FreeRepublic wrote on 09/30/23 at 15:15:46:

Fischer definitely seemed to favor the positional lines. A search of the Mega Database fails to find any examples of him playing 7.Qg4 in a mainline Winawer. The closest he came was 7.Qg4 in the Armenian Variation in his famous game vs. Tal at the 1960 Leipzig Olympiad. Another game with Fischer playing a Qg4 maneuver that readily comes to mind is Fischer-Hook (in the eponymous Portisch-Hook variation) at the 1970 Siegen Olympiad . Fischer's most frequent adoption of Qg4 was in the several games where he employed 4.a3 vs. the Winawer, most notably in Fischer-Kovacevic, Rovinj/Zagreb 1970. However, as noted, not of these is the mainline Winawer with 7.Qg4.

Some comments by Edmar Mednis in his How to Beat Bobby Fischer may shed some light on Fischer's avoidance of 7.Qg4. In his annotations to Fischer-Geller, Monte Carlo 1967, Mednis describes the situation after Black's 18...Bb7 as "a crazy, wild, complicated, unbalanced position...Yet there is no strategically clear way to proceed with the attack. What is required is an intuitive juggling of a number of tactical opportunities. This means the position is not one where Fischer is at his best: yes, he's good in it but not great. It is truly a "Tal type" position..." While the Fischer-Geller game was a Sicilian, the description "crazy, wild, complicated, unbalanced" could easily fit the Winawer Poison Pawn. 

At another point, when commenting on the Tarrasch Variation vs. the French, Mednis expressed the opinion that it "allows White to retain a small, but riskless initiative for many moves. This is exactly the situation that Bobby strives for with White, thus the Tarrasch should suit him to a T". 

Combining the two Mednis comments leads to the thought that Fischer avoided the 7.Qg4 variation due to its murkiness and because he apparently thought he would have better control of things with less risk in the positional lines.  Also, Fischer famously called the Winawer "anti-positional", so what better way to combat it than the positional variations. 

One last Mednis observation: he noted that Fischer ventured the Tarrasch only once (in Fischer-R. Byrne, US Championship 1965) despite its stylistic fit. According to Mednis, Fischer persisted in playing against the Winawer because it was "the only opening where the mature Fischer still shows signs of 'youthful stubbornness'".

FreeRepublic wrote on 09/30/23 at 15:15:46:

Apparently the positional lines were the main lines in Fischer's heyday since they were given at the very end of the ECO section on the Winawer. That's based on my impression that ECO organizes variations so that early deviations and sidelines come first, with the absolute main line coming at the end. The positional lines were assigned C19, the last of the ECO codes for the French, when the code system first appeared.  I was surprised to see ECO Volume C (5th edition 2006) had switched to 7.Qg4 as C19, and had reassigned C18 to the positional lines.  However, not everyone has followed suit. For example, ChessBase still has the positional lines classified as C19. Apparently the ECO folks made the switch in recognition of 7.Qg4 becoming the main line. I wonder if other changes are on the way. For example, the last code for the Queen's Gambit, D69, corresponds to the ancient Orthodox Defense, complete with Capablanca's freeing maneuver, which just isn't played much any more, at least at the top levels. And C99 is the Chigorin Variation in the Closed Ruy. Perhaps this should be superseded by the Berlin 

Nernstian59 wrote on 09/20/23 at 22:35:44:


I seem to recall that Fischer was suspicious of 7Qg4 and preferred positional lines. Theory does not stand still and 7Qg4 became the main line. Ponomoriov has a White repertoire against the French Defense at modern-chess. He continues with 7h4. More recently GM Kalyan Arjun released his repertoire against the French (modern-chess). From the description, his choice is 7a4 Qa5! 8. Bd2 Nbc6 9. Nf3. I only consider and play the French from time to time. So for me, it seems that a line out of vogue for decades has come back into fashion.

