help with traxler opening variation #f7

Hey all,

Been seeing lots of online opponents (low rating like me ) try to quick mate with the attack on f7 using bishop and knight. Granted, it's pretty simple to just defend and develop, but this gets me so frustrated every time and I wish I could effectively counterattack.

One of the methods I really liked was the Traxler Counter Attack because of the sacrifices and tempo. The example I saw was:

Stockfish suggests also the knight sacrifice 8...Qe7 9. Kxe4 Rf8.

IMBacon: It's not sure at all that 7. Kg1 is better than 7. Ke3 and that "white is winning". In a game between human players both moves will create something messy with an unpredictable issue.

I may have been a bit overly ambitious with my assessment, and will admit that at the OP's level anything can and will happen.  My assessment that "white is winning" was from my point of view.  

I don't know whether 7. Kg1 draws or loses for White but it doesn't win. (Black can play Qh5 and if then simply take it g3 and there's obviously at least a draw there but the last time I went down a long Traxler line that Stockfish thought White was winning in these forums I showed a draw when Black actually had a win if I'd been a bit more creative and looked a bit further so I'm not going to go there again just now.) Anyway 5. Nxf7 has been analyzed to a draw; that's why White should take with the Bishop.

Thanks guys. I didn't really expect Kg1 and white seems to be able to recover. Not certain if these are optimal moves:

I can't seem to get the white bishop off that diagonal. I'm enticed by how white's king looks so vulnerable but I don't see an approach after Kg1 that stands out (messy indeed).

Hm...what about 12.d5 13.Bxd5 Bh3+?

After 8. g3 Stockfish wants to take with the knight, and it assesses the position as equal (triple zeros) after:

1. e4 e5 2. Nf3 Nc6 3. Bc4 Nf6 4. Ng5 Bc5 5. Nxf7 Bxf2+ 6. Kxf2 Nxe4+ 7. Ke3 Qh4 8. g3 Nxg3 9. hxg3 Qd4+ 10. Kf3 O‑O 11. Rh4 e4+ 12. Kg2 Qxc4

All of white's moves from 8. g3 onwards are pretty much forced.

zchickenwang: As tmkroll said, experienced players will deviate earlier and simply play 5. Bxf7+ instead of messy 5. Nxf7 then retreat the LSB to, let's say, b3.

5. Nxf7 is messy but should only give white a draw with best play (though it's very messy).

5. Bxf7+ or 5. d4 are the two ways for white to play for an advantage. After 5. d4 black's only good defense starts with 5... d5.

It seems that black can also play 5...Nxd4.

I'm really surprised at all of these conclusive opinions being given when there has been plenty of analysis given in the forums already showing that black is fine after 5.Nxf7.

I'm not sure what Bacon is saying when white has an overwhelming advantage after 5.Nxf7 and then also brings up 7.Kg1 as being completely winning.  If anything, black gets a draw in hand attack and white has zero winning chances.  It's much too drawish with perpetuals everywhere.

1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5 Bc5 5.Nxf7 Bxf2+ 6.Kxf2 (if white wants to play on, it's better to keep winning chances with 6.Kf1 Qe7 7.Nxh8 d5 8.exd5 Nd4 where we reach the main tabiya of the Traxler 5.Nxf7 variation) 6...Nxe4+ 7.Kg1 Qh4 8.g3 Nxg3 9.hxg3 (A poor choice) 9...Qxg3+ (Better than 9...Qd4+ because white failed to remove the rook with 9.Nxh8) 10.Kf1 Rf8 and black has an extremely dangerous attack.  Simply 11.Qh5 d5 and the above mentioned 11.Ke2 d5 should be just about winning.

To avoid this white is just about forced to play 9.Nxh8.

9.Nxh8 Nd4 and white should just settle for a perpetual with 10.hxg3 Qxg3+ 11.Kf1 Qf4+ otherwise he will lose.

Theory has black playing 10...d5 here, but 10...0-0 can transpose back into the main line.

10...0-0 11.Rh4 e4+ 12.Kg2 and now 12...d5 reaches the main tabiya.  12...Qxc4 might be an inaccuracy because white may try to play 13.Ng5.

13.Ng5 Nd4 (what else?) 14.Rxh7 Rf5 15.Nc3 where it becomes really awkward to try to hold the balance as black.  White probably has a sizable advantage here.

13. Ng5 Qd4 is Stockfish's answer to "what else?"

LoL Almost nobody wanna search and read old threads. Tomorrow someone can easily post a new Traxler topic without even looking at this one.

