 14 program you can access the Live Database of 8 million games, and receive three months of free ChesssBase Account Premium membership and all of our online apps! Have a look today!

More...

ChessBase Magazine #177

Review by Nagesh Havanur

ChessBase Magazine #177 (DVD + Booklet)
Languages: English, German
Delivery: Download, Post
Level: Any
Price: €19.95 – €16.76 or $18.10 without VAT (for customers outside the EU)

Tan Zhongyi who was crowned Women’s World Champion this year has graced the cover of this issue. It also carries all the games from the event. As is known, the contest for the title was marred by controversy. However, that should not detract from the extraordinary effort that the players made to win and qualify round after round in this knockout event.

Perhaps the toughest encounter was the last round between Tan Zhongyi and Anna Muzychuk. In this issue the game is annotated by GM Moradiabadi. Here I shall limit myself to the final phase that proved decisive.

[Event "Women's World Championship"] [Site "?"] [Date "2017.03.03"] [Round "?"] [White "Tan, Zhongyi"] [Black "Muzychuk, Anna"] [Result "1-0"] [Annotator "Nagesh Havanur"] [PlyCount "83"] [EventDate "2017.??.??"] 1. d4 d5 2. c4 c6 3. Nf3 Nf6 4. g3 Bf5 5. Nc3 e6 6. Nh4 dxc4 7. Nxf5 exf5 8. e3 Nbd7 9. Bxc4 Nb6 10. Be2 Bd6 11. Bf3 O-O 12. O-O Re8 13. Qc2 Qd7 14. b3 Re7 15. Na4 Rae8 16. Nc5 Qc8 17. Bd2 Nbd5 18. Rac1 Ne4 19. Bg2 g6 20. b4 Qc7 21. Nxe4 fxe4 22. b5 Ba3 23. Rb1 cxb5 24. Qb3 Qd6 25. Qxb5 Rc8 26. Qb3 Kg7 27. Bc1 Bxc1 28. Rbxc1 Rc6 29. Rxc6 Qxc6 30. Qa3 Qb6 31. Rc1 {[#]} Qb4 {A dangerous gamble. Moradiabadi rightly questions this move.} ({It didn't hurt to play} 31... a6 { If} 32. Rc5 Qb1+ 33. Bf1 Rd7) 32. Qxa7 Nc3 {He is also critical of this move. However, the move is not without merit. Black gains tempo targeting both the a-pawn and the rook. The knight also releases the castle from defending the pawn on e4.} 33. Rf1 $6 {seeking safety first} (33. Re1 $4 {loses to} Nb5 $19) (33. Bf1 $2 {costs the exchange after} Rc7 $1 34. Kg2 Nb5 35. Bxb5 Rxc1) ({ If White wanted more, he had to try} 33. a3 $1 Qb2 34. Rf1 {with favourable chances, though much play is left in the position after} Qb5 ({or} 34... Re6 35. Qc5 b6 36. Qb4 Qxb4 37. axb4 $14)) 33... Qc4 $2 {A plausible move targeting the a-pawn and also making way for his b-pawn. However, it allows the White queen to re-enter the diagonal and invade dark squares with great effect.} ({Moradiabadi recommends} 33... b5 $1 {Now} 34. Qa8 {appears intimidating as neither the knight nor the rook can move away without losing the e-pawn. However, Black has the saving move} Qa4 $1 35. Qxa4 bxa4 $11) 34. Qa3 $1 Re6 35. Re1 Ra6 36. Qe7 Rxa2 $6 {Rushing to make it to the time control. But the e-pawn is too important to lose.} ({Moradiabadi again gives} 36... b5 $1 {Here Black will have to tread carefully.} 37. Qe5+ ({not} 37. Bxe4 $4 Re6 $19) 37... Kg8 38. d5 $1 Nxd5 (38... Rxa2 39. d6 Ne2+ 40. Kh1 Qc3 41. Qe8+ Kg7 42. Rf1 Qf6 43. Qxb5 {is rather dangerous.}) 39. Rd1 Nf6 40. Bf1 Qxa2 41. Bxb5 Rb6 42. Qc5 Ng4 43. Qc8+ Kg7 44. Qxg4 Rxb5 45. Qxe4 Rb2 46. Rf1 $14) 37. Bxe4 $6 (37. Qxb7 $1 {would have precluded the counterplay that followed with the advance of the b-pawn.}) 37... b5 38. Bf3 b4 $2 {"This move allows a dangerous check" -Moradiabadi} ({Instead he suggests} 38... Qe6 39. Qb4 Qf6 40. Qxc3 Qxf3 41. Rf1 Qd5 {with good drawing chances.}) 39. Qe5+ Kh6 $4 (39... Kg8 $1 { would have prolonged resistance, though the advance of White's d-pawn would have outweighed any chances of Black with his b-pawn.}) 40. g4 $1 f6 41. Qxf6 Ra5 42. h4 {and a despairing Muzychuk resigned.} 1-0

A traumatic defeat for Anna Muzychuk! However, Tan Zhongyi held steady nerve that enabled her to win.

In retrospect I think it’s time to have a level playing field for men and women in chess. There is plenty of chess talent among the latter and as long as it is confined to “girls only” events, it would have a stunted growth. Let parents encourage their daughters to play in open contests with boys, and in course of time they would be following the footsteps of Judit Polgar.

This brings me to tournament games in the magazine. Among others they cover battles from Wijk aan Zee, Gibralter and FIDE Grand Prix events. As is known, the Wijk aan Zee Tournament was won by Wesley So ahead of Carlsen with a half point margin. Magnus had the discomfiture of being beaten by Richard Rapport (an elegant performance annotated by Aleksandr Lenderman in this issue). Rapport who finished at the bottom of the tournament table also had the distinction of playing the craziest game of the tournament with Wesley So. Here I shall limit myself to one line pointed out by the commentator in this issue.

[Event "79th Tata Steel GpA "] [Site "?"] [Date "2017.01.16"] [Round "?"] [White "So, W."] [Black "Rapport, R."] [Result "1-0"] [ECO "E18"] [WhiteElo "2808"] [BlackElo "2702"] [Annotator "Nagesh Havanur"] [PlyCount "95"] [EventDate "2017.??.??"] 1. Nf3 Nf6 2. g3 b6 3. d4 Bb7 4. c4 e6 5. Bg2 Be7 6. O-O O-O 7. Nc3 Ne4 8. Bd2 Bf6 9. Rc1 Nxd2 10. Qxd2 d6 11. d5 e5 12. e4 Nd7 13. h4 a5 14. Bh3 Nc5 15. Kg2 Bc8 16. Rh1 Bxh3+ 17. Rxh3 Qd7 18. Nh2 h5 19. f3 g6 20. g4 Kh7 21. Rh1 Rg8 22. Qd1 Bg7 23. Nf1 Bh6 24. Kf2 Bc1 25. Ng3 Bxb2 26. Nb5 Bc1 27. gxh5 Bf4 28. Nc3 f5 29. hxg6+ Rxg6 30. Nxf5 Rag8 31. Kf1 b5 32. cxb5 Rg2 33. Qb1 {[#]} Qf7 ({ Now the clinical} 33... Rd2 $1 {followed by the invasion by the rook or the knight would have won as pointed out by many commentators.}) ({In this magazine Igor Stohl offers one more possibility.} 33... c6 $5 34. b6 (34. dxc6 Qa7 35. b6 Qa6+ 36. Qb5 Qxb5+ 37. Nxb5 Nd3 $19) 34... Qb7 {followed by ...Qa6 winning. That's more aesthetically pleasing.}) 34. Ne2 Qg6 35. Ne7 Rf2+ 36. Kxf2 Qg2+ 37. Ke1 Rg3 38. Rxg3 Qxh1+ 39. Rg1 Qxf3 40. Nxf4 Qe3+ 41. Ne2 Nd3+ 42. Qxd3 Qxd3 43. Ng8 Qf3 44. h5 Kh8 45. Rg6 Qh1+ 46. Kd2 Qxe4 47. Nf6 Qb4+ 48. Ke3 1-0

A terrific essay in tactical fantasy that failed to meet the reality test! M. Tal would have sympathized. Igor Stohl has analysed the game in great depth in this issue. Danny King has also made a video presentation of the same. What we have seen here is the proverbial tip of the iceberg.