Nernstian59 wrote on 06/07/23 at 21:35:27:


Thanks for the excellent summary. The positional lines have always been interesting.

kylemeister - Thanks for those informative quotes! They do show that 10 ...f6 was fairly credited to Korchnoi in the years right after its introduction.  For whatever reason, that doesn't seem to be the case with more recent opening works. The comment from Timman is particularly interesting since it directly addresses the misgivings about opening the position expressed by Moles.   

kylemeister - While writing my previous post, I had a nagging feeling that I had seen an annotated French Winawer game where Kavalek played White. My recollection was too vague to allow me to find that game, let alone to know if it was the one with Uhlmann, so I didn't mention it in my post. Now that I've seen Kavalek's article (BTW - Thanks for the link!), I'm fairly sure I looked it up years ago to see the new analysis of Botvinnik-Capablanca AVRO 1938.

When you noted that Timman was Kavalek's second for the Manila Interzonal, I immediately wondered if Timman's experiences with the Winawer had been useful in preparing Kavalek for his game with Uhlmann.  Sure enough, Kavalek confirms this in his article.

Another interesting nugget from the article is that Kavalek credits Korchnoi with introducing 10...f6, the current main move. It's certainly plausible since the 1976 Timman-Korchnoi game I mentioned in Reply #16 is the earliest instance of the move in Megabase. However, I haven't seen the attribution of 10...f6 to Korchnoi anywhere else.

kylemeister - Thanks for pointing out those Uhlmann games with 10...c4. The one against Kavalek from 1976 is an interesting mix of new (Giri's plan of ...Kb8 + ...Nc8 and defending the g7-pawn with a rook when it's attacked with Nh5) and old (Botvinnik's snatch of the a4-pawn from the 1940s). With 20/20 hindsight, it appears that Uhlmann's transfer of a knight to g6 wasn't a good idea. It became a target and was later lost. Lc0 came up with a different knight maneuver - playing it from c6 to a7 (after ...a7-a6) and then to b5 to gang up on the c3-pawn.

I wonder about the reason for Uhlmann's reversion to 10...c4 in the 2000 game against Schulte. Did Uhlmann want to avoid a tactical clash against an opponent who was close to thirty years younger? Or did he want steer clear of a theoretical battle and head for a maneuvering game in a position with which he had past familiarity?  Whatever the reason, I liked how Uhlmann repositioned his knight with ...Ne7-f5-d6.  I don't recall seeing that way of getting to d6 before.

Out of curiosity, I got out my copy of the French Defense book published by RHM in 1975. Gligoric is given as the author, although Uhlmann is credited as his collaborator.  They give 10...c4 as the main line, just as Moles did in his book, which appeared in the same year.  The analysis of 10...c4 in the RHM book follows Moles' preferred line (see my post in Reply #16) up to 15.g3. It then deviates with 15...f6.  The line ends here with the comment "with an unclear position". The engines seem to think that Black's position is OK, evaluating it as =.

The RHM book doesn't mention 10...f6, although that's understandable given that it was an untried move at the time and that the authors were trying to fit the entire French Defense in a single volume. Moles had the luxury of only covering the main line of the Winawer.  Even the secondary Winawer lines were left to his second book.

While looking through The French Defense Revisited, I noticed that Jacimovic + Zlatanovic ("J&Z") offered an interesting recommendation in the 7.a4 positional Winawer variation, which begins with 1.e4 e6 2.d4 d5 3.Nc3 Bb4 4.e5 c5 5.a3 Bxc3+ 6.bxc3 Ne7 7.a4. J&Z then follow mainstream theory for a few moves: 7...Qa5 8.Bd2 Nbc6 9.Nf3 Bd7 10.Be2. Now their suggestion is 10...c4 instead of the well-established 10...f6. 

