7... exd4
8. Nxd4

This left:

click for larger view

As Tarrasch rightly said, White is surely better here, but Black's position is sound and not without resources. The lost position into which Lasker soon fell cannot be attributed to his choice of opening.

8... 0-0

In Game 4 of the match, Lasker played 8...NxN which Wilson called "am improvement." While, given what followed here, it is not surprising that Lasker looked for something better. But as will be seen 8..0-0 was not the source of Lasker's troubles in this game and is probably best here.

9. NxN

Winter noted that 9. BxN is more usual here, but explained Tarrasch's move as based on Tarrasch's "great affection for Bishops."

Tarrasch acknowledged that "It is difficult to determine the best move here." He later claimed that 9. Bg5 would have been very strong, but his analysis of that move was dreadful: 9. Bg5 Nxe4? [this loses straight off; Black is fine--and better than in the game--with 9...h6 or 9. Re8] 10. BxB NxN (10...NxB is perhaps an improvement, but Black is lost anyway) 11. BxQ? [missing the win with 11. NxN!] 11...NxQ 12. BxN? (this lets Black off the hook. White can probably still win here with 12. Be7!) bxB 13. Be7 winning the exchange as Tarrasch says but after 13...Nxb2 14. BxR KxB 15. Rab1 Nc4 Black emerges with two pawns for the exchange and reasonable chances of holding the game.

Not one of Tarrasch's best instances of analysis!

Reti conjectured that Tarrasch played 9. NxN because he had "achieved a magnificent victory" with this move against Steinitz at Vienna 1898. But Reti's further claim that Lasker "refuted" 9. NxN here is quite wrong. Lasker got a decent game with his response, but even then White was better.

Probably best for White here is 9. a4 and if then 9...a6 the trade of two sets of minor pieces on c6 or just 10. Bc4, in either case leaving White with the somewhat better game.

After 9. NxN, the position was:

click for larger view

9... BxN

Tarrasch said this was better than 9...bxB as played by Steinitz in the above-references game, since it "gives greater freedom to Black's game." But many commentators strongly disagreed.

Hoffer, knowing that Lasker soon afterward got in big trouble, claimed that Steinitz's 9... bxN was "preferable and perhaps necessary, keeping the Bishop to prevent White's Knight being posted at f5." Winter, who seconded Hoffer's analysis, reported that 9...BxN was "widely condemned." But Reti sided with Tarrasch on this one, saying that "After the text, White cannot in the long run stop his opponent from easing his game by ...d5."

For what little it is worth (software can be a treacherous guide in evaluating opening strategy), Fritz slightly prefers 9...bxN while Stockfish gives the text the nod.

10. BxB bxB

click for larger view

11. Ne2

"A clever little move. If 11...Nxe4?? Nd4 etc." (Schroeder).

"Well played." (Hoffer)

"Since Black's ...d5 could not be stopped anyway White indulges in a bit of finessing." (Reti)

"En route for the important square [f5]." (Winter)

"A good move that leads to an attacking formation for White. The g7 square is the target." (Tarrasch)

"This maneuver was new to me...White gains time to place the Knight advantageously." (Lasker)

The text is certainly subtle and apparently a surprise for Lasker. Objectively, 11. Qd3 or 11. Be3 were at least as good.

The position after 11. Ne2 was:

click for larger view

11... Qd7

"?"--(Tarrasch)

Tarrasch suggested that 11...d5 might be better, giving the line 12. e5 Ne4 13, Nd4, but after this move Black--Tarrasch notwithstanding--would be OK with 13...Qd7 (and not Tarrasch's alternative 13...Bc5 which would run into 14. Be3 with a big edge for White. But after 11...d5 White can improve with 13. Be3 with the superior chances.

Reti agreed that 11...d5 would have been an improvement, and said that Lasker avoided this "simpler" move because "true to his complicating style" he wanted to develop his Rooks first. Reti said he agreed with Lasker's move and his idea here, but contended that Lasker now got into trouble by incorrectly posting his Rooks on d8 and e8 instead of on b8 and d8.