13.Ng5 Qd4 14.Qe2 and I guess black has to do something regarding h7 so now 14...h6 15.Rxe4 Qd5 and now 16.Nf3 looks like black has insufficient compensation for the material deficit. 

Considering that I have Stockfish as well, I turned it on to compare results.

It reads this as +0.82 at depth 25 so I'm wondering if your silicon is falling into a perpetual evaluation trap.

EDIT: Apparently the engine I used was Komodo 8.  

I ran it with the correct Stockfish 5 instead and got +0.87 at depth 25.

I just started running it (Stockfish 8) on the position after 13. Ng5, and noticed that the evaluation was very favorable for white until depth 31, when 13... Qd4 became the preferred move and it jumped to very close to zero (+0.07 at depth 31). So you have to run a deep search I think. I'll keep going with the search and report back later.

Interim result: At depth 35 it prefers 13... Qd4 with an eval of 0.11. Second choice is 13... h6 with eval 0.30. Prior to depth 31 it preferred h6.

Huh, it's kind of a curious result.  I'm really not sure what it is seeing that makes the position seem pretty even.  I'm running a deeper study of the the position to see what line it prefers.  Granted I have version 5 vs version 8, but that still is a rather large difference.  What line is being given?  I assume there probably is an improvement to 15...Qd5, but I'm failing to see the compensation.

My silicon is recommending that white play 15.Nxe4 over 15.Rxe4 to save a tempo I guess.  I still like the plan of 15.Rxe4, Nf3 plan though.

The line I have is at depth 31 at +0.06 is:

13...Qd4 14.Qe2 h6 15.Nxe4 Qe5 16.Nec3 Qxe2+ 17.Nxe2 and it thinks that the position is equal after 17...b5.  I suppose the threat is to mobilize the QB to the a8-h1 diagonal and with the lack of development of the white camp, black has a sort of dynamic equality.  It's really hard to understand, but sort of beautiful.  That line would need some rigorous testing to figure out what's happening.  It feels like that whole setup is blown up by the Qe5, b5 plan so why not try to stop it with 15.Rxe4 instead?  What does your rig read after that?

After 14. Qe2 h6 it doesn't care whether white plays Rxe4 or Nxe4, it's giving 0.00 either way. But in response to 15. Rxe4 only 15... Qc5 preserves that eval, any other move and white is winning. Here's what I've got currently (assuming the deep evals after only a short time are due to stored positions from previous searches to reach this position).

1. e4 e5 2. Nf3 Nc6 3. Bc4 Nf6 4. Ng5 Bc5 5. Nxf7 Bxf2+ 6. Kxf2 Nxe4+ 7. Ke3 Qh4 8. g3 Nxg3 9. hxg3 Qd4+ 10. Kf3 O‑O 11. Rh4 e4+ 12. Kg2 Qxc4 13. Ng5 Qd4 14. Qe2 h6 15. Rxe4

15... Qc5 16. d4 Nxd4 17. Be3 b6 18. Bxd4 Qxg5 19. Bxg7 Rf5 20. Bc3 Bb7 21. Nd2 Raf8 22. Re1 Qg6 23. b4 Bxe4+ 24. Qxe4 Rf2+ 25. Kh3 Qh5+ 26. Qh4 Qf5+ 27. Qg4+ Qxg4+ 28. Kxg4 d5 29. b5 R8f5 30. Bb4 a5 31. bxa6 Rg5+ 32. Kh3 Rh5+ 33. Kg4 Rg5+
= (0.00) Depth=40/68 0:04:15 1498 MN

15... Qf6 16. Qc4+ d5 17. Qxd5+ Kh8 18. Rf4 Bf5 19. Rxf5 Qxf5 20. Qxf5 Rxf5 21. Ne4 Raf8 22. b3 Nd4 23. Na3 Nf3 24. Bb2 Nxd2 25. g4 Rf4 26. Nxd2 Rf2+ 27. Kg3 Rxd2 28. Re1 Rdf2 29. Re3 b6 30. Be5 c5 31. Bb8 R8f7 32. Rc3 Re2 33. Nc4 Rff2 34. Ne3 Rf7 35. a4 Kg8 36. Nc4 Rb7 37. Kf3 Re1 38. Kf2 Rc1 39. Bd6 Rf7+ 40. Kg3 Rg1+ 41. Kh3 Rc1 42. Bb8 Kh7 43. Kg2
+ (0.56) Depth=39/68 0:04:15 1498 MN