This brings me to other sections of the Magazine. There are eleven opening surveys ranging from the Sicilian to the Slav. For a change I would mention the survey by Michal Krasenkow on an unorthodox line in the English Opening: 1.Nf3 Nf6 2.c4 c5 3.Nc3 d5 4.cxd5 Nxd5 5.e4 Nb4 6.Bc4 Be6 (A34)

 

White’s opening play circumvents Black’s attempt to initiate the Grünfeld with 3…d5 followed by…g6, Bg7 and 0-0. Is it worth it? Krasenkow shows how White can run into some serious danger with this line. I have simplified his analysis for younger readers unfamiliar with theory and also indicated improvements for White

[Event " "] [Site "?"] [Date "????.??.??"] [Round "?"] [White "Avoiding Grünfeld"] [Black "English Opening"] [Result "*"] [ECO "A34"] [Annotator "Nagesh Havanur"] [PlyCount "38"] 1. Nf3 Nf6 2. c4 c5 3. Nc3 d5 4. cxd5 Nxd5 5. e4 {A move introduced by Aron Nimzowitsch.} (5. d4 Nxc3 6. bxc3 g6 7. e4 {leads to regular Grünfeld lines.}) 5... Nb4 (5... Nxc3 6. dxc3 Qxd1+ 7. Kxd1 {favours White according to theory.}) 6. Bc4 (6. d3 Bg4 7. Be3 e6 8. Rc1 N4c6 $11 {is said to be innocuous. However, the fight is still ahead.} ({Not} 8... N8c6 9. a3 Na6 {and the knight is out of play.})) (6. Bb5+ N8c6 7. O-O {deserves attenion-NSH}) 6... Be6 {This move pioneered by Kurt Richter way back in 1930 deserves a revival according to Michal Krasenkow.} (6... Nd3+ 7. Ke2 Nf4+ 8. Kf1 Ne6 {is in vogue.}) 7. Bxe6 ( 7. Bb5+ N8c6 {should be OK for Black.}) 7... Nd3+ ({After} 7... fxe6 8. O-O { White has easier development-NSH}) 8. Kf1 fxe6 9. Ng5 Nc6 $5 $146 {sacrificing a pawn for initiative} (9... Qd7 {is the older line.} 10. Qg4 (10. Qf3 { is met by} Ne5) 10... Nc6 11. Qxe6 (11. Nxe6 $2 {loses to} Nd4 12. Nxg7+ Bxg7 13. Qxg7 O-O-O $19) 11... Qxe6 12. Nxe6 Kd7 13. Ng5 ({Not} 13. Nxf8+ $2 Rhxf8 14. f3 Nd4 15. Ne2 Nxf3 $1 $17) 13... Nd4 $44) 10. Nxe6 Qd7 11. Nd5 {This side line may help White to make a safe landing.} ({The main line of Krasenkow's analysis runs} 11. Nxc5 Nxc5 12. Qh5+ g6 13. Qxc5 e6 14. Qc4 ({Or} 14. Qb5 a6 15. Qe2 Bc5 16. d3 O-O $1 $44) 14... Bg7 15. h4 O-O 16. Rh3 Nd4 $44) 11... Rc8 12. Qh5+ ({The more obvious} 12. Qg4 $2 {fails to} Nd4 13. Nxg7+ Bxg7 14. Qxg7 Rf8 15. f3 Qe6 $1 16. Qxh7 Rf7 17. Qh4 Qa6 $19) 12... g6 13. Qh3 c4 14. b3 Nd8 15. bxc4 $1 ({Better than} 15. Nxd8 Qxh3 16. gxh3 Rxd8 $17) 15... Nxc1 16. Nxd8 Qxh3 17. gxh3 Rxd8 18. Rxc1 e6 19. Nc7+ Ke7 $44 *

A minefield for White! However, if he survives, it does not look so bad in the end. Here I would put more emphasis on the side lines of Krasenkow’s analysis. White can transpose to the Grünfeld with 5.d4 as suggested by him. Or he can try 6.Bb5+Nbc6 7.0-0, a line that is less explored.

Apart from these surveys, there are regular exercises in opening traps, middle game tactics and endgame technique. Here I would single out Mihail Marin’s essay on the move …d6-d5, Black’s breakthrough in the Spanish with illustrative examples from great masters of the opening.

In all there are 2284 OTB games of which 149 are annotated. Commentators include Wesley So, Levon Aronian and Mihail Marin among others. I missed Telechess section. Hopefully, the grandmasters would be back next time to present games from correspondence chess.

Recommended – more info on the DVD is here

 


The editor’s top ten

  1. Finally winning Wijk! Wesley So presents his game against Wojtaszek and explains why victory in the classic tournament in the Netherlands was so important to him.
  2. An “oddball” idea: Levon Aronian explains how with 8.Na3 in the Catalan he achieved a brilliant win against Giri.
  3. French aggression: attack together with Simon Williams "Move by Move“ – just like Ju Wenjun did in her victory over Hou Yifan!
  4. Spanish temptation: let Rainer Knaak show you how to entice your opponent into the 7.Ng5 trap with 6...d7-d6.
  5. “Raving Rooks”: along with Oliver Reeh light some tactical fireworks with double threats and multi-purpose moves (interactive video).
  6. For and against the Bogo-Indian: GM Postny sums up the trends of recent years and shows you what is at present recommended.
  7. "Portuguese Scandinavian": let Petra Papp show you how best to meet the gambit 1.e4 d5 2.exd5 Nf6 3.d4 Bg4.
  8. A "mad way to win“: Anish Giri explains how he got Andreikin into trouble with the Scotch and what brilliant engine move he missed at the end.
  9. Fantastically shattered: enjoy Rapport's brilliancy against So in Daniel King’s video analysis.
  10. Surprise in the Queen’s Indian: let top Swiss player Yannick Pelletier unravel Aronian's 9.Be1 (video).

All opening articles in ChessBase Magazine #177

Lampert: English A32 (Recommendation for White)
1.d4 Nf6 2.c4 c5 3.Nf3 cxd4 4.Nxd4 e6 5.g3 Qc7

In this variation, classified under the English Opening, but more likely having arisen from a Declined Benoni, White usually sacrifices his c4-pawn. As Jonas Lampert demonstrates in his article, in return he obtains compensation or even something more.

Krasenkow: English A34 (Recommendation for Black)
1.c4 c5 2.Nf3 Nf6 3.Nc3 d5 4.cxd5 Nxd5

As Michal Krasenkow explains, every fan of the Grünfeld has to be prepared for White choosing a different starting move order and delaying d2-d4 so as not to allow an original Grünfeld. One of these variations is the subject of the present article by our Grünfeld expert.

Moskalenko: Dutch Defence A80 (Recommendation for Black)
1.d4 f5 2.Nc3 d5 3.Bg5 Nf6 4.Bxf6 exf6 5.e3 Be6

Many completely avoid the position in the diagram (see Part 2 in CBM 172). But nowadays the variation is dealt with slightly differently: above all c7-c6 is absolutely avoided. Viktor Moskalenko evaluates the ensuing positions as slightly static but very playable for Black.

Papp: Scandinavian Defence B01 (Recommendation for White)
1.e4 d5 2.exd5 Nf6 3.d4 Bg4

Petra Papp has against the Portuguese Variation 3...Bg4 a clear plan which gives White an advantage in all lines. The very starting move 4.Bb5+ is in her opinion somewhat more accurate than the popular 4.f3.

Breutigam: Pirc Defence B08 (Recommendation for Black)
1.e4 d6 2.d4 Nf6 3.Nc3 g6 4.Nf3 Bg7 5.Be2 0-0 6.0-0 a6

The move 6...a6 in the Pirc Defence has been known for a long time, but now this interesting variation has been enriched by remarkable facet: the idea, after Nb8-c6 followed by d4-d5, of retreating with the knight to a7. Martin Breutigam is enthusiastic about it.

Petrov: Sicilian Defence B23 (Recommendation for White)
1.e4 c5 2.Nc3 Nc6 3.Bb5 Nd4 4.Nf3

As Marian Petrov explains in his article, the whole variation is above all based on the fact that by playing the “natural” 4...Nxb5 Black brings difficulties down on his own head, because White takes advantage of his lead on development for a rapid d2-d4.

Szabo: Sicilian Defence B53 (Recommendation for Black)
1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Qxd4 Nc6 5.Bb5 Bd7

The Hungarian Variation 4.Qxd4 is employed above all to avoid the Najdorf System. Krisztian Szabo presents a plan for Black with which he obtains at least a level game against both 6.Bxc6 and the recently popular 6.Qd3.

Ris: French Defence C11 (Recommendation for White)
1.e4 e6 2.d4 d5 3.Nc3 Nf6 4.e5 Nfd7 5.f4 c5 6.Nf3 Nc6 7.Be3 a6 8.Qd2 b5 9.dxc5 Bxc5 10.Bd3

As Robert Ris explains, the main idea behind Karjakin’s 10.Bd3 Qb6 11.Bf2! is above all that Black will have a few difficulties in the possible endgames. There may be quite good alternatives, but they too are not very easy to play.

Kuzmin: Queen's Gambit D37 (Recommendation for Black)
1.d4 d5 2.c4 e6 3.Nf3 Nf6 4.Nc3 a6

Facing the unusual 4...a6 White mainly transposes with 5.cxd5 exd5 6.Bg5 to an Exchange Variation. As Alexey Kuzmin shows in his article on the DVD, Black develops with 6...Be6 and obtains very playable positions, though perhaps not complete equality.