Lasker said he rejected 11...d5 because of 12. e5 Ne4 13. Nd4 Bc5 14. Be3, missing--as did Tarrasch--the superiority of 13...Qd7 in this line.

Lasker could have avoided all these issues (and the problems that followed after his actual move 11...Qd7) with 11...Re8 here, which would have left him in decent shape.

12. Ng3

12. f3 would also have given White the edge, but the text--as will be seen--had more than a drop of poison in it, the position after 12. Ng3 being:

click for larger view

12... Rfe8

"?"--(Reti)

"A mistake. If the Rook moves at all it must go to d8." (Lasker)[I quote the translation of Lasker's remarks as given here by <TheFocus>].

"Black now has a very cramped game." (Tarrasch)

Winter also disliked Lasker's move since Tarrasch's response did not allow Lasker to complete his plan to play Bf8; g6; and Bg7.

The text was indeed hardly best, though hardly an sort of catastrophic mistake. Best was probably 12...Qe6.

13. b3

click for larger view

13... Rad8

This move, which further cramped Lasker's position, has escaped criticism. But surely 13...Qg4 or 13...Qe6 were better.

14. Bb2

click for larger view

This, as it turns out, was the critical position in the first stage of this game.

Lasker was unquestionably worse at this point. But was he lost?

Reti thought so, and said that Lasker was here in a "hopelessly restricted position."

Tarrasch called Lasker's position "beleaguered."

Lasker himself said that he did not believe he had any satisfactory move here.

14... Ng4?!

"?"--(Tarrasch)

"An error..." (Tarrasch)

"Though this move is unsatisfactory, Black has none better." (Lasker)

"Simply a blunder." (Hoffer)

"...Black is in a hopelessly restricted position...[so] that great psychologist [Lasker] chooses the right moment for mixing things up. His next move looks like an oversight, but in point of fact it is a very shrewd sacrifice of a Pawn, aiming at the isolation of the White e-pawn..."

"Had anyone else but Dr. Lasker made this move it would be put down as a blunder, but White has already secured a very strong game of the type which he played perfectly, and it is quite possible that the champion deliberately made the move in order to bring about complications in which his opponent might go astray." (Winter)

"...a psychological masterstroke." (<birthtimes>)

"A blunder, Full stop." (<keypusher>)

What a collection of differing assessments. Confused? Should we try to psychoanalyze what was going on in Lasker's mind? No, I think not.

I reach the same conclusion that <keypusher> did, but let's take this one step at a time. There are a few points on which I hope we can agree:

1) If Lasker was not lost before this move but was lost afterwards, then his move was a mistake.

2) If Lasker overlooked Tarrasch's crushing reply [15. Bxg7], then his move can be called a blunder. On the other hand, if he saw 15. Bxg7 coming and played 14...Ng4 anyway because he overlooked the strength of 14...Qe6, then, once again, Lasker erred and his move can be called a blunder

If these points are granted, then perhaps we can answer the questions posed concerning 14...Ng4.

So, was Lasker lost before this move. Having read all of the analysis on this site and by other commentators, and having conducted deep searches with Fritz15 and Stockfish, the answer is NO. Lasker is not lost with 14...Qe6. As <Pawn and Two> and <keypusher have shown, and as my analysis has confirmed, Lasker would have had a playable--though admittedly difficult--position after 14...Qe6.

<crwynn> suggests that Black is lost after 15. Qd4. But such is not the case (as Stockfish and Fritz confirm on very deep searches and as now seems apparent to me). After 15. Qd4 Black can play (following the analysis of 15...d5 (15...c5 also draws [e.g., 16. Qc3 d5 17. exd5 Qxd5 18. Re5 Qc6], but I agree with <Pawn and Two> and <keypusher> that 15...d5 is simpler and better) and now if:

(A) 16. exd5 then 16...Qxd5 [much better than 16...Qd7 which is much more difficult for Black] and now if:

(i) 17. Qxa7 Ng4 18. Qa4 (or 18. c4 Bc5 19. cxQ BxQ 20. Nh1 [best] RxR+ 21. RxR cxd5 with only a small edge for White.

(ii) 17. Qa4 Qc5 18. Rac1 h6 and Black should be able to hold.

