Latvian Lacerated

[Stephen Ham]

Hi Eelco,

Thanks for your analyses.

However, as I human, I'm shocked that your two silicon analyzers failed to spot the obvious refutation of 1e4 e5 2 Nf3 f5 3 Nxe5 fxe4?. Surely Black's reply loses to a move that all humans should find instantly - 4 Qh5+. Now, 4...g6 5 Nxg6 hxg6 6 Qxh8 Kf7 7 Nc3 basically ends the game...although Black can struggle until eventually mated.

A human can see this immediately. So, this raises the bigger question: why didn't Crystal 7 see this after 8 minutes of computation and a ply depth of 37?

All the very best,
-Steve-

[Peter Berger]

Dear Peter Berger,

Long time, no talk, buddy!

You wrote,
"1. E4 e5 2. Nf3 nc6 3. C3
1. d4 Nf6 2. C4 e6 3. F3
I was absolutely sure that both of these should be a draw, as white didn’t do that bad. I couldn’t confirm it though – I didn’t find a convincing way for white to equalize."

The top line represents the Ponziani Opening. At least one very strong fellow Correspondence Chess GM loves it and thinks White is better. I also played it as White back in my OTB chess days, decades ago. So if you believe that White can't even equalize, then please show us Black's best line(s). My opinion is that Black struggles but can equalize only with 3...d5 and 3...Nf6.

The bottom line is interesting but quite uncommon and indeed probably dubious after 3...d5. Still please share your line(s) here too, Peter.

Thanks in advance.

All the very best,
-Steve-

Dear Steve,

let's start with the latter. I saw the slightly different versions of your post and you getting way less interested over a short period of time. We can probably simply agree that it doesn't work too well after 1. d4 Nf6 2. c4 e6 3. f3 d5, which still surprises me personally - and leave the challenge to convert to anyone meeting it in a real game .

The Ponziani is certainly much more of a challenge. I face the problem here, that I don't own any recent human opening books, so I am not really sure, how strong humans want to make it work. Is Graham's post good guidance - and white wants to continue with 5. d3 after 1. e4 e5 2.Nf3 Nc6 3. c3 d5 4. Qa4 f6, which would be my starting point to reply to?

I appreciate talking to you again after all this time, too.

All the best,
Peter

[Stephen Ham]

Lieber Peter,

I own many old books on the Ponziani, and played it back in the 70s. I know it very well, even though I've switched to 1 d4 exclusively since the 80s. That said, my bias is that White is objectively equal against Black's best continuations, with great practical chances, especially at club-level play.

After 1 e4 e5 2 Nf3 Nc6 3 c3 d5 (3...Nf6 is Black's only other path to equality) 4 Qa4 f6 Graham's 5 d3 is less frequently played and seems unnecessarily passive. White's mainline continuation is the more natural 5 Bb5. Black has several options here but the mainline continues 5...Ne7 6 exd5 Qxd5 when it branches with both 7 d4 and 7 0-0.

Now, buddy, the burden is on you to defend your claim that White can't equalize.

Alles Gute,
-Steve-

[Eelco de Groot]

Hi Eelco,

Thanks for your analyses.

However, as I human, I'm shocked that your two silicon analyzers failed to spot the obvious refutation of 1e4 e5 2 Nf3 f5 3 Nxe5 fxe4?. Surely Black's reply loses to a move that all humans should find instantly - 4 Qh5+. Now, 4...g6 5 Nxg6 hxg6 6 Qxh8 Kf7 7 Nc3 basically ends the game...although Black can struggle until eventually mated.

A human can see this immediately. So, this raises the bigger question: why didn't Crystal 7 see this after 8 minutes of computation and a ply depth of 37?

All the very best,
-Steve-

Hello Stephen,

it is possible that standard NNUE from Crystal is not very good here, it is the same as the Stockfish NNUE. But I think it does okay actually, and Rebel EAS NNUE from Ed and Chris Whittington may be even better because it defeated Stockfish 10 a while ago. But I suspect you now have no idea what I am talking about because you did not see the diagram... And also not the pgn above that should be visible. That is the same as in OTB play when people walk away from the board because they think they have already seen everything in their mind. Some super GMs can do that and it actually helps their concentration perhaps, like I read from Vasyl Ivanchuk. They are the exceptions! Well I am not good at blind chess at all. But please revert to http:// for a moment if you want to see the diagram Eight Knights chess was introduced here but Graham thinks we are better playing Tiddlywinks if we don't want to play chess anymore. Maybe he's right. We call it 'Vlooienspel' in Dutch. It may never be solved by computers, that is a plus.