Marin: Slav/Grünfeld D90 (Recommendation for Black)
1.d4 d5 2.c4 c6 3.Nf3 Nf6 4.Nc3 g6

The relationship to the Schlechter Defence cannot be missed, but White has not yet played e3 and has some additional possibilities. But as Mihail Marin demonstrates, there is no need for Black to fear these, though knowledge of the Grünfeld is required.

Stohl: Catalan E06 (Recommendation for both sides)
1.d4 Nf6 2.c4 e6 3.g3 d5 4.Bg2 Be7 5.Nf3 0-0 6.Qc2

The immediate 6.Qc2 (instead of 6.0-0 dxc4 7.Qc2) is in no way a new idea, but some recent games (especially So-Wojtaszek, Wijk aan Zee 2017) have once again attracted attention to the whole variation. Igor Stohl sums up the latest state of affairs in his article on the DVD.

Buy ChessBase Magazine 177 in the ChessBase Shop



Prof. Nagesh Havanur (otherwise known as chessbibliophile) is a senior academic and research scholar. He taught English in Mumbai for three decades and has now settled in Bangalore, India. His interests include chess history, biography and opening theory. He has been writing on the Royal Game for more than a decade. His articles and reviews have appeared on several web sites and magazines.
Discussion and Feedback Join the public discussion or submit your feedback to the editors


Discuss

Rules for reader comments

 
 

Not registered yet? Register

Petrarlsen Petrarlsen 5/31/2017 09:09
@ lajosarpad :

"(...) I think we did not drift totally away from the article."

I agree.

This discussion derived directly from an element of the article, so, for me, it wasn't misplaced here.

I can perfectly understand that the author would have prefered to see more elements of his article discussed here, but no-one can really control this, in my opinion...
lajosarpad lajosarpad 5/31/2017 08:56
@Petrarlsen

Unfortunately our debate is not welcome here, therefore I have to refrain from answering, even though you had some interesting points and I would have enjoyed discussing them. So, for now let's agree to disagree.

@Chessbibliophile

First of all, I would like to apologize for being part of a debate which was not 100% about the content of the article. Yet, you need to understand that the discussion which was started here was generated by your comment of

"In retrospect I think it’s time to have a level playing field for men and women in chess."

so I think we did not drift totally away from the article. Drcloak seems to not like feminism and he considered this quote to be his daily doze of feminism. The problem with his comment was its style, I believe. I understand you that you were insulted by Jackie's defamatory comment as well. However, your response was not too much better either:

"You could apply to Mumbai University and seek information about my writing
(a comparative study of three works of European fiction dealing with the Russian Revolution). My more recent work was editing an Indian language work about saint literature. It is not something that you can understand, let alone relate to."

I know it is easier said than done, but I recommend avoiding trolling as a response to trolling, especially if one is the author of the article being discussed.
chessbibliophile chessbibliophile 5/31/2017 07:30
So what is the better course for women players in the long term?
In the review I wrote, “Let parents encourage their daughters to play in open contests with boys, and in course of time they would be following the footsteps of Judit Polgar.”
It’s easier said than done. In many developing nations the outlook is conservative. Parents are uneasy about their daughters mixing with grown up boys and men. Their comfort level is with girls-only tournaments. There is also one more factor to be considered. If a girl is talented, she scores better with other girls and the results are immediate. Many parents crave for instant success for their daughters rather than risk failure and disappointment that comes in events common to both boys and girls.
The conscious decision that the Polgar parents made with Judit is still the best: allowing the child to play in only common events. If parents of the present generation do that, the results would be visible in the next ten years.
The ultimate handicap for women players would be raising a family and it’s at this point that many give up professional chess. That’s what happened to Judit Polgar. The next generation of female players may take a different path and pursue their career notwithstanding family obligations. They may not win the world championship, but would compete among the elite.
I have explained a great deal and do not intend to return to this page. May I still ask readers to comment on games and analysis in this magazine? That would help others rather than a gender discussion that goes on forever.
chessbibliophile chessbibliophile 5/31/2017 07:29
Dear readers,

Most of the discussion here has related to female players. Unfortunately, much of the statistics does not relate to the main issue: How well have the best women players performed with the best among men? (Thanks, Petralarsen)
Here the paradigm has to be at the highest level, one being Judit Polgar and the other being Hou Yifan.
With a peak performance at 2735 (current average being 2675) Judit Polgar was among the top ten around the year 2003 With marriage and motherhood her career took a nosedive and she came last in the 2005 World Championship. However, she regained her place among the top ten when she participated in the World Cup in 2011. Judit’s score against most of her contemporaries, Kasparov, Kramnk and Anand is negative. Each of them crossed 2800 barrier, she didn’t. But if you look at the games she drew or won with them, they are fabulous. No amount of statistics can measure her talent or style of play.
With all respect to Hou Yofan she is not a prodigy like Judit and does not possess that kind of natural gift. She is talented, hard working and determined to do well. But let us be realistic. With her current rating of 2611 she is not among the top 100 players who have 2700+ ratings. As for 2800 + rating achieved by Carlsen and his peers, it’s beyond her reach. She did not do well in Gibralter Open (+5 -3 =2) and FIDE GP Sharjah (+0 -1 = 8). She did better in Grenke (+2 -2 = 3) and FIDE GP Moscow (+3 -2 =4). In both she finished in the middle of the score table. Her results are mixed. She has beaten Caruana and Nepomniachtchi. She has also lost to Aronian and Grischuk. It would take a long time for her to be among the top players. Here even a conservative figure of top 20 would be doubtful as her male peers would not be sitting idle and they would also be improving. So the gap would be constant.
chessbibliophile chessbibliophile 5/31/2017 07:28
Dear @Jackie

As I mentioned to dr.cloak, I happen to be the author of the review. May I ask why I should satisfy the idle curiosity of an anonymous reader who indulges in insinuation, innuendo and abuse? Anyone can see me here along with my name, personal information and writing. The same cannot be said of you. Could you please tell readers who you are? Let them know. Meanwhile I draw your attention to Rules for Comment:
“Any and all comments and contributions must be legitimately related to the article in question. We place a great deal of importance on comments being objective, as well as on the tolerance and respect of the opinions of others.”
Do these rules apply to you or are you a law unto yourself?
OK, as you have had the temerity to question my bonafides, let me enlighten you a bit. Much of my work was done in the pre-internet era and so you were not able to find it on Google. You could apply to Mumbai University and seek information about my writing
(a comparative study of three works of European fiction dealing with the Russian Revolution). My more recent work was editing an Indian language work about saint literature. It is not something that you can understand, let alone relate to. Now I am into bi-lingual projects in history and Indology. I see no reason to go into details about them on a chess platform.
Be that as it may, posting a comment on a public forum is an act of responsibility. Have you had any when you made wild and reckless allegations about the author of this review? Google is not the be-all and end-all of search for information, let alone knowledge. If you were decent and serious about your question I could have given you a better answer. Sadly, you don’t deserve any.
chessbibliophile chessbibliophile 5/31/2017 07:27
Dear @dr.cloak,
I happen to be the author of this review. Now do you know me personally? You don’t. So what gives you the presumption that you can address me with first name familiarity?
What is all this stuff about my giving you a “daily” dose of feminism? Daily? I don’t even know you in the first place. As for feminism, it doesn’t come into the picture. Of course if you are a misogynist you would see it everywhere.
Courtesy and civility are not part of your make-up. As such, your comment does not merit a reasoned response. You may be tempted to write an aggressive reply and have the last word. However, I have neither time nor inclination to come back here, waste my time and that of readers in pointless exchange of words with you.
chessbibliophile chessbibliophile 5/31/2017 07:25
Dear friends,
As the author of this review, I did expect to see comments related to the games and analysis mentioned above. As of now I have yet to see anything relating to the same. That’s a pity. However, I did find two comments in bad taste and the second of them practically amounts to libel. In any case both are a violation of ChessBase Rules for comment:
http://en.chessbase.com/pages/discussion-rules

So there is a direct reply to these readers first and the general issue of women players would be taken up next.
Petrarlsen Petrarlsen 5/31/2017 06:29
@ lajosarpad (6/6) :

- In my opinion, the problem, when using your conception of a "gap" between men and women in Chess, is that it is much to much linked with subjective criterion.

For example, if you would consider (and why wouldn't you ??) that "serious chess" begins at a 2400+ IM level, then, there is absolutely no gap between women and men in chess.

So I don't think this method is a good method, for comparing men and women in chess.

- "Even the numbers I have given you before show that top women chess is getting better."

For me, you didn't show that "top women chess is getting better" but only that, quite probably, the best women's play is stronger now that in 2000.

But, as I said previously, in my opinion, this doesn't mean anything : perhaps that Hou Yifan would beat Capablanca, today, but that wouldn't mean at all that she is a "better" player, only that, globally, chess has much evolved, since Capablanca's time. And yes, the best women's play is indeed stronger now than in 2000, but this is only a natural reflection of the global increase in level of chess in general.