(iii) 17. Qh4 h6 18. Re5 Qd7 and Black should survive.

(B) 16. e5 Nd7 17. Qxa7 Bb4 18. Re2 Nc5 19. a3 Ra8 20. Qxc7 [or 20. axB RxQ 21. RxR Ne4 with only a small edge for White in this wild and crazy position] Nxb3! 21. cxN Bc5 22. Nf5 Ra7 [22...Rec8 probably also holds for Black] 23. Ne7+ RxN 24. Qd8+ Re8 25. Qg5 Be7 26. Qd2 c5 with excellent counter-play for the pawn minus.

In sum, Lasker was NOT lost and could have held the game with 14...Qe6. Since, as will be shown, Black is objectively lost after 14...Ng4; it follows that Lasker's move was a mistake.

Moving on to the next question, did Lasker see 15. Bxg7 coming? Here, certainty is not possible, but since even I saw the move and Lasker was obviously much stronger than I am or ever will be, I must assume Lasker saw it too. So does this mean that Lasker was using psychology instead of chess?

Nope. From Lasker's comments, it appears that he either overlooked 14...Qe6 or misjudged its strength. In either case, his move must be classified as an objective blunder.

So now let's return to the actual game after Lasker's 14...Ng4. The position was now:

click for larger view

15. Bxg7!

click for larger view

"!"--(Schroeder)

"This powerful move Black must have overlooked." (Hoffer)

If now 15...KxB White wins back the piece with 16. Nf5+ followed by 17. QxN with an overwhelming attack in addition to his pawn plus.

15... Nxf2

click for larger view

Arguably the best practical chance, but Tarrasch should have brushed this aside with ease.

16. KxN?

While this may still leave Tarrasch in a theoretically winning position, the killing move--as every commentator has pointed out--was 16. Qd4. Black then has no satisfactory recourse: 16...Ng4 17. Nf5 [17. Nh5 also wins] Qe6 [better than 17...f6 18. Rf1] 18. Qc3 [18. Qxa7 probably also wins, but 18. Qc3 is stronger; 18. Re3 c5 throws away the win for White]. For a discussion of a deeper search of this line, please see the analysis on this site by [Pawn and Two].

16... KxB

After 16. KxN, the position was:

click for larger view

Many of the contemporary commentators thought that Tarrasch had now missed his chance. But much of the analysis on this site has brought to light demonstrations that Tarrasch--despite missing the best line on move 16--still had much the better game and probably a theoretical win or at least very strong winning chances.

17. Nf5+

<keypusher> has shown that 17. Qd4+ is inferior to the text (refuting contrary analysis of Hoffer and Napier). <keypusher> also shows that 17. Rf1 is not best.

17... Kh8

click for larger view

18. Qd4+

Gunsberg preferred 18. Re3, but this move looks silly if Black responds 18...Bg5.

18... f6

Obviously forced to avoid immediate checkmate, and leaving:

click for larger view

19. Qxa7

It must have been tempting to grab the a-pawn, but 19. Rad1 as given by <Pawn and Two> was better, as was 19. Re3. The text gave Lasker counterplay as will be seen.

White has good winning chances after 19. Rad1, but perhaps not a theoretical win. Black's best then is 19...c5 as given here by <Pawn and Two>. White would likely then play 20. Qc3 and then Black can not play 20...Bf8? [as recommended by <Pawn and Two> since after 21. Qxf6+ Bg7 White would not play 22. Qh4 as appears in the line given by <Pawn and Two. but the simple 22. QxB+ QxQ 23. NxQ KxN 24. e5 with a clearly winning double Rook ending. Better for Black then would be 20...Qe6. Perhaps Black can then hold the position, but the White Knight on f5 would make that--at best--a horrendously difficult task.

But Tarrasch still had much the better of the game and perhaps still winning chances after 19. Qxa7

click for larger view

Most of the older commentators thought that Lasker had at least equalizing counterplay. Typical was the comment by Reti (with which Lasker essentially agreed):

"Tarrasch may well be pardoned for having flattered himself at this juncture for not being merely a very healthy pawn up but for having a promising attack as well against Black's dangerously exposed King. But this soon turns out to be a fallacy. Owing to the isolation of the e-pawn, e5 has become a very strong square whence the White attack will soon be brought to a standstill."

