Smith Morra Bust?

[zullil]

18-year old USCF Expert Elijah Logozar claims to have refuted the Smith Morra Gambit. His mainline goes:
[pgn]1.e4 c5 2.d4 cxd4 3.c3 dxc3 4.Nxc3 Nc6 5.Nf3 e6 6.Bc4 Nge7!? 7.Bg5 h6!{N}[/pgn]
US Chess is offering an unspecified prize to the best counter-analysis submitted before Aug 15, 2019.
See https://new.uschess.org/theory/bust-smith-morra-gambit/ for the teen's comments on the line.

h6 is not a novelty

I can find in chesstempo that
White won against h6 in the following human-human game in 1994

Carrion-Lara, Luis vs Schermer, Anton

1. e4 c5 2. d4 cxd4 3. c3 dxc3 4. Nxc3 Nc6 5. Nf3 e6 6. Bc4 Nge7 7. Bg5 h6 8. Nb5 d5 9. exd5 exd5 10. Bf4 Ng6 11. Nc7+ Ke7 12. Nxd5+ Ke8 13. Nc7+ Ke7 14. Qe2+

[pgn][Event "Alkmaar"] [Site "Alkmaar"] [Round "1"] [Date "1994.??.??"] [White "Carrion-Lara, Luis"] [Black "Schermer, Anton"] [Result "1-0"] 1.e4 c5 2.d4 cxd4 3.c3 dxc3 4.Nxc3 Nc6 5.Nf3 e6 6.Bc4 Nge7 7.Bg5 h6 8.Nb5 d5 9.exd5 exd5 10.Bf4 Ng6 11.Nc7+ Ke7 12.Nxd5+ Ke8 13.Nc7+ Ke7 14.Qe2+ 1-0 [/pgn]

[Ovyron]

I find it difficult to believe that you'd beat anyone with 7...h6. Even at 12+2 time control.

Oh, I've beaten people with much worse lines, at 50days/10 moves with 50 day increment every 10 moves time control. How bad people play is surprising, and I still don't know why they play so badly, or why they have 300 elo points higher than me (Ozymandias has theorized about this on Rybka Forum). But sometimes they play really badly and it's not enough to defeat them, so it's better to play good lines against them, just in case.

I guess the question is, if I faced an opponent in the Smith Morra, would I be better playing 7...h6 or 7...f6 ? Even if h6 is better, its mainline has been published, so they could just copy and paste moves, and then I'm just playing the position after 13...g6. The mainline for 7...f6 remains private, so who knows if they can find the moves by themselves.

I guess the sad part about this is that people are playing so badly that it may not matter what move I play. And all my analysis is going to waste because someone with Leela and a decent GPU could beat them at 12+2 time control (vs. 50/day repeating time control), but, just because Stockfish at depth 32 finds the correct move in a line every time, doesn't mean contending people are going to find it...

[zullil]

I find it difficult to believe that you'd beat anyone with 7...h6. Even at 12+2 time control.

Oh, I've beaten people with much worse lines, at 50days/10 moves with 50 day increment every 10 moves time control. How bad people play is surprising, and I still don't know why they play so badly, or why they have 300 elo points higher than me (Ozymandias has theorized about this on Rybka Forum). But sometimes they play really badly and it's not enough to defeat them, so it's better to play good lines against them, just in case.

I guess the question is, if I faced an opponent in the Smith Morra, would I be better playing 7...h6 or 7...f6 ? Even if h6 is better, its mainline has been published, so they could just copy and paste moves, and then I'm just playing the position after 13...g6. The mainline for 7...f6 remains private, so who knows if they can find the moves by themselves.

I guess the sad part about this is that people are playing so badly that it may not matter what move I play. And all my analysis is going to waste because someone with Leela and a decent GPU could beat them at 12+2 time control (vs. 50/day repeating time control), but, just because Stockfish at depth 32 finds the correct move in a line every time, doesn't mean contending people are going to find it...

Sounds like you're wasting your efforts if the opposition is that bad!

In any case, here's what Cfish thinks of your 7...f6. Just a quick first look:

0.00 8. Be3 Ng6 9. O-O Bb4 10. Rc1 b6 11. Qb3 Bxc3 12. Qxc3 O-O 13. Bxe6+ dxe6 14. Qxc6 Bd7 15. Qd6 e5 16. Rfd1 Bg4 17. h3 Qxd6 18. Rxd6 Bxf3 19. gxf3 Rfc8 20. Rxc8+ Rxc8 21. Rd7 Rc2 22. Rxa7 Rxb2 23. a4 f5 24. exf5 Nh4 25. Kf1 Nxf3 26. Rb7 Ra2 27. Rxb6 Ra1+ 28. Ke2 e4 29. Bd2 Ra2 30. Rd6 Kf7 31. Ke3 Nxd2 32. Rxd2 Rxa4 33. Rd6 Ra3+ 34. Kxe4 Rxh3 35. Rb6 Rh6 36. Rxh6 gxh6 (depth 44, 0:09:45)

[Ovyron]

Sounds like you're wasting your efforts if the opposition is that bad!

That's the ICCF for you.