- "Difference answers the question of expected result in a game, while ratio answers the question of relative strength."

I still don't see how these two elements can be really different : for me, it stands to reason that, if chess strength grow, the expected results will also become better...
Petrarlsen Petrarlsen 5/31/2017 06:28
@ lajosarpad (5/6) :

- "I will acknowledge there is an inflation if the rating limit to achieve a GM title increases."

To increase this rating limit would also be a "political" decision. And it is quite simply possible that FIDE will not increase this limit exclusively for "political" reasons (and not at all for scientific reasons).

- "The level of chess games in general has increased."

In my opinion, this is quite sufficient to explain why there is a (slight) increase in number of women playing at GM level, without any link with any narrowing of the gap between men an women in chess.

- "(...) the number of female GMs is increasing, which, in my opinion means that the gap between male and female chess is getting narrower."

These are my preliminary (imaginary, obviously) data :

In 2020 : 15 female GMs for 800 male GMs.

In 2030 : 20 female GMs for 1600 male GMs.

In 2040 : 25 female GMs for 3200 male GMs.

In 2050 : 30 female GMs for 6400 male GMs.

Following your reasoning, these data would show a continuous narrowing of the gap between male and female chess.

But, for me, to go from 1.875 % of female GMs in 2020 to 0.46875 % of female GMs in 2050 wouldn't be at all what I would call a "narrowing of the gap".

So I still can't say that I am convinced by your reasoning on this point.

- "Then we have a different definition for the term "gap".
Standard definition: https://en.oxforddictionaries.com/definition/gap"

One of the definitions on this page of the Oxford Dictionary is : "A space or interval". I don't see why "a space or interval" couldn't be between two points ?? (following my own reasoning)
Petrarlsen Petrarlsen 5/31/2017 06:27
@ lajosarpad (4/6) :

- "The player of today has better access to engines, games, analysis, theory, books, opening theory, etc., etc. So how is a 2500 rating of today less strong than 17 years ago?"

In my opinion, this reasoning is wrong.

For example, very probably, if Eljanov was to play today Capablanca, with his "access to engines, games, analysis, theory, books, opening theory, etc.", he would probably beat him soundly.

Does this make Eljanov a better player than Capablanca ? In my opinion, clearly, no.

Perhaps even Hou Yifan, for example, could beat Capablanca (if you take into account all the "engine-related" elements). But would this mean that Hou Yifan is better than Capablanca ?? For me, very obviously, no...

So, in my opinion, it is just wrong to compare STRENGTH of players of different periods. Because they aren't in the same circumstances. It is as if someone would reproach Archimedes not to know everything that a modern scientist Nobel prize laureate would know ; it would not be coherent, in my opinion : Archimedes could NOT know all this, at his time... And thus, this would not be a reason to consider that he is necessarily weaker that this given Nobel prize laureate.

For example, IF (I don't know the numbers and this is just a completely imaginary example, certainly quite far from the true numbers) between 2000 and 2017 the number of grandmasters went from 50 to 800 active male grandmasters, it would seem perfectly obvious to me that the (real, this time) increase from 10 to 13 active 2500+ women players
wouldn't mean AT ALL that women players compete more easily at GM level. For me, it isn't possible to affirm that there is an increase in level (at GM level) for women without taking into account the global number of grandmasters at the same period.

I don't know the exact numbers, and I am not absolutely sure of it, but I think that, at grandmaster level, the proportion of women has greatly decreased between 2000 and 2017. If this is indeed true, could it really be possible to affirm, as you do : "Therefore the number of female GMs is increasing, which, in my opinion means that the gap between male and female chess is getting narrower." ? For me, no. Because, yes, the number of female GMs is increasing, but, if the PROPORTION between male and female GMs is greatly decreasing, this also means that, still in PROPORTION, women are less and less present in GM level chess. So the gap between male and female chess cannot be considered as getting narrower. (If, to use a comparison, the female grandmasters are literally "submerged by a flood of new male grandmasters", I don't see how this could constitute a narrowing of the gap between men and women in chess.)
Petrarlsen Petrarlsen 5/31/2017 06:27
@ lajosarpad (3/6) :

- About the use of ratios for comparisons between different players or levels :

I am not sufficiently competent in mathematics to analyze this in details.

But, for me, it seems simply impossible, for example, to say that, with a ratio of 0.9090909 between a 1000 and a 1100 player and a ratio of 0.9642857 between a 2700 and a 2800 player, the 2700 player must be considered as nearer (in one way or the other) to the 2800 player than the 1000 player is from the 1100 player.

As I see things, if in both cases (1000 to 1100 and 2700 to 2800) the expected score for the better player is 64 %, I simply find at first sight that such a result couldn't be right ; there must be a flaw somewhere.

For me, it seems really absurd that, for example, the average result (in terms of points) of a match between a 1000 and a 1100 player would be the same as the average result of a match between a 2700 and a 2800 player, and that the ratios would show that, in a way, the 2700 player is MUCH nearer to the 2800 player than the 1000 player to the 1100 player (as there is really a big difference between 0.9642857 and 0.9090909).

It is not at all a direct comparison, but, for me, it is a bit as if a mathematician was explaining that two runners finished a given competition with exactly the same time, but, for some complex mathematical reason, the result for this same competition of one of these runners was MUCH better than the other. I couldn't perhaps demonstrate in mathematical terms what would be wrong, but I would nonetheless KNOW that this couldn't be right either, because, if they finish the competition with exactly the same time, it is simply impossible that a mathematical demonstration shows that one of them has a much better result than the other.

Perhaps there is something that I don't understand somewhere, but, to me, this seems like a complete impossibility...
Petrarlsen Petrarlsen 5/31/2017 06:26
@ lajosarpad (2/6) :

- "I do not think that the favorability of the result should have any importance when we determine the methodology."

For me, the objective is to compare the level of the very best men to the level of the very best women. And it seems to me to stand to reason that to compare a long section of the quite stable men's list to an equivalent section of the completely nosediving women's list (exaggerating a little bit, I must admit !...) is not at all fair for women : it does not reflect a true comparison between the maximum level for men and the maximum level for women. In fact, it essentially reflect what we did know from the start : that, precisely, the women's lists decreases much faster than men's lists (and this last element has absolutely no link, in my opinion, with the maximum level of women compared to the men's level).

In my opinion, 10 players represents an interesting equilibrium : it isn't so dependent to personal factors than to compare the best players one with the other (the best man against the best woman), and it doesn't take into account players that are clearly not at all competing with the top men or women (for example, who would say that Daria Pustovoitova, Ding Yixin or Yuliya Shvayger are directly in competition with Hou Yifan ? I think that such players have nothing to do into a comparison between top men and women players - there are much too far from the real top players).

In fact, for me, the problem is that if a too long section of the women's list is used, a totally extraneous factor (in comparison to the maximum level of mens and womens) is included into the comparison : the "speed" with which the respective lists decrease. And I think that, as much as possible, this factor must be taken out of the comparison.
Petrarlsen Petrarlsen 5/31/2017 06:26
@ lajosarpad (1/6) :

- About potential differences in value for Elo ratings between men an women, I rather think that there are much more "links" (by the open tournaments in which women participate) between the women and men players of one given country, than between American and Russian players, for example. And I rather think that, globally, it must not be very often that "international players" play in succession with American and Russian players (apart from 2700+ grandmasters, and, to some extant, certain 2600+ grandmasters).

So, I don't see how it could be certain that there is necessary a equivalence between (for example) American and Russian players, and not between men and women ? For example, if we admit that, on the one hand, in a given country ("country A"), all the women participate regularly to open competitions, and that, on the other hand, no player from the country B ever plays in the country C, that, also, no player from the country C ever plays in the country B, and that there is only one time by decade an "international player" that plays successively in "country B" and in "country C", my impression is clearly that there is much less "links" between players from "country B" and "country C" than between men and women from "country A".

In my opinion, it is theoretically possible to have many "links" between men and women on the one hand, and very few "links" between men of two different countries, on the other hand, and, in this last hypothesis, it seems to me to stand to reason that the Elo systems of these countries could have a tendency to evolve differently.

By the way, the same could be said also about Junior competitions as for women competitions.

But I think that we agree on one thing : that the differences are very probably not very important for a same Elo rating value, between men and women.

- "(...) therefore even if Americans and Russians rarely or never play with each-other, they "play indirectly"."

I don't see the difference with women competitions : there are also quite a quantity of women that play both in women competitions and in open competitions, so that the other women and the men "play indirectly" by the intermediary of the women that play in both types of competitions.
Petrarlsen Petrarlsen 5/31/2017 06:26
@ chessbibliophile :

"Unfortunately, much of the statistics does not relate to the main issue. How well have women players performed with men?"

For me, the two most significant subjects are this subject (the very best women in history vs. the very best men, if I undestood well your meaning) and, at a given moment, the "top women" vs. the "top men". These are in my opinion the two elements that really permit an assessment of the situation between men and women chess players.