Of course you are right if this were normal chess Latvian Gambit!

Good to see you here and all the best from me too!

[Peter Berger]

Lieber Peter,

I own many old books on the Ponziani, and played it back in the 70s. I know it very well, even though I've switched to 1 d4 exclusively since the 80s. That said, my bias is that White is objectively equal against Black's best continuations, with great practical chances, especially at club-level play.

After 1 e4 e5 2 Nf3 Nc6 3 c3 d5 (3...Nf6 is Black's only other path to equality) 4 Qa4 f6 Graham's 5 d3 is less frequently played and seems unnecessarily passive. White's mainline continuation is the more natural 5 Bb5. Black has several options here but the mainline continues 5...Ne7 6 exd5 Qxd5 when it branches with both 7 d4 and 7 0-0.

Now, buddy, the burden is on you to defend your claim that White can't equalize.

Alles Gute,
-Steve-

Dear Steve,
my computer thinks that white’s position is plain horrible after 1. E4 e5 2. Nf3 Nc6 3. C3 d5 4. Qa4 f6 5. Bb5 Ne7 6. Exd5 Qxd5 7.O-O
A sample line:
Analysis by Rebel-16.2:

7...e4 8.Se1 Ld7 9.f3 a6 10.Le2 f5 11.d4 exd3 12.Sxd3 0-0-0 13.Db3 Sa5 14.Dc2 Sg6 15.b4 Sc4 16.a4 Df7 17.Sa3 Le6 18.Sxc4 Lxc4 19.Te1 Ld6 20.Le3 The8 21.Lf1 Se5 22.Sxe5 Txe5 23.Lg5 Lb3 24.Db2 Tde8 25.Txe5 Lxe5 26.Lh4 Lc4 27.Lf2 Df6 28.Dc2 Lxc3 29.Tc1 Lxf1 30.Kxf1 Le5 31.b5 Lf4 32.Td1 axb5 33.axb5 Td8 34.Txd8+ Kxd8 35.Db3 Da1+ 36.Ke2
Schwarz steht deutlich besser: -/+ (-1.53) Tiefe: 50/83 09:33:04 65963MN
I basically agree (without looking at things too far down the line, that never make much sense).
7. d4 clearly looks somewhat better. But again after 7. …e4 I fail to see a really satisfactory development for the white side.
As you don’t play this yourself – maybe you can give some line that you think works better, and I can give a point where I disagree in response?
All the best,
Peter

[Stephen Ham]

Lieber Peter,

I agree that 7 0-0?! is the inferior choice after 7...e4. Although the -/+ evaluation is a bit extreme. Opening theory continues many moves further, leading to an edge for Black after 8 Nd4, which is superior to 8 Ne1?! and leads to a clear Black advantage.

Peter, White's best continuation is instead 7 d4. Now you like 7...e4. OK, next look at the uncommon 8 Ng1!? a6 9 Bc4 Qf5 10 Qc2 b5 11 Be2 Qg6 12 Nd2 continuation. I think it's fine for White, but again have not played it in decades. Perhaps you can find Black improvements.

Peter, it's interesting that you're using Rebel for analysis. Isn't the latest Stockfish development iteration with the large NNUE superior? Normally I'd use mine to join you, but my engine is busy on an ICCF game.

Alles Gute,
-Steve-

[Ajedrecista]

Hello Peter:

Dear Steve,
my computer thinks that white’s position is plain horrible after 1. E4 e5 2. Nf3 Nc6 3. C3 d5 4. Qa4 f6 5. Bb5 Ne7 6. Exd5 Qxd5 7.O-O
A sample line:
Analysis by Rebel-16.2:

7...e4 8.Se1 Ld7 9.f3 a6 10.Le2 f5 11.d4 exd3 12.Sxd3 0-0-0 13.Db3 Sa5 14.Dc2 Sg6 15.b4 Sc4 16.a4 Df7 17.Sa3 Le6 18.Sxc4 Lxc4 19.Te1 Ld6 20.Le3 The8 21.Lf1 Se5 22.Sxe5 Txe5 23.Lg5 Lb3 24.Db2 Tde8 25.Txe5 Lxe5 26.Lh4 Lc4 27.Lf2 Df6 28.Dc2 Lxc3 29.Tc1 Lxf1 30.Kxf1 Le5 31.b5 Lf4 32.Td1 axb5 33.axb5 Td8 34.Txd8+ Kxd8 35.Db3 Da1+ 36.Ke2
Schwarz steht deutlich besser: -/+ (-1.53) Tiefe: 50/83 09:33:04 65963MN
I basically agree (without looking at things too far down the line, that never make much sense).
7. d4 clearly looks somewhat better. But again after 7. …e4 I fail to see a really satisfactory development for the white side.
As you don’t play this yourself – maybe you can give some line that you think works better, and I can give a point where I disagree in response?
All the best,
Peter