In any case, here's what Cfish thinks of your 7...f6. Just a quick first look:

0.00 8. Be3 Ng6 9. O-O Bb4 10. Rc1 b6 11. Qb3 Bxc3 12. Qxc3 O-O 13. Bxe6+ dxe6 14. Qxc6 Bd7 15. Qd6 e5 16. Rfd1 Bg4 17. h3 Qxd6 18. Rxd6 Bxf3 19. gxf3 Rfc8 20. Rxc8+ Rxc8 21. Rd7 Rc2 22. Rxa7 Rxb2 23. a4 f5 24. exf5 Nh4 25. Kf1 Nxf3 26. Rb7 Ra2 27. Rxb6 Ra1+ 28. Ke2 e4 29. Bd2 Ra2 30. Rd6 Kf7 31. Ke3 Nxd2 32. Rxd2 Rxa4 33. Rd6 Ra3+ 34. Kxe4 Rxh3 35. Rb6 Rh6 36. Rxh6 gxh6 (depth 44, 0:09:45)

I'm not going to discuss any line for obvious reasons

zullil, do you contend? I've been thinking that we could play a correspondence chess game, I'm willing to show my lines in a game from the opening position against you.

[Uri Blass]

I find it difficult to believe that you'd beat anyone with 7...h6. Even at 12+2 time control.

Oh, I've beaten people with much worse lines, at 50days/10 moves with 50 day increment every 10 moves time control. How bad people play is surprising, and I still don't know why they play so badly, or why they have 300 elo points higher than me (Ozymandias has theorized about this on Rybka Forum). But sometimes they play really badly and it's not enough to defeat them, so it's better to play good lines against them, just in case.

I guess the question is, if I faced an opponent in the Smith Morra, would I be better playing 7...h6 or 7...f6 ? Even if h6 is better, its mainline has been published, so they could just copy and paste moves, and then I'm just playing the position after 13...g6. The mainline for 7...f6 remains private, so who knows if they can find the moves by themselves.

I guess the sad part about this is that people are playing so badly that it may not matter what move I play. And all my analysis is going to waste because someone with Leela and a decent GPU could beat them at 12+2 time control (vs. 50/day repeating time control), but, just because Stockfish at depth 32 finds the correct move in a line every time, doesn't mean contending people are going to find it...

I am not surprised that you have beaten people with much worse lines.

I believe that chess engines with no help for a long time can get at least 2500 ICCF rating.

The way to get high ICCF rating relatively fast is simple.

First step is to win the correspondence championship of your country
Second step is playing for your country in the olympiad.

This is at least how I did it in the past and my first ICCF rating was 2585 if I remember correctly(when I do not include provisional rating of 2200 that I had in the first olympiad).

see https://www.iccf.com/event?id=35844

[carldaman]

It's hard to envision a typical "engine slave" (someone who just relays the engine's move without significant analytical interaction - ovyron's description is spot-on) having a 2000+ rating on ICCF. However, I'll concede that it may be possible if the engine is left on to analyze for days at a time on very high-end hardware.

[Ovyron]

It's hard to envision a typical "engine slave" (someone who just relays the engine's move without significant analytical interaction - ovyron's description is spot-on) having a 2000+ rating on ICCF.

They seem to have over 2400 rating.

Also, there's even higher rated people that seem to not even use engines, and the theory is that they achieved this rating long time ago (like, +20 years ago) and the games they play haven't hit their rating as fast as they could (or their opponents see high rating and are scared so they play for a draw.) That's why chess engines with no help can reach 2500.

[jp]

"Bust" is a bit of an extreme word in this case.

Nevertheless, I think he correctly claims that in his lines, black is the one pressing and white is (successfully, with good play) trying to hold, which is a pretty good outcome for black in chess.

I agreed. Most of gambits are unsound in today technology with tons of analysis.

Because chess is an extremely drawish game with big margin of error.

I don't know "unsound" is the right word either if you could show that with best play it's still a draw. To the computer, all draws are equal, and we wouldn't say the computer plays deliberately unsoundly.

[Ovyron]

I don't know "unsound" is the right word either if you could show that with best play it's still a draw.

Yeah, what we need is a way to measure how difficult is it for white/black to play this drawing line. There's 2.00 scored positions where drawing is very easy, and 0.00 positions where drawing is very hard, we need some other mechanism to score chess positions (or maybe we don't... if engines get good at knowing what positions are easy to draw and aim for them it may be the end of computer chess...)

[zullil]

Sounds like you're wasting your efforts if the opposition is that bad!

That's the ICCF for you.

In any case, here's what Cfish thinks of your 7...f6. Just a quick first look:

0.00 8. Be3 Ng6 9. O-O Bb4 10. Rc1 b6 11. Qb3 Bxc3 12. Qxc3 O-O 13. Bxe6+ dxe6 14. Qxc6 Bd7 15. Qd6 e5 16. Rfd1 Bg4 17. h3 Qxd6 18. Rxd6 Bxf3 19. gxf3 Rfc8 20. Rxc8+ Rxc8 21. Rd7 Rc2 22. Rxa7 Rxb2 23. a4 f5 24. exf5 Nh4 25. Kf1 Nxf3 26. Rb7 Ra2 27. Rxb6 Ra1+ 28. Ke2 e4 29. Bd2 Ra2 30. Rd6 Kf7 31. Ke3 Nxd2 32. Rxd2 Rxa4 33. Rd6 Ra3+ 34. Kxe4 Rxh3 35. Rb6 Rh6 36. Rxh6 gxh6 (depth 44, 0:09:45)

I'm not going to discuss any line for obvious reasons

zullil, do you contend? I've been thinking that we could play a correspondence chess game, I'm willing to show my lines in a game from the opening position against you.

No, I've never played correspondence chess. And with players able to access all sorts of software/hardware, I don't think I understand the point of such a competition.