But I think that there were not many comments on your subject (the very best women in history vs. the very best men) because it is much simpler to assess and much better known, quite simply.
lajosarpad lajosarpad 5/30/2017 12:00
@Petrarlsen

"Clearly, if what Wikipedia affirms is true, in my example, between the 2020 ratings and the 2040 ratings, there would be no change in terms of real difference between men's and women's ratings. But your system would show a supposed decrease of this difference. So I think that this tends to show that it is not an appropriate system."

The thing Wikipedia affirms is true. However, difference is the input of a comparison which gives you a chance of results in a game, or a match. How do you measure the amount by which a player is better than the other in case of a difference? If you have a good formula, then it might be better than ratio, but at least I do not know of such a formula. I know how can it be used to estimate the results of a match, but do not know how this can be used to determine the amount a player is better than the other. Of course, if a person has better chance of succeeding than the other, then he/she is better. But by what amount? Ratio will not tell you the chance that someone will have a good result against the other. It tells you how much points does have a person compared to the other.
lajosarpad lajosarpad 5/30/2017 11:53
@Petrarlsen


"For example, in a Formula One competition, I don't see why you couldn't use the term "gap" about the distance and time that separate the first and the second drivers in a given competition. "

When I measure the gap between the first and the second pilot, I will look at the back of the first and the front of the second, not to their fronts. To measure the distance between the front of the two cars in determining the length of the gap does not really make sense for me. When the front of the second car is ahead of the back of the first, but the first is still slightly ahead, then I would say that the first is ahead of the second, but would not say there is a gap between the two. Since with time we have moments, that is points and not intervals, the difference will show a gap if they do not have the exact same time. Yet, when we talk about points which define intervals (like in the case of male and female chess), by "gap" I always mean the ratio or difference between the lowest point of the better and the highest point of the worse. When analysing whether the gap between male and female chess is narrowing or not, comparing Carlsen's rating to Hou Yifan and using a very small set of 10 players for each does not seem a plausible methodology for me.

"The advantage of this conception of the "gap between men and women" in chess is that, for me, it reflects much more precisely the maximum potential of men and women players at a given moment."

I surely did not address that. That's the gap between maximum potentials. I meant the gap between male and female chess, that is, in a GM tournament would the best women be whitewashed more than before. The answer seems to be no.

"Yes, it is possible that woman chess progressed until the beginning of "Polgar's reign", but, since then, I don't think that there are many elements that show significant signs of progress for woman chess. "

Even the numbers I have given you before show that top women chess is getting better. Top male chess has been getting better in a more rapid pace, but I do not think that to be very relevant when we look at the gap between male and female chess. Yes, there is an increasing number of male GMs. But the limit is still 2500. If that is not accurate, then we need to change it. And I was meaning a global trend. 100 years ago for instance there was practically no women chess at all.

"And if, from 2000 until now, no real progress can be seen in this domain, it doesn't seems to be a given that this progress will start again in the future... "

I think there was a progress, as I have shown before. And if you want to compare the very top with the very top, I think the next rating list will appeal ladies better, since Hou Yifan has "stolen" some points from male players. But I find that not really relevant either. I consider the ability to play on GM level to be the most relevant.

"IF this is true (Wikipedia isn't certainly an universal and absolute reference !), it DOES seem that it is more exact to use rating differences measured in terms of number of points rather than ratios, as you did previously... "

The difference is direct input in determining the expected score indeed. Yet this does not tell me too much in terms of average ratings. If the average rating of a set of people is 90% of the other compared to the previous 91%, that is much more telling for me in terms of difference than the expected score. Here I am not predicting a match, but comparing players. If someone has 75% chance to beat the other, then which is the amount that he is better? Difference answers the question of expected result in a game, while ratio answers the question of relative strength. I would however be interested to know the statistician's statement you have talked about, as I might be wrong, this is not exactly my field.
lajosarpad lajosarpad 5/30/2017 11:36
@Petrarlsen

"This because, if you do the same research for men's lists, you find even much more significant results"

The level of chess games in general has increased. The fact that male players at the very top are getting better than before is a natural consequence. However, the fact that the gap between the top male players and the top female players has widened does not mean that the gap between male and female chess has widened as well. A GM is a player who achieves GM norms and 2500 rating. The number of women achieving this level, or being very close to it, as I have shown previously increased. Therefore the number of female GMs is increasing, which, in my opinion means that the gap between male and female chess is getting narrower. And, as I stated before, the gap is not between the very top players. Take the intervals of [6, 10] and [33, 35]. The gap between the intervals is between the minimum of the higher and the maximum of the lower, which is 33 - 10 = 23 and not between the maximum values, which is 35 - 10 = 15. If the intervals change to [8, 15] and [33, 55], then the gap is getting narrower, since 33 - 15 = 18 < 23 and not longer because the top of the higher interval had a greater increase. The gap should always be calculated between the closest edge points. And in our case, comparing the very top male players to the very top female players do not answer whether the gap between male and female chess has increased or not. We need to compare the weakest of the strongest male players to the strongest female players. The strongest males are GMs. Weakest GMs are on 2500 and below. Women chess is approaching to this standard. Comparing the very top players will answer the question whether the absolute best that men can provide is better than the absolute best that women can provide. In our case the answer is yes and it has an increasing trend.

"So I think that, even this last element seems to show that the gap between top male and top women players is rather growing. "

Yes, indeed. The gap between the top players is growing. My statement was

"As a trend, the relative strength of women chess players compared to male chess players has been increasing and the gap between the two has been getting smaller."

Here I meant that women players are increasingly capable to compete with men. I did not address the very top male players against the very top female players. I addressed male chess and female chess in general. And I was really meaning the increase of female players playing successfully in open GM tournaments. I have given a few examples for the phenomenon and I have shown that female chess is approaching to a level where they will have a lot of GMs, the highest title achievement as a player, besides World Champion.

"For me, the gap between the women and the men is the gap between the best level where you find permanently women (2600, approximately, today) and the best level where you find permanently men (2800, still approximately, today)."

Then we have a different definition for the term "gap".
Standard definition: https://en.oxforddictionaries.com/definition/gap
lajosarpad lajosarpad 5/30/2017 11:15
@Petrarlsen

Open tournaments are accessible to women, women tournaments are closed to men. There are some women who play in open tournaments as well, this helps a lot in making men and women ratings comparable. I do not think a woman with a 2500 rating would play on a very different level than a man with 2500 rating, but there is probably some difference. If there is a difference, I am not sure about its size and direction. American and Russian players would be fully comparable to each-other even if an American would never play with a Russian, since the international tournaments involves some non-American-non-Russian players who play with the respective players and therefore even if Americans and Russians rarely or never play with each-other, they "play indirectly".

" I also think that to take the 10 best men and the 10 best women is better, if you want to really compare the level of the "top women" to the level of the "top men"."
I think the 50th man is among the best men and the 50th woman is among the best women, therefore 50 is a more significant seed. But if you do not agree with this and you want to compare the top 10, then you can do so. My script could come handy for this purpose as well.

"So, to use 50 players instead of 10, for comparisons between men and women, is quite unfavorable to women, and doesn't, in my opinion, reflect the true difference between the very best men and the very best women."

I do not think that the favorability of the result should have any importance when we determine the methodology. If one thinks that the 11th player is not among the best, that's ok, but I think we should favor reasons which are independent of the result which might arise.

The formula for updating the rating is

R'A = RA + K(SA - EA)

If we accept that the system is more-or-less precisely yielding the players' strength, then I would be really interested to know why would one think that we cannot compare the strength of players (in our case with aggregation) using ratios, why wouldn't a value which tells us how much percent is the average women ÉLŐ compared to men be meaningful.

A rating difference of 200 is not telling me very much. Let's suppose that from an average of 2600 and 2400, respectively the averages reach 2800 and 2600, respectively. The difference is the same, yet the ratio is different. For me the percent of how good are women in average in comparison to men is much more telling. I agree with you that the difference and the ratio are answers to different, yet, related questions.

"So, either, really, a given difference always has the same significance, be it, for example, a 100 points difference between 1000 and 1100 players or the same difference between 2700 and 2800 grandmasters, and it isn't possible to use ratios as you do, or, otherwise, a given difference hasn't the same significance in every part of the "Elo spectrum", but the one doesn't seems to be compatible with the other. "

The difference is direct input for the formula to determine the chance a player will win/draw against the other. The ratio is determining how good a player is, compared to the other and does not deal with the chances they will beat each-other.