The Fifth Edition (2006) of the Volume C of the Encyclopedia of Chess Openings gives a line with 7.- d4 at note 14 of ECO C44, but none with 7.- O-O. The given game is Eales — Beliavsky; Groningen 1969/70:

1.- e4, e5; 2.- Nf3, Nc6; 3.- c3, d5; 4.- Qa4, f6; 5.- Bb5, Nge7 (sic, even with Nc6 pinned; Arena 2.0.1 GUI agrees with ECO notation); 6.- exd5, Qxd5; 7.- d4, Bg4; 8.- Bc4, Qa5; 9.- Qxa5, Nxa5; 10.- Bb5+, c6; 11.- Be2, Bxf3; 12.- Bxf3, exd4; 13.- Na3! (white stands slightly better).

Eales won this game, although not all the moves in the opening were optimal according to SF 16.

The first game that I found with 7.- ..., e4 is a white win in 1986:

https://lichess.org/ImOdMOvx
https://old.chesstempo.com/gamedb/game/1917064

Where black blundered twice.

Regards from Spain.

Ajedrecista.

[Peter Berger]

Hi folks,
regarding notation. ECO (and Arena AND all Chessbase programs) are wrong here. We can look at the rules of chess to clarify. This is covered in Appendix C of the FIDE chess laws in C.10 :
„If two identical pieces can move to the same square, the piece that is moved is indicated …“
But the other knight can’t move to this square, the poor thing being pinned. The rule is as clear as it gets (in fact you could even create a case how adding the additional letter would VIOLATE the rules and may even be punished written down on your sheet according to FIDE rules). I am nearly 100% sure that a ChessBase program started with this useless additional letter, and there should be some ancient talkchess discussions about this (I remember Bob Hyatt being vocal about the ChessBase implementation). Of course, this doesn’t matter that much.
Regarding ECO C. I own the 4th edition from 2000, that used to be well-known to be pretty useless – it has the same line from Eales-Beliavsky – and nothing else, so this was obviously just copied to the next edition.
And regarding my useage of Rebel 16.2. I am currently doing a longish correspondence chess experiment where I am only allowed to use this engine by design. That I used it here is just coincidence, as it is so easily availlable on my computer – and I really like it anyway. I don’t think it matters at all – Rebel is supposedly like 80 points weaker than Stockfish ELO-wise – this is like nothing at a longer time control.
No time for chess today, so back to the Ponziani later, thanks for your interesting replies.
All the best.
Peter

[Werewolf]

Hi Eelco,

Thanks for your analyses.

However, as I human, I'm shocked that your two silicon analyzers failed to spot the obvious refutation of 1e4 e5 2 Nf3 f5 3 Nxe5 fxe4?. Surely Black's reply loses to a move that all humans should find instantly - 4 Qh5+. Now, 4...g6 5 Nxg6 hxg6 6 Qxh8 Kf7 7 Nc3 basically ends the game...although Black can struggle until eventually mated.

A human can see this immediately. So, this raises the bigger question: why didn't Crystal 7 see this after 8 minutes of computation and a ply depth of 37?

All the very best,
-Steve-

Because he's not using regular chess. It's 960.

[Peter Berger]

..

Peter, White's best continuation is instead 7 d4. Now you like 7...e4. OK, next look at the uncommon 8 Ng1!? a6 9 Bc4 Qf5 10 Qc2 b5 11 Be2 Qg6 12 Nd2 continuation. I think it's fine for White, but again have not played it in decades. Perhaps you can find Black improvements.

..

Dear Steve,

I have to admit that my file on the Ponziani contained some sloppy analysis. My feeling is, that white is the one suffering in general, but for the time being I don't have anything ready, that would prove that White can't equalize. I could work on it of course - but then, you could do that, too. So, I'll just eat my words: it is conceivable that White can equalize with the Ponziani.

Thanks for pointing this out and all the best to you.

Peter