19... Bf8

"!"--(Reti)

"Threatening Rxe5 followed by QxN+." (Reti). See, similarly, Tartakower and Winter.

20. Qd4 Re5

"!"--(Reti)

click for larger view

Reti, though conceding that White still "has the better of it," contended that Black's edge in the center gave Lasker an opportunity to drawn on his skill in dynamic play "which as dogmatic a player as Tarrasch was ill equipped to cope with."

This description of so great a player as Tarrasch was grossly unfair, not to mention the fact that at this point White may well still enjoy a theoretically winning position.

As will be seen, Lasker erred in the upcoming play, and it was perhaps time pressure that may explain Tarrasch's missing some excellent chances in what ensued.

21. Rad1

Tarrasch had a difficult choice here. He later said that 21. g4 was best, but that seems doubtful. Following the line given by <Pawn and Two>: 21. g4 d5 [anything else loses] 22. Rad1 Bd6 23. h3 Qe6 24. Kf3 and now instead of 24...dxe4+ suggested by <Pawn and Two> 24...Re8 piles up on the e-file and allows Black to hold. (See also the analysis of 21. g4 here by <crwynn>).

Best here for White was 21. Nh4, as given by <Pawn and Two> and many years ago by Winter. If then 21...d5 (Winter's move) 22. Nf3 Bd6! ( the best try as given by <Pawn and Two> and certainly not Winter's 22...Rxe4 which gets smashed by 23. RxR dxR 24. Qxd4) 23. NxR BxN 24. Qe3 BxR 25. RxB and White should win the ending (for which see the longer analysis given here by <Pawn and Two>).

But, as <Pawn and Two> has also pointed out, Black can do better with 21...Bg7 after which--as discussed at length by <Pawn and Two> here, White has a much tougher time but good--but not certain--winning chances.

Tarrasch still had much the better of the position even after his likely weaker 21. Rad1:

click for larger view

21... Rde8

Hardly best. As both <crwynn> and <Pawn and Two> have shown, Lasker should have played 21...d5! after which it does not appear that White can win against best play.

After 21...Rde8, Tarrasch now had a strong line that he missed in the following position:

click for larger view

22. Qc3

<Pawn and Two> and <crwynn> suggest 22. g4, but then Black seems to able to survive with 22...h5 23. h3 Qh7 or 23. gxh5 d5.

White's best winning chances here lay in 22. Nh4, which does seem to lead to a likely win for White: 22. Nh4 c5 [22...Bh6 may be a better try for Black but White still seems clearly for choice] 23. Qc4 [23. Qc3 may be even better than the move given by <Pawn and Two> Qg4 24. Nf3 Rxe4 and now White wins not with 25. RxR as suggested by <Pawn and Two> but with 25. Qf7! to which I am unable to find a tenable defense for Black.

Even with 22. Qc3, Tarrasch still had the better of the game:

click for larger view

22... Qf7?

The older commentaries pass over this lemon by Lasker without a peep of disapproval. But, as been discussed on this site [see e.g., <Pawn and Two> and <crwynn>, Tarrasch would have had decent chances of survival (against best play) with 22...d5!

After the text, Tarrasch should have put Lasker away in this game, the position now being:

click for larger view

23. Ng3

This move probably did not blow the win for White (yet!), but it clearly was not best.

Many earlier commentators and several contributors on this site have claimed that 23. Qxc6 was best. But this is wrong. Black has chances to survive after 23...Rc5 24. Qa6 Rxc2+ 25. Re2 Rc3 26. Rd4 [26. Kg1 may be better, but even then Black may well be able to hold on: 26...Re5 27. Qa8 Re8 28. Qd5 QxQ 29. RxQ c6 30. Rdd2 d5 though it will be dicey for him] Qh5 27. Rc4 RxR 28. QxR Re5 with some counterplay that may help balance against White's extra pawn.