"This is true, but seems to me to be related not to a decrease of the gap between men and women, but to, either a global increase in level (for men AND women), or an "Elo inflation". "

I have expected this argument. I am convinced that a 2500 player of today would easily beat a 2500 player of 17 years ago. The player of today has better access to engines, games, analysis, theory, books, opening theory, etc., etc. So how is a 2500 rating of today less strong than 17 years ago? I will acknowledge there is an inflation if the rating limit to achieve a GM title increases.
Petrarlsen Petrarlsen 5/30/2017 05:59
@ lajosarpad : I read this just now in Wikipedia (at the article : "Elo rating system") : "The difference in the ratings between two players serves as a predictor of the outcome of a match. Two players with equal ratings who play against each other are expected to score an equal number of wins. A player whose rating is 100 points greater than their opponent's is expected to score 64%; if the difference is 200 points, then the expected score for the stronger player is 76%."

IF this is true (Wikipedia isn't certainly an universal and absolute reference !), it DOES seem that it is more exact to use rating differences measured in terms of number of points rather than ratios, as you did previously...

I remind you what I said in one my previous posts of today :

"For an example, we will admit that in 2020, there will be a 200 points difference between the best female players and the best male players, with 2600 for the best women, and 2800 for the best men, and that (supposing that there would be a very significant "Elo inflation" between the two, which isn't at all certain or even probable, but which I don't suppose is either impossible) in 2040, the difference between the best women and the best men will still be 200 points, but with 2700 (and not anymore 2600) for the women, and 2900 (and not anymore 2800) for the men.

With this example, the difference between men and women would be considered to be the same when taking into account the difference in terms of number of Elo points (200 points), but not in terms of ratios (the difference would seem to diminish : 0.9285714 in 2020, and 0.9310344 in 2040), thus, the two systems cannot be used indifferently, as the results obtained with each of them are not the same."

Clearly, if what Wikipedia affirms is true, in my example, between the 2020 ratings and the 2040 ratings, there would be no change in terms of real difference between men's and women's ratings. But your system would show a supposed decrease of this difference. So I think that this tends to show that it is not an appropriate system.
Petrarlsen Petrarlsen 5/30/2017 04:13
@ lajosarpad :

"The very fact that Hou Yifan played decently at the Grand Prix, Polgár was 8th in the world and invited to play directly for the open World Champion title, Polgár achieved bronze in the European Chess Championship not long before she retired, Hou Yifan was shared first at Gibraltar in the not distant past shows us that there are some women who are able to compete at the highest level and achieve better than par results."

Yes, it is possible that woman chess progressed until the beginning of "Polgar's reign", but, since then, I don't think that there are many elements that show significant signs of progress for woman chess.

And if, from 2000 until now, no real progress can be seen in this domain, it doesn't seems to be a given that this progress will start again in the future...
Petrarlsen Petrarlsen 5/30/2017 03:55
@ lajosarpad :

Erratum... : in my first post of today, read : "(...) to compare the 50 best women to the 50 best men", instead of : "(...) to compare the 50 best men to the 50 best men" (I don't think that this last comparison would be very useful !!).
Petrarlsen Petrarlsen 5/30/2017 03:45
@ lajosarpad (3/3) :

- "In our case, the gap between male and female chess is the difference in abilities between the worst of the best male players and the best of the best female players."

The first problem is that the "worst of the best male players" can be defined in many different ways. There can be the last of the 2800+ GMs, the last of the 2700+ GMs, the last Top 100 GMs, the last of the 2600+ GMs, the last of the 2500+ GMs, or the last of the 2400+ IMs, for example.

For me, the gap between the women and the men is the gap between the best level where you find permanently women (2600, approximately, today) and the best level where you find permanently men (2800, still approximately, today). It is a gap between two points, but I don't see why a gap couldn't be between two points rather than between two groups of players... For example, in a Formula One competition, I don't see why you couldn't use the term "gap" about the distance and time that separate the first and the second drivers in a given competition.

The advantage of this conception of the "gap between men and women" in chess is that, for me, it reflects much more precisely the maximum potential of men and women players at a given moment.

And, in my opinion, to compare the number of women that can compete at GM level between two different periods, as 2000 and 2017, without taking into account the (probable - one more time, I don't know for the moment how to check this...) considerable increase in number of men GMs in the same period is not a significant comparison.

For example, if, today, there are 800 active 2500+ GMs, whereas, in 2000, there were 400 active 2500+ GMs (and I think that, probably, the increase is even greater than that), the fact that, today, there are 13 active 2500+ women, whereas in 2000, there were only 10 active 2500+ women doesn't mean at all that women can more easily compete with "standard men GMs" (this would indeed rather mean the opposite...).
Petrarlsen Petrarlsen 5/30/2017 03:45
@ lajosarpad (2/3) :

- About the use of ratios (as, for example, "W / M = 0.91304997608799617407938785270206"), if I remember well, I read an article by a statistician, where it was explained that the Elo rating system is designed in such a way that to use such ratios for comparisons isn't possible.

As I am not a mathematician, and as I read this article (I think on ChessBase) rather long ago, I don't remember the details, but you can quite certainly check this easily yourself without using any exterior sources (as you are, precisely, a mathematician yourself...).

What seems to me certain is that if what I said is true (about the fact that a given difference - for example, 100 Elo points - has the same "impact" in every part of the "Elo spectrum"), it isn't indifferent to use ratios, like you, or differences in term of Elo points, like me.

For an example, we will admit that in 2020, there will be a 200 points difference between the best female players and the best male players, with 2600 for the best women, and 2800 for the best men, and that (supposing that there would be a very significant "Elo inflation" between the two, which isn't at all certain or even probable, but which I don't suppose is either impossible) in 2040, the difference between the best women and the best men will still be 200 points, but with 2700 (and not anymore 2600) for the women, and 2900 (and not anymore 2800) for the men.

With this example, the difference between men and women would be considered to be the same when taking into account the difference in terms of number of Elo points (200 points), but not in terms of ratios (the difference would seem to diminish : 0.9285714 in 2020, and 0.9310344 in 2040), thus, the two systems cannot be used indifferently, as the results obtained with each of them are not the same.

So, either, really, a given difference always has the same significance, be it, for example, a 100 points difference between 1000 and 1100 players or the same difference between 2700 and 2800 grandmasters, and it isn't possible to use ratios as you do, or, otherwise, a given difference hasn't the same significance in every part of the "Elo spectrum", but the one doesn't seems to be compatible with the other.

- You say that, in 2017, there where 13 women at 2500 or above, and 67 at 2400 or above, whereas in 2000, there where only 10 women at 2500 or above and 38 at 2400 or above.

This is true, but seems to me to be related not to a decrease of the gap between men and women, but to, either a global increase in level (for men AND women), or an "Elo inflation".

This because, if you do the same research for men's lists, you find even much more significant results :

* In July 2000 : 1 man at 2800 or above, 11 at 2700 or above, and 28 at 2660 or above (2660 because, in 2017, 2650 players don't figure anymore in the Top 100).

* In May 2017 : 4 men at 2800 or above, 43 at 2700 or above, and 95 at 2660 or above.

So, in these three categories (+ 2800, + 2700, and + 2660), there are more than three time more players in 2017, compared to 2000 !

And, today, there are 43 2700+ players, whereas, in 2000, there where only 33 2650+ players (so there are more 2700+ players in 2017 than 2650+ players in 2000 - a quite significant difference in my opinion).

This is in fact a much more considerable increase than the women's increase (10 to 13 and 38 to 67) in the same period.

So I think that, even this last element seems to show that the gap between top male and top women players is rather growing.

And, also, the fact (for example) that the number of female 2500+ players has grown from 10 to 13 between 2000 and 2017 would be completely annihilated if, for example, the number of male GMs has tripled in the same period - I don't know for the moment where to find this information, but I am nearly sure to have read that, in the last decades, the total number of GMs has increased considerably).
Petrarlsen Petrarlsen 5/30/2017 03:44
@ lajosarpad (1/3) :

- "I never said that the same male rating means a different thing than its female correspondent."

In fact, I didn't thought at all that you gave your opinion on this subject ; it was just a step in my own reasoning (and even if I agree with you when you say that men's and women's ratings aren't probably completely comparable, I don't think that the difference must be really significant, as there are no "men-only" competitions : there are "women-only" competitions, and open competitions, and as the women players can - and do - participate in men competitions, the two systems aren't closed one to the other ; by the way, otherwise, the same could be said about men players in very different countries, as, for example, the USA and Russia : an American player doesn't play, most of the time, in the same tournaments as a Russian player).

- "(...) with the difference that I have used 50 as seed size instead of 10, since it obviously yields a more reliable result."

On this, I don't agree.

I think that, indeed, it is also interesting to compare the 50 best men to the 50 best men, but I also think that to take the 10 best men and the 10 best women is better, if you want to really compare the level of the "top women" to the level of the "top men".

Why ? Because women's ratings drop MUCH quickly in the lists than men's rating (for example, in May 2017, the difference between the n° 1 and the n° 50 is 139 points for the men and 232 points for the women, and, in July 2000, the difference between the n° 1 and the n°50 was 216 points for the men and 278 points for the women ; in May 2017, the difference between the n° 1 and the n° 50 for the men is the same as between the n° 1 and the n° 12 - !!! - for the women, and, in July 2000, the difference between the n° 1 and the n° 50 for the men was the same as between the n° 1 and the n° 22 for the women).

