italien opening

[Uri Blass]

Ovyron claimed that the italien opening is good for black.

GM larry kaufman claims that it is good for white

his sample pages about his new repertoire has the following information:

"So how much of an edge can you get as White with my repertoire? Here
are the Elo advantages for White after my preferred response to the main
lines (per Hiarcs db):
- Caro-Kann Two Knights +40;
- French Tarrasch +45;
- Sicilian 2...♘c6 3.♗b5 +37;
- Sicilian 2...d6 3.♗b5+ +40;
- Petroff main line +37;
- Italian +30;
- Spanish Berlin 5.♖e1 with ♗f1 +42;
- Spanish 6.d3 +26"

What is your opinion?
I guess that most agree that it is a draw with perfect play but I added also 2 options that white is winning or black is winning.

[Ovyron]

My claims about it has been about unassisted engine chess (blitz from 5 0 to 12 2 time control) and assisted chess at correspondence time controls (+1 year per game), based mainly in positions where Stockfish prefers white but it's black that has the advantage, and the critical positions contain material imbalances that engines don't seem to play well from the white side. An opening book with these lines will score better from the black side of the Italian against itself and will outperform old books unprepared (I can give dates, it seems the relevant lines were found/uncorked in Febrary 2019 and were polished up to May 2019 - since then the Italian has become rare and black scores better when it is tried - again, because of unprepared opponents with old books.)

However, we're discussing apples and oranges as Larry Kaufman is talking about the Italian in OTB human games where engines are forbidden, in such cases he's probably right, and it's possible the engine-engine lines useful in corr chess have no use in practice of human games (black needs engine-like tactical awareness to succeed.) I still expect that a black win on the Italian will be played in an important human game, and to win one at corr chess, but black's advantage in it is so small I won't hold my breath.

[Ajedrecista]

Hello:

Ovyron claimed that the italien opening is good for black.

GM larry kaufman claims that it is good for white

his sample pages about his new repertoire has the following information:

"So how much of an edge can you get as White with my repertoire? Here
are the Elo advantages for White after my preferred response to the main
lines (per Hiarcs db):
- Caro-Kann Two Knights +40;
- French Tarrasch +45;
- Sicilian 2...♘c6 3.♗b5 +37;
- Sicilian 2...d6 3.♗b5+ +40;
- Petroff main line +37;
- Italian +30;
- Spanish Berlin 5.♖e1 with ♗f1 +42;
- Spanish 6.d3 +26"

What is your opinion?
I guess that most agree that it is a draw with perfect play but I added also 2 options that white is winning or black is winning.

I have done a basic statistical analysis using Chess-DB web of human games (1. e4 e5 2. Nf3 Nc6 3. Bc4 Bc5).

Code: Select all

Score = White + Draw / 2
Std. Dev. = sqrt{ [ Score * (1 - Score) - (Draw Ratio) / 4 ] / (Games - 1) }

Elo: compute Elo of Score.
Error bars (95% confidence level): compute Elo of Score ± 1.96 * (Std. Dev.).

Code: Select all

Chess-DB. Opening and Position Explorer.

Rating 2500 ± 100

-------------------------------------------------------
Move    Eval    Depth   Games   White%   Draw%   Black%
-------------------------------------------------------
Nc3     +0.3    28      290     26%      32%     42% 
d3      +0.26   28      1014    31%      38%     31% 
O-O     +0.24   28      1205    37%      31%     32% 
c3      +0.22   26      4370    36%      34%     30% 
b4      -0.6    24      947     36%      23%     41% 
d4      ...             29      24%      24%     52% 
Qe2     ...             9       44%      55%      1% 
a3      ...             4       50%       0%     50% 
Bb3     ...             4       75%       0%     25%
-------------------------------------------------------

Games: 7872

White ~ 35.13%
Draw  ~ 32.61%
Black ~ 32.26%

Score     ~ 51.44%
Std. Dev. ~  0.46%

Elo (95% confidence level) ~ +10.0 ± 6.3

Code: Select all

Chess-DB. Opening and Position Explorer.

Rating 2600 ± 100

-------------------------------------------------------
Move    Eval    Depth   Games   White%   Draw%   Black%
-------------------------------------------------------
Nc3     +0.3    28      213     28%      32%     40% 
d3      +0.26   28      696     30%      40%     30% 
O-O     +0.24   28      815     36%      34%     30% 
c3      +0.22   26      3153    35%      37%     28% 
b4      -0.6    24      601     35%      25%     40% 
d4      ...             18      16%      16%     68% 
Qe2     ...             6       33%      50%     17% 
a3      ...             3       33%       0%     67% 
Bb3     ...             3       33%       0%     67%
-------------------------------------------------------

Games: 5508

White ~ 34.18%
Draw  ~ 35.34%
Black ~ 30.48%

Score     ~ 51.85%
Std. Dev. ~  0.54%

Elo (95% confidence level) ~ +12.8 ± 7.4

Code: Select all

Chess-DB. Opening and Position Explorer.