But even if White can indeed win, there must be something better than 23. Qxc6.

<crwynn> proposes 23. g4, and that attacking move does indeed appear to be an improvement on 23. Qxc6; e.g., 23...Rc5 24. Qd3 h5 25. h3 Qh7 26. Re2 Rce5 27. Qf3 and Black is in trouble.

But better still is 23. Qf3! which according to <Pawn and Two> was suggested by Victor Korchnoi. Unfortunately, Korchnoi's follow-up to his brilliant idea seems flawed: 23...d5? 24. exd5 cxd5 25. RxR RxR 26. Re1 after which White is certainly winning. But Black can improve with 23...Qd7 [or with 23..Qg6 or 23...Qg8 for that matter]. Even then, however, White is winning after 24. g4! after which it is hard to find a semblance of a decent line for Black.

But let's get back to Tarrasch's actual 23. Ng3 which was certainly not a blunder and at a minimum keep White close to a win, the position now being:

click for larger view

23... Bh6?

Another weak move by Lasker that was ignored in the older commentaries.

Tarrasch and Lasker both praised the threats presented in this move [see <keypusher> and <TheFocus> for their reasoning. But the move was hardly best.

Only <Pawn and Two> suggests an alternative, but his 23...d5 also leaves White in control after 24. exd5 cxd5 25. RxR RxR [not 25...fxR+ 26. Kg1 d4 27/ Qd3 as in the post of <Pawn and Two> after which White is clearly winning] 26. Kg1 d4 27. Qf3.

23...Rc5 is a better try for Black, and best is 23...Qg6 (Stockfish's move).

24. Qf3

"Presumably to prevent Bf4. But Black now has a compensating attack for the pawn minus, White's King being in an exposed position. 24. Re2 was necessary." (Hoffer)

Wrong on both counts. If 24. Re2, Black is very much back in the game with 24...Bf4. Moreover, and as will be seen, after the text Black is almost certainly lost with best play, the position now being:

click for larger view

24... d5?

"!"--(<crwynn>)

The above approval notwithstanding, this move which allowed White to play the crushing 25. Nf5! was a mistake. Black would have had better chances with 24...Qg6 or 24...Bg5 or even 24...Bf8, though Black was probably lost anyway.

Lasker said his intention was to play Qe6, but that doesn't work after 25. Nf5.

After 24...d5, Tarrasch had what may have been his last chance to win the game, the position now being:

click for larger view

25. exd5?

"?"--(Schroeder)(<crwynn>)(<Pawn and Two>

"Now 25. Nf5 is necessary." (Schroeder)

With the exception of Schroeder, none of the older commentators had any criticism of Tarrasch's 25. exd5. According to Winter: "There is nothing satisfactory now." Yet 25. Nf5 is a clear win.

As <keypusher> reports, Tarrasch expressly said that 25. Nf5 would not have led to a win, giving the following variation: 25. Nf5 Qg5 26. NxB QxN 27. exd5 Qh4+ 28. g3 after which Black draws by perpetual check beginning with 28...Qxh2+

But, as <keypusher> has correctly pointed out, Tarrasch overlooked 28. Qg3 after which there is no perpetual check and White wins. Black also loses, as <keypusher> has also pointed out if he plays instead 27...cxd5 28. RxR RxR.

<Pawn and Two> (who provides analysis showing the win beyond 28. Qg3) and <crwynn> also have pointed out that 25. exd5 was a mistake.

This is one of many examples that show that the contributors on this site have--often with the help of computer analysis--provided far better commentaries on this game than did their illustrious predecessors. The comparison is of course unfair since Stockfish, Fritz, and Shredder were not available in days of yore. But anyone interested in discovering the truth of the positions in this game will do far better to look at the commentary for this game appearing on this site back in 2006-2009.

While 25. Nf5 would have given Tarrasch a won game, he was still much better and retained at least practical winning chances--this is, after all, a very difficult game, especially over the board!

One thing is for sure, even with 25. exd5, Tarrasch was not lost or even worse as some of the older commentators seemed to think, the position now being:

click for larger view

The position was double-edges, but Tarrasch was still up a Pawn with a scary-looking passed a-Pawn.