So, to use 50 players instead of 10, for comparisons between men and women, is quite unfavorable to women, and doesn't, in my opinion, reflect the true difference between the very best men and the very best women.

And, in practice, the difference is quite significant : with you system, the difference between men and women is 233 points in 2000, and 263 points in 2017. With my system, the difference is 215 points in 2000, and 237 points in 2017. So there is a 18 points difference in 2000 and a 26 points difference in 2017, if you use your system or mine.

So I really don't think that to compare the 50 best men to the 50 best women is optimal...
lajosarpad lajosarpad 5/29/2017 10:49
@Petrarlsen

Now, let's focus on male and female ratings. Since ratings determine the relative strength of players towards each-other and not necessarily the absolute strength, in the case when there is a segregation between two large communities, who will never play each other, the same rating could mean a completely different level of play. ÉLŐ is meaningful precisely for the reason that it is self-learning. However, without teaching it by having many games between males and females it might lose from its meaningfulness. In our case there is no total segregation, as some women are playing in open tournaments against male players and therefore there is some realistic adjustment between the two groups, the two groups are partially segregated and therefore I cannot say for sure that a male player with a rating of X will roughly play at the same level than his female counterpart having the same rating. If this is the case in this moment, then it is due to the fact that the two groups play occasionally with each-other. I would, however watch a match between 50 males and 50 females with roughly the same ÉLŐ points, that would be much more telling than our opinions about how the same ratings are meaning the same in the two partially segregated groups or how they are different.
lajosarpad lajosarpad 5/29/2017 10:41
@Petrarlsen 2/3

Let's define the notion of gap. A gap between X and Y is the distance between X's closest point to Y to Y's closest point to X. In our case, the gap between male and female chess is the difference in abilities between the worst of the best male players and the best of the best female players. When the best of the best female players reach the consistent level of the worst of the best male players (for instance all the women players in their top 100 list will be above 2500), then there will be no longer gap. But even then women chess will be inferior to male chess if the best of the best women players will still be worse than the best of the best male players (it is theoretically possible that they will get better than us eventually, yet this possibility is far fetched at the moment). Or, to put it more simply with a geographical analogy, let's imagine two countries, maybe Germany and Russia. The gap between the two is their minimal distance and it is a bad approach to measure the distance between Germany's easternmost point and a point in Russia's coast to the Pacific Ocean. This kind of comparison is not really meaningful when we try to measure the gap between the two. The same is the case with women chess vs. male chess. If women players are able to compete in GM tournaments with decent results, then the gap is narrower than before, when they would have been whitewashed. I do not think anyone would seriously believe that Chiburdanidze, for example would be able to ever beat even at her best Carlsen. Yet, Hou Yifan could beat him once in a while.

So, again: I am not stating that a male rating means a different thing than its equal female rating and I repeat that comparing the very top with the very top will not answer the question of whether the gap between male and female chess is narrowing or not. For instance, June's list will probably say that the difference between top male chess and top female chess was decreasing compared to the May list. However, I find the relevance of this kind of comparison to be not very high. In my book if women players are able to win GM tournaments (and not necessarily top GM tournaments), then the gap between the performance of the two genders has gotten narrower compared to the times when the very best 5 female players would have not stood a chance against a normal GM opposition of 2500 on average, not mentioning super GM opposition. Yet, I think women chess will never have the same quality as male chess. Women have some "problems" on monthly bases, also, if they get pregnant, then they will get out of form and they tend to be more sensitive and therefore it is more difficult for them to get back to normal after a loss. Yet, I do not know the final answer of women's ability in chess versus male abilities.
lajosarpad lajosarpad 5/29/2017 10:20
@Petrarlsen 1/3

I think you misunderstood my reasoning. I never said that the same male rating means a different thing than its female correspondent. While this is a very interesting topic as well and I am not sure it has a simple answer, I did not address this. I have basically done comparison according to your idea of comparing the average of top female players to top male players (with the difference that I have used 50 as seed size instead of 10, since it obviously yields a more reliable result).

This

W / M = 0.91304997608799617407938785270206

means that in a given period (July 2000) the average rating of the top 50 women players was ~= 91.305% of the average of the top 50 male players, that is, if you take the men average for the top 50 and multiplicate it with W / M in the given period, then M * (W / M) will yield W, the Women average. The other number, namely 0.90366724598735230973696893575382 shows that this ratio has decreased, so the average of Women rating for the top 50 now is ~ 90.367% of the top 50 male players' average now. Again, these numbers are not representing anything like what is worth a male ÉLŐ compared to the same value in case of females. If there was any difference taken into account, the formula would have been different than W / M. And this kind of division is much more meaningful than calculating the difference as you suggested. When comparing, we need to decide whether we use ratio or difference comparators. Here ratio is superior, as being able to give a proportion is much more meaningful than the difference of averages. Knowing that 17 years ago the top females' average was approximately 91% of the male average, yet this proportion decreased to 90% now is telling me much more than a difference.

You also state that my opinion was not confirmed by the objective elements we found. While there are objective elements that show the contrary, there are objective elements which are confirming my proposition as well. For example

https://ratings.fide.com/toparc.phtml?cod=2

here you can see 10 players above 2500 and 38 above 2400

https://ratings.fide.com/top.phtml?list=women

here you will find 13 players above 2500 and 67 above 2400

which shows that there is an increasing number of women being able to compete at or near GM strength. Which means that the gap is getting closer, since there is an increasing number of women players being able to compete at or near GM strength, that is, there is an increasing number of female players who are able to play decent chess against the weakest of the strongest male competitors. The very fact that Hou Yifan played decently at the Grand Prix, Polgár was 8th in the world and invited to play directly for the open World Champion title, Polgár achieved bronze in the European Chess Championship not long before she retired, Hou Yifan was shared first at Gibraltar in the not distant past shows us that there are some women who are able to compete at the highest level and achieve better than par results.
Petrarlsen Petrarlsen 5/29/2017 04:39
@ lajosarpad and @ drcloak :

A new idea : I think that, in our previous exchanges, there are in fact two very different ideas :

1) Is the number of women playing chess at a "very serious level" (= in my opinion GM level, but this is indeed quite subjective) increasing or decreasing ? ( and to this corresponds directly, for example, lajosarpad's idea : "What I do consider to be relevant is that there is an increasing number of women playing GM level chess.", or my own idea to calculate a ratio between the total number of male grandmasters and the total number of female grandmaster)

2) Is the global level of "top-level female chess" increasing or decreasing ? (and to this corresponds for example lajosarpad's observation : "The rating gap between the very top male players and the very top female players has increased lately.", or to my own idea to calculate the difference between the average of the 10 best players in general with the 10 best women players at a given moment, and to use those results for comparisons between different periods)


I think that the "1" is about "quantity", and the "2" about "quality".


The results of the "1" tell us if there are more women who play very good chess than before, and the "2" if the best women are getting nearer the best men.

To go back to the beginning, we were discussing a sentence in one of lajosarpad's posts : "As a trend, the relative strength of women chess players compared to male chess players has been increasing and the gap between the two has been getting smaller."

I think that this is more a question of "quality" than "quantity" ("strength (...) increasing" and "gap (...) getting smaller" do seem to indicate that "quality" is the subject, and not "quantity").

As our data seem to show that the answer to the "2" is that the global level of "top-level female chess" is (slightly) decreasing, I think personally (and for the moment) that lajosarpad's sentence ("As a trend, the relative strength of women chess players compared to male chess players has been increasing and the gap between the two has been getting smaller.") hasn't been confirmed by the objective elements that we find.
Petrarlsen Petrarlsen 5/29/2017 04:38
@ lajosarpad :

"What I do consider to be relevant is that there is an increasing number of women playing GM level chess." For me, the obvious way to check this is to calculate the ratio between the total number of male GMs and the total number of female GMs. Cf. my previous post.

Perhaps one nuance : for this calculation, I think that it would very probably be better to keep only the active players (to have a result more closely linked to the "chess world of today").
Petrarlsen Petrarlsen 5/29/2017 04:37
@ lajosarpad :

- "There is an increasing number of women who are able to play in GM tournaments."

"Women players will get closer to male players when the strongest women players will get closer to the weakest strong male players, that is, see women beat male grandmasters, or hold them to a draw."

I frequently agree with you, but I must say that, this time, I am highly doubtful about your reasoning !

I am under the impression that this exact question is much less complicated than what can be understood from your post.

This is my own reasoning...

The Elo system and the Elo based titles attribution system for IM and GM titles seems to me to be quite sufficently reliable to give an absolute result, quite indepentenly from the fact that a player is a man or a woman. So a 2500 man exactly corresponds to a 2500 woman, and, as GM and IM title are Elo-based titles, a male IM or GM exactly corresponds to a female IM or GM.