I was puzzled about this recommendation because I had recently read Steve Giddins' account of the historical evolution of the 7.a4 line in his The French Winawer Move by Move. The story starts with Botvinnik's successes with ...c5-c4 in the 1940s.  Giddins then notes that it took quite a while for White players to develop an effective counter. By the early to mid-1970s, White began to meet 10...c4 with 11.Ng5 h6 12.Nh3, with the intention of playing Nf4-h5 to force a weakening of Black's kingside pawn structure. Giddins cites two Planinc-Timman games (Wijk aan Zee 1974 and Amsterdam 1974), where Black tried 12...Ng6 to keep the white knight out of f4, only to beaten handily after 13.Bh5. In both games, Black tried to reinforce his g6-knight with 13...Nce7, but White proceeded with a winning plan that consisted of Qe1 to guard c3, allowing Bc1-a3, combined with play down the b-file and an eventual advance of the f-pawn to threaten the knight on g6 and hit the base of Black's pawn chain. Giddins notes that these two defeats, and similar ones, caused Black players to become dissatisfied with 10...c4, resulting in a switch to the more dynamic 10...f6.

John Moles' landmark work The French Defence - Main Line Winawer was published in 1975, right around the time that the transition from 10...c4 to 10...f6 was occurring. Thus, it's informative to see Moles' opinions on these two lines. He has 10...c4 as his main move, with the comment "Closing the centre before embarking on counterplay on the wings, in good classical style". Moles describes 10...f6 as a suggestion by (Rolf) Schwarz and only gives a brief treatment, saying that after 10...f6 11.c4 Qc7 12.exf6 gxf6 13.cxd5, the opening of the position will enhance the strength of White's bishop-pair. 

Moles cites the two Planinc-Timman games, but didn't care for Timman's 12...Ng6, marking it ?! and giving the reply 13.Bh5 an exclam and an evaluation of ±. Moles preferred to allow 13.Nf4, meeting it with 13...g6 ("Black stops Nh5, saddling himself with a suspect pawn structure but gaining time to secure active counterplay") 14.h4 Rdg8 15.g3 Kb8! 16.Bg4 Nc8 17.h5 g5 18.Ne2 f5 19.exf6 Nd6 20.f4 Ne4!, as in the game Bannik-N.Levin, Ukraine Team Ch, Ternopol 1965. Moles' evaluation: "a complex position with about level chances.". Stockfish gives Black a slight edge, but also points to improvements for White on moves 17 and 18. 

Moles' approach didn't seem to catch on, but 10...f6 did, apparently rather quickly, since Giddins' illustrative game for this move has Korchnoi playing it successfully in 1976, just two years after the two Planinc-Timman contests. Korchnoi's opponent? - none other than Timman, who had the misfortune of now losing a positional Winawer game with the white pieces! (Timman-Korchnoi 3rd match game Leeuwarden 1976). Korchnoi's use of 10...f6 in his 1977 Candidates match with Spassky may have also accelerated its adoption as Black's main move, a status which has seemingly endured to the present day.

This impression is reinforced by 10..f6 being the recommended move in the Black repertoires offered by Watson in the 4th edition of Play the French, by Berg in his GM Repertoire book and by Giri in his Chessable course. However, a bit more research shows some grumbles of discontent. The most pronounced criticism of 10...f6 comes from Miedema, who choose not to recommend this move in his The Modernized French Defense (2019) because of the line 11.c4 Qc7 12.exf6 gxf6 13.cxd5 Nxd5 14.c3 0-0-0 15.0-0 Rhg8 16.a5!N e5 17.a6 Bh3 18.g3!!, when he could not find a way to prevent White from gaining an advantage. Miedema's comment about the position arising from 10...f6 is interesting: "I felt there was something fishy about an open position with the bishops and opposite-side castling." He seems to have the same misgivings that Moles had back in the 70s.  Fortunately, Giri's course offers an antidote for Miedema's line, although it's a rather concrete variation that was likely produced via engine-assistance.  