25... Be3+
26. Kf1 cxd5

click for larger view

27. Rd3?

"?"--(<crwynn>)

As Tarrasch correctly pointed out, 27. Nf5 was his last chance to try for a win, and gave the following line: 27...d4 [27...Bc5 as <keypusher> has pointed out, would not have been a success; e.g., 28. RxR RxR 29. Re1! and Black is dead] 28. NxB dxN 29. Re2 (or perhaps better still 29. Rd3) and White is clearly for choice, even if probably not winning.

<crwynn> aptly says that after the text, "White's advantage is gone," the position now being:

click for larger view

The claim by Tarrasch that White was now lost was simply wrong, but seems to have persuaded older commentators to cease much serious analysis of the rest of the game.

The best assessment of the position at this point is that of <Pawn and Two> that the game is theoretically equal but difficult for White to play.

As a side note, the concept that the game was now hopeless for White led Schroeder to recommend sacrificing the exchange with 27. RxB?!. White can probably still survive (e.g., 27...RxR 28. Qxd5 Re1+ 29. Kf2) but there was no need for such drastic remedies at this point, though perhaps in practice Tarrasch would have found the position less challenging and might have drawn had he adopted this off-beat course.

27... Qe6
28. Re2

As <Pawn and Two> notes, Black is now slightly better. Tarrasch likely had better tries here (28. c3 or perhaps 28. a4 or 28. a3). But even after the text, White was hardly lost:

click for larger view

28... f5?!

As <Pawn and Two> has said, Lasker's move here, though praised to the hilt by older commentators such as Winter ("Lasker conducts the counter-attack with superb vigour") and Lasker ("the advance of the f-pawn is decisive"), was in fact inferior to 28...d4! and gave Tarrasch an easier route to a draw, the position (after 28...f5) now being:

click for larger view

29. Rd1?

As <Pawn and Two> has noted, Tarrasch could have drawn here with 29. Nh1! Ba7 (White is also fine after 29...f4 30. Rd1--KEG) 30. Rdd2 [30. RxR also draws--KEG] Qh6 31. RxR RxR 32. Rxd5 RxR 33. QxR Qc1+ 34. Ke2 Qe3+

After the text, Tarrasch was now in jeopardy:

click for larger view

29...f4?

For almost the entirety of the balance of the game, the older commentators--perhaps cowed by the analysis provided by Tarrasch and Lasker--simply missed the boat. By contrast, the analysts on this site have discovered many fascinating possibilities. For example, here--as pointed out by <Pawn and Two>--Lasker should have played the much stronger 29...d4! The text did not blow the win since White was not yet lost, but it did (or should have) eased Tarrasch's efforts to save the game.

30. Nh1!

The only saving move. Better late than never!

click for larger view

30... d4

<keypusher> reports that his version of Fritz gave 30...Re4 as best, but as he points out Lasker's move was better, e.g., 30...Re4 31. Nf2 Re5 32. Kg1 and suddenly White is back on top (though still hardly winning). For whatever it's worth, Fritz15 gives 30...Rg8 as best with 30..d4 ranked nearly as good (but neither giving Black much of an edge). On this point, therefore, Lasker does indeed "has the last laugh" as <keypusher> says.

31. Nf2

Now, according to Reti, Lasker was able to "force the decision." In fact, the game was now a theoretical draw, at least until Tarrasch's next move.

31... Qa6

click for larger view

"!"--(Reti)

"Threatening BxN." (Winter)

"Now, by means of two fine and deeply considered moves, (31st and 33rd) Lasker forces the decision." (Reti)

<crwynn> here faulted <keypusher> for suggesting that the game was about equal and claimed that at this point Tarrasch was fighting for his life. As a practical matter, Tarrasch may indeed have been overwhelmed by Lasker's attacking efforts, but as <Pawn and Two> has shown, and as I will discuss in my next post on this game, Tarrasch could still have achieved near equality here. It was only after his next move that Tarrasch was doomed, though--as we will see--even then Lasker subsequently gave him chances to save the game. So far as I have found, none of these possibilities were discovered before the analysis presented on this site.