So, for me, to compare top women players to "standard" male grandmasters, the simpler solution is to calculate the ratio between the total number of male grandmasters, and the total number of female grandmasters. For the moment, I don't know where to find these numbers (could somebody help ??), but the calculation itself seems to me to be extremely simple.

And I don't think either that there is a significant difference between comparing top women and top men players levels, on the one hand, and top women players and "standard" men GMs, on the other hand, because, as a GM is very approximately a 2500+ player, and an "absolute top level player" a 2800+ player - still quite approximately -, to compare top women players with top male players is approximately the same as comparing top women players with "standard GMs" + 300 Elo points.

So I think, globally, that to compare top level women to top level men, or top level women to "standard men GMs" is rather indifferent ; the idea is approximately the same, because the level of a "standard male GM" is very closely linked to the level of the top-level players (the link being, more or less, a difference of 300 Elo points).

Anyway, I don't know what the result would be of the ratio between the total number of male grandmasters and the total number of female grandmasters, but, taking all those elements into account, I don't quite see, at least at first sight, how the gap between top women players and "standard male GMs" could become narrower, if the gap between top women players and top men players is (slightly) widening...

- "W / M = 0.91304997608799617407938785270206" and "W / M = 0.90366724598735230973696893575382"

I am nearly sure that the Elo system works in such a way that a given difference (for example, 100 Elo points) have the same meaning in every part of the "Elo spectrum". If this is really true, it is better to retain the Elo points difference than the ratio between the two given Elos (for example, if I am right, 1400 to 2100 Elo represents the same difference than 2100 to 2800 ; with your calculation system, 1400 to 2100 would correspond to 2100 to 3150, a big difference with 2100 to 2800).

In the present case, the difference wouldn't be very important, but, nonetheless, still if I am right, it would be better to retain the Elo differences, namely, 232.72 Elo in 2000, and 263.34 Elo in 2017.
drcloak drcloak 5/28/2017 11:29
@lajosarpad

This is absolutely fascinating; I truly underestimated your power.
lajosarpad lajosarpad 5/28/2017 09:43
@Petrarlsen & @DrCloak

The script I have written to do the comparisons is

var context = document.getElementsByClassName("mainbody")[1].getElementsByTagName("table")[1].getElementsByTagName("tr")[1].getElementsByTagName("tr");
var sum = 0;
for (var i = 1; i <= 50; i++) {
sum += parseInt(context[i].getElementsByTagName("td")[4].innerText.trim());
}
console.log(sum / 50);

I am giving this to you to simplify the comparison, to not have to manually add the numbers and divide with 50. When there are women in the list, we just change the for cycle to

for (var i = 1; i <= 51; i++) {

and subtract the women rating from sum, like

var context = document.getElementsByClassName("mainbody")[1].getElementsByTagName("table")[1].getElementsByTagName("tr")[1].getElementsByTagName("tr");
var sum = 0;
for (var i = 1; i <= 51; i++) {
sum += parseInt(context[i].getElementsByTagName("td")[4].innerText.trim());
}
sum -= 2656; //Polgar
console.log(sum / 50)

The URLs are like this:

https://ratings.fide.com/toparc.phtml?cod=1

where the number at the end (1 in this case) is defining the list. I was using top 50 data because back in 2000 the women list contained only the top 50 women players. My opinion is that top male players got better compared to top women players, but top women players start to be good-enough to compete against male GMs.
lajosarpad lajosarpad 5/28/2017 09:36
@Petrarlsen

Many decades ago there were virtually no women players. With the appearance of Vera Menchik we can more-or-less say that women chess had been founded. Since the very first step of overcoming absolute 0, women chess has gone a long way. Menchik, Chiburdanidze were not real dangers for the absolute top and women chess was much much weaker than now, a woman beating a top 10 player would have been unimaginable back then. In 2005 Polgár entered top 8 and became a challenger for the World Championship by rating. She was the last at the San Luis tournament, but even then, getting there was a great individual achievement, which was never achieved by any women before. On the other hand, this performance was not repeated since then. However, I think it is an early sign that we might see a woman with the title some day. If we take a look at the rating list progress, we will observe two things:

1. The rating gap between the very top male players and the very top female players has increased lately
2. There is an increasing number of women who are able to play in GM tournaments

1. I have taken more comparisons, but I consider this to be the most significant:

July 2000
Men average for top 50 (excluding Polgár): 2676.48
Women average for top 50: 2443.76
W / M = 0.91304997608799617407938785270206

May 2017
Men average for top 50: 2738.84
Women average for top 50: 2475.5
W / M = 0.90366724598735230973696893575382

So we can say that in the last 17 years the very top female players' ratings did not increase at the pace the very top male players' ratings increased. We could argue based on the absence of Polgár, but Kasparov has retired in this period as well, although, Polgár's rating has increased female average than Kasparov's the male rating average. I would like to see a comparison to 1990s, 1980s and 1970s as well, as my comparisons are only meaningful to very recent times.

2. But even though the very top male players had an increase compared to the very top female players, I think this is not the right way to compare male chess to women chess. Women players will get closer to male players when the strongest women players will get closer to the weakest strong male players, that is, see women beat male grandmasters, or hold them to a draw. Some women are already capable to do so and their number is increasing. Hou Yifan's recent performances in tournaments against male players were pretty decent for women competing against top male players. Lázárné has held Anand to a draw. I think women chess is simply weaker than male chess in this moment, however, I see the potential of increasing number of women players who could perform decently in a GM tournament against males, a thing which was less imaginable for me 10 years ago. I will compare June's ratings to May and to 2000 as well and while I believe the numbers will appeal women chess in that comparison (since Hou Yifan will be 84th, "stealing" some points from male players), I do not consider that future result to be relevant either. What I do consider to be relevant is that there is an increasing number of women playing GM level chess. So the facts are not unanimously backing up my opinion, but the very thing we experience seeing tournament games and results were the foundations of my opinion. I could be wrong though.
drcloak drcloak 5/28/2017 08:26
@Petrarlsen

Very interesting analysis. What could be a factor, and something that is confusing things a bit, is cyclic in nature. I think what we have (based on your analysis and my intuition) is an overall lessening of the rating gap, then a period of time or cycle within that gap that it increases, then another overall lessening of the gap over a greater period of time, then an increase, etc.
Petrarlsen Petrarlsen 5/28/2017 01:17
@ drcloak and @ lajosarpad :

One last (for the moment...) comparison : The number of women with an Elo difference of less than 300 Elo with the average Elo of the 10 better players in the general list between July 2000 and May 2017 : In July 2000, there was 16 women players in this category ; in May 2017, 13.

So this comparison also tend to show a (slight) widening and not a narrowing of the gap between men and women, in chess...
Petrarlsen Petrarlsen 5/28/2017 01:00
@ drcloak and @lajosarpad :

I did the calculations for one comparison (I have no time for the moment to do more...) : the difference between the average of the 10 best players in general with the 10 best women players, in May 2017 (the last list), and in July 2000 (the first official list that can be found on the FIDE website).

This is obviously only one isolated result, but it certainly doesn't show a narrowing of the gap between men and women : the gap was 215 points in 2000, and is, now, 237 points. So, taking these two results, it even widened a little...

I don't know if lajosarpad has more extensive data on this subject ??
Petrarlsen Petrarlsen 5/27/2017 10:46
@ drcloak : Another idea : The number of women with an Elo difference of less than 300 Elo with, not, this time, the World Elo Number One, but with the average 10 better players in the general list (for example, now, with an average of 2798 for the Top 10, there would be 13 women in this category).
Petrarlsen Petrarlsen 5/27/2017 10:36
@ drcloak : Another possible element : The number of women with an Elo difference of less than 300 Elo with the World Elo Number One (for example if the global n° 1 is 2850, the number of women above 2550 would be retained for this comparison ; for example, at the present moment - Carlsen being 2832 -, there would be 9 women in this category).
Petrarlsen Petrarlsen 5/27/2017 10:18
@ drcloak :

About the male-female chess strength gap, your argument does tend to show that this gap doesn't tend to get smaller. But there can also be other data that go in the opposite direction (for example, the rating difference between the 10 better male players and 10 better female players since the beginning of the Elo rating system - I don't know anything at all about this element, but, if nobody knows for the moment, I will certainly search for this some time or other ; perhaps also, for example, the Elo average difference between the best man and the best woman player calculated by periods of 10 years, the interest of using a rather long period of time being that it would tend to diminish the influence of very special players - as, for example, Polgar or Kasparov - in the global result).
drcloak drcloak 5/27/2017 08:48
@Petrarlsen

Consider: There have only been 3 women that have broken into the top 100 overall in the world, and they are from 3 different generations. Maia Chiburdanidze, Judit Polgar & Hou Yifan. Although Yifan is not currently in the top 100, she did achieve it in 2014 for a time.

In order for me to be convinced that the male-female chess strength gap is closing, there needs to be more than 1 female per generation to break into the top 100 overall in the world.