Even back in 2013, Watson, while recommending 10...f6, spent a couple pages to present "a way for Black to play more safely and solidly". His suggestion? - 10...c4! In other words, the old main line. Watson notes that this approach "goes back to some old moves'. His main line is 10...c4 11.Ng5 ("This has long been regarded as the theoretical drawback to 10...c4" - Watson) 11...h6 12.Nh3 Ng6 13.0-0 0-0 14.Qc1 Rfb8. He also covers 13.Bh5 (Planinc's killer move), answering it with 13...Nce7 14.Qg4 0-0. Watson gives this comment: "This is the point! 10...c4 has mostly been associated with ...0-0-0, especially in the very few old games with 12...Ng6. But I think the king is happy enough on the kingside, and this opens up possibilities on the other side of the board" This is a new approach (at least for me), and one can see how 14...Rfb8 supports ...b5, seeking play on the queenside. A search of Megabase fails to find any games where Black went for Watson's approach.

In his Chessable course, Giri has the following comment on the position in the 7.a4 Winawer after 10.Be2: "Black ... is always choosing between the direct ...f6, allowing a mess after c3-c4, and starting with ...c5-c4 first, which is slower but more reliable". So there seems to be a consensus growing that 10...f6 is getting a bit hot too handle for some, and that something more solid would be desirable.  While devoting the bulk of his analysis to 10...f6, Giri also offers some coverage of 10...c4, meeting 11.Ng5 with 11...h6 12.Nh3 0-0-0, allowing White to execute his threat with 13.Nf4 Kb8 14.Nh5. Now he suggests 14...Rdg8 to guard the g7-pawn without weakening Black's pawn structure. After 15.0-0 Nc8, Black can play the ...f5 advance, meeting exf6 with ...g6 and picking up the pawn on f6 later. Also the knight was played to c8 so it can now go to d6. Note that Giri uses the same ...Kb8 + ...Nc8 maneuver that Moles suggested back in the 70s!

So it would seem that J&Z's recommendation of 10...c4 is consistent with this trend of bypassing a tactical melee and seeking more strategic play in the 7.a4 Winawer. However, J&Z don't say anything to this effect. They just note that 10...c4 leads to well-known positions with White a tempo up, although they don't feel that the tempo is important. Presumably, J&Z are referring to the positions where White plays 10.Bd3 instead of 10.Be2, and now 10...c4 11.Be2 is the same position as 10.Be2 c4, but with Black a tempo up.  It's noteworthy that J&Z don't cover 11.Ng5 in response to 10...c4. Watson and Giri are apparently more aware of the theoretical history of the line since they devote some analysis to this relevant move. J&Z cover 11.h4, 11.0-0, and 11.Nh4, which are reasonable enough ideas, but neglecting the most theoretically challenging line (and a top choice of the engine) is a disappointment.

Perhaps some idea of how J&Z would handle 11.Ng5 is given in the analysis of another positional Winawer variation, the one where White plays 7.Nf3 instead of 7.a4. One of J&Z's lines continues 7...Bd7 8.Bd3 c4 9.Be2 Ba4 10.Ng5 h6 11.Nh3 Nbc6 12.Nf4 and now the amazing 12...Kd7! intending the famous Yusupov queen transfer to the kingside with ...Qg8, possibly with a subsequent ...Qh7 to add to the pressure on the c2-pawn. J&Z say that the black king is quite safe on d7, though a search of Megabase shows that no one has been brave (or foolhardy?) enough to try this move in this specific position.

It's interesting that 7...Bd7 was Moles' recommendation against the positional 7.Nf3 Winawer. And since J&Z also suggest Moles' 10...c4 in the 7.a4 line, it seems as though the two authors have opted for a decidedly retro approach to handling the positional Winawer. J&Z might feel justified in this decision since it provides the opportunity for Black to play a more strategic game, one that would fit their target audience of club players better than long, concrete engine lines.