32. Nd3?

"?"--(<crwynn>)(<Pawn and Two>)

"The Knight abandons defense of the K-side and this loses quickly." (<crwynn>)

"A serious error." (<Pawn and Two>)

The text was indeed a crucial error that should have allowed Lasker to win almost forthwith. But what else could Tarrasch have done?

The only alternative considered by <crwynn> is 32. Rde1, but that--while not losing--was not satisfactory. It was however, not as awful as <crwynn> suggests. After 32...Rg5 33. Nd1 is better than 33. Ng4 as given by <crwynn>. Even then White is not lost; e.g., 33. Ng4 Qg6 34. Nf2 (forced) Qh6 [34...Rg8 is better but even then White can probably hold with 35. Nd1 or perhaps 35. Ne4] 35. Nh3 Rh5 and now instead of the tricky 36. Rd2 as given by <crwynn> White can play 36. a4 with reasonable chances of survival.

But the above is all beside the point, since--as <Pawn and Two> has noted--White is OK after 32. Kg1!, e.g., Qxa2 33. Kh1 Qa8 34. Nd3 Ra5 35. b4 QxQ 36. gxQ Raa8 37. Ree1 Ra2 38. Nxf4 Re5 39. Ng2 Rxc2 40. NxB with a likely drawn double Rook ending no matter how Black recaptures.

By contrast, after the text, I agree with <Pawn and Two> that White appears to be lost.

32... Rg5

"Intending ...Qh6. The h-pawn is lost since it must guard the g3 square and therefore must not advance." (Lasker, as quoted here by <TheFocus>.

click for larger view

33. Ra1

"An attempt by the King to escape to the queenside. The terrible Bishop commanded him to stop." (Tarrasch translation again by <keypusher>)

The text, however, was quite hopeless.

"?"--(<keypusher>)(<Pawn and Two>)

"It is impossible to suggest any valid defense now. White's position is hopeless." (Hoffer)

White is indeed probably lost here, but the text should have hastened his defeat.

<keypusher> suggests that Tarrasch only here began to fall into serious danger of losing, but in fact that happened on Tarrasch's last move.

<keypusher> also claims that Tarrasch's position would be "far from hopeless" with 33. h3. But that move would also be a mistake: e.g., 33...Rg3! 34. Qh5 Re7 35. Qf5 Re6 36. Nc5 Rf6 37. NxQ RxQ 38. Ree1 f3 39. gxf3 and now what follows is brutal: Rxf3+ 40. Ke2 Rf2+ 41. Kd3 Rd2+ 42. Kxd4 Bc3+ 43. Ke4 BxR 44. RxB Rg6 45. Nc5 Rxc2 and White, though having a pawn for the exchange, can safely give up.

33. a4 was better than the text, but as Winter pointed out, 33...Qh6 is now murder since the White King has no escape.

After the text, <Pawn and Two> is surely correct that Lasker (though still down a pawn) had a clearly winning position:

click for larger view

33... Qh6

"!"--(Reti)

click for larger view

"The White h-pawn can no be defended." (Reti)

The text is indeed crushing, but--following Lasker's dictum---found something even better: 33...Reg8 34. Ke1 (there is nothing better) Qg6.

34. Ke1

Now, as pointed out long ago by Tartakower, if 34. h3 [??] Rg3 is a killer. If 35. Qxf4 Rf8 wins the Queen. If 35. Qb7 then 35...Qxh3 wins since White cannot play 36. gxQ because of 36...Rg1 mate. And if 35. Qd5 f3 wins for Black (but not, as noted by Tartakower 35...Qxh3 because of 36. Qxd4+! Kg8 [not 36...BxQ?? 37. RxR + and White wins]).

34... Qxh2
35. Kd1

click for larger view

35... Qg1+

As noted on this site, 35...Qh6 also wins. Both moves should be sufficient to finish off White.

36. Ne1

The best try, though inadequate.

As Tarrasch pointed out, if instead 36. Re1 Black wins with 36...Qxg2 after 37. QxQ [the best move available for White] RxQ since if now 38. Nxf4 Rd2+ 39. Kc1 Re2+ wins (40. Kd1 RxR+ 41. KxR BxN+ wins a piece)

36... Rge5
37. Qc6

click for larger view

37... R5e6

As <Pawn and Two> has discussed, there is still a win for Black after this move. But the real killer is the subtle 37...Kg8!

<Pawn and Two> claims that 37...d3 [he says "e3" but that I assume is a typo] also wins. But that gives White a draw after 38. cxd3 and now if 38...R8e6 (given as correct by <Pawn and Two] White just plays 39. Qxc7 [and not the suggested 39. Qf3 which gets smoked by 39...Bc5!].

38. Qxc7

click for larger view

38... R8e7?

As has been pointed out by the fine analysts on this site, the text blows the win. The winning line for Black here was 38...f3! 39. gxf3 Qg6! 40. Qc4 [40. Qc7 is no better] Bf2 41. RxR RxR 42. Kc1 (best but unavailing) RxN+ 43. Kb2 RxR 44. KxR h5 and Black wins; his extra piece and passed h-pawn are too much for White nominal two extra pawns.

Needless to say, this was all hard to find over the board. But some of the contemporary analysts with plenty of time at their disposal might have found the win, or at least noted that Tarrasch now could have saved the game, the position now being:

click for larger view

39. Qd8+?

This mistake finished off Tarrasch, and Lasker now wrapped up the game in three moves. But, as has been pointed out on this site (but never before so far as I am aware), Tarrasch might have saved the game with 39. Qc8+ instead.

To be fair, in Tarrasch's own commentary (helpfully translated here by <keypusher>), he said that "it was better...to keep the Rook on e6 under threat, for example by Qc4 or Qc8+" but concedes that by this point that "under the relentless pressure of the black pieces [unwilling to credit Lasker by name] I had long since thrown in the towel."

39. Qc4 would have lost after 39...Re5 40. Qc8+ Kg7 41. Qg4+ Kf7 [41...Kf6 also wins] 42. Qh4 [or 42. Qf3 d3] h4 43. Qh3 Kg6

But, for reasons Tarrasch apparently overlooked, 39. Qc8+ would likely have allowed him to draw.

Why is Qc8+ correct but the seemingly similar Qd8+ [Tarrasch's actual move] wrong? Because Black has a Black-square Bishop that comes to g5 and--as will be seen--allows a crushing discovered attack on the White Queen (the combination Lasker unleashed in the game)

Even so, does 39. Qc8+ actually save the game? The line given here by <Pawn and Two> based on Fritz seems convincing, although for quite a while I didn't believe it and tried (without success I might add) to win the game as Black.

I first tried 39...Kg7 but after 40. a4 with b4 and a5 to come I had to give up my quest.

A better try is 39...Re8 as given by <Pawn and Two> But 40 Qc7 or 40. Qd7 [see line given by <Pawn and Two> appear to hold for White.

Once again, I agree with <keypusher> that the analysis of the finale of this game given on this site "should be included in any future books on this game.

But back to the actual game after Tarrasch's actual 39. Qd8+?:

click for larger view

Lasker now made short work of White's now hopeless position:

39... Kg7

The White Queen is now helpless to save the game, and it's vulnerable position was now ruthlessly exploited by Lasker:

40. a4

Nothing else is better.

40... f3!

click for larger view

41. gxf3 Bg5!

click for larger view

Now, if 42. RxR RxR the g5 Bishop springs to life, and--to quote Winter--after 43. Qh5 Bh4 "wins everything (and 43...Bf4 is even more devastating).

So:

0-1

I am sad to report that I had to endure this form of torture more than once, but that is the price I had to pay for taking lessons (always ending in a game) against my grandmaster teacher. Luckily for me, these games are only reported in a folder I will never share with anyone (except for one of the very few wins I scored that is the closest thing to a brilliancy I will probably ever manage to execute).

(But remember the, er, <parity rule>(*) - to match that brilliancy, one of your more memorable losses should also be submitted - in all fairness y'know).

And if I may remind you of the <parity-rule> and this tidbit:

Christopher Chase (kibitz #6)