Rating 2700 ± 100

-------------------------------------------------------
Move    Eval    Depth   Games   White%   Draw%   Black%
-------------------------------------------------------
Nc3     +0.3    28      136     32%      33%      35% 
d3      +0.26   28      377     28%      44%      28% 
O-O     +0.24   28      424     32%      41%      27% 
c3      +0.22   26      1935    32%      44%      24% 
b4      -0.6    24      254     34%      31%      35% 
d4      ...             6        0%       0%     100% 
Qe2     ...             2        0%      50%      50% 
h3      ...             2        0%       0%     100% 
Bf1     ...             2        0%       0%     100%
-------------------------------------------------------

Games: 3138

White ~ 31.56%
Draw  ~ 41.93%
Black ~ 26.51%

Score     ~ 52.52%
Std. Dev. ~  0.68%

Elo (95% confidence level) ~ +17.6 ± 9.3

************************

SUMMARY:

Code: Select all

Rating 2500 ± 100 ==> +10.0 ± 6.3
Rating 2600 ± 100 ==> +12.8 ± 7.4
Rating 2700 ± 100 ==> +17.6 ± 9.3

I hope no typos. Italian Opening is a very balanced opening IMHO and it is (it should be) a draw with perfect play / correct preparation, just like many other regular and well studied openings. It gives a good game to both sides.

Last but not least, a MultiPV = 7 search up to depth 26 with SF 10 (2 cores, non deterministic) returns a white advantage of 25 cp, which is similar to the white advantage on the opening position.

Code: Select all

FEN: r1bqk1nr/pppp1ppp/2n5/2b1p3/2B1P3/5N2/PPPP1PPP/RNBQK2R w KQkq - 4 4 

Stockfish_10_x32:

---------------------------------------------------------------------------
[...]
---------------------------------------------------------------------------
 26/34	08:19	 266.650.428	534.379	+0,13	Nb1c3 Ng8f6 d2d3 OO OO h7h6 Nc3a4 Bc5e7 Bc1d2 d7d6 h2h3 Nc6d4 Na4c3 Nd4e6 a2a4 c7c6 Rf1e1 a7a5 Nc3e2 d6d5 e4xd5 c6xd5 Bc4b3 e5e4 d3xe4 d5xe4 Nf3e5 Ne6c5 Bd2e3 Nc5xb3 c2xb3
 26/34	08:19	 266.650.428	534.379	+0,16	h2h3 Ng8f6 d2d3 OO c2c3 d7d6 OO h7h6 Rf1e1 a7a5 a2a4 Bc8e6 Bc4xe6 f7xe6 Nb1a3 Bc5xa3 Ra1xa3 Nf6d7 Qd1b3 Nd7c5 Qb3c4 Qd8e8 Bc1e3 b7b6 Ra3a1 Qe8g6 Be3xc5 b6xc5 Qc4b5
 26/34	08:19	 266.650.428	534.379	+0,18	a2a4 Ng8f6 d2d3 OO OO h7h6 c2c3 a7a5 Nb1d2 d7d6 h2h3 Bc8e6 Bc4xe6 f7xe6 Nd2c4 Qd8e8 Kg1h1 Ra8d8 Nc4xa5 Nc6xa5 b2b4 d6d5 b4xc5 d5xe4 Nf3xe5 e4xd3 Ne5xd3
 26/39	08:19	 266.650.428	534.379	+0,18	Bc4b3 d7d6 OO Ng8f6 d2d3 OO c2c3 h7h6 Rf1e1 a7a5 Nb1d2 Bc8e6 Nd2f1 Bc5b6 Bb3a4 Nc6e7 h2h3 c7c6 Nf1g3 Ne7g6 d3d4 e5xd4 c3xd4 d6d5 e4e5 Nf6e4 Ng3xe4 d5xe4 Re1xe4 Be6d5
 26/37	08:19	 266.650.428	534.379	+0,23	c2c3 Ng8f6 d2d3 d7d6 OO OO Nb1d2 a7a5 a2a4 h7h6 Rf1e1 Nc6e7 Nd2f1 Ne7g6 Nf1g3 c7c6 h2h3 d6d5 Bc4b3 Qd8b6 d3d4 e5xd4 c3xd4 Bc5b4 Re1e2 Qb6a7 e4xd5
 26/34	08:19	 266.650.428	534.379	+0,25	OO Ng8f6 d2d3 d7d6 Nb1c3 OO Nc3a4 Bc5b6 c2c3 Nc6a5 Na4xb6 a7xb6 Bc4b5 Bc8d7 Bb5xd7 Qd8xd7 Bc1g5 Qd7e6 b2b4 Na5c6 a2a4 h7h6 Bg5d2 d6d5 e4xd5 Qe6xd5 b4b5 Nc6e7 c3c4 Qd5xd3 Nf3xe5
 26/34	08:19	 266.650.428	534.379	+0,25	d2d3 Ng8f6 OO d7d6 Nb1c3 OO Nc3a4 Bc5b6 c2c3 Nc6a5 Na4xb6 a7xb6 Bc4b5 Na5c6 Rf1e1 Bc8d7 a2a4 h7h6 Bc1d2 Nc6e7 Bb5c4 Qd8e8 Bc4b3 b6b5 a4xb5 Ra8xa1 Qd1xa1 Bd7xb5 Bb3c2
---------------------------------------------------------------------------

************************

[...] and the critical positions contain material imbalances that engines don't seem to play well from the white side. [...]

Just by curiosity, could you post some of these critical positions, please? It is never too late to learn something new. Thanks in advance.

Regards from Spain.

Ajedrecista.

[Uri Blass]

My claims about it has been about unassisted engine chess (blitz from 5 0 to 12 2 time control) and assisted chess at correspondence time controls (+1 year per game), based mainly in positions where Stockfish prefers white but it's black that has the advantage, and the critical positions contain material imbalances that engines don't seem to play well from the white side. An opening book with these lines will score better from the black side of the Italian against itself and will outperform old books unprepared (I can give dates, it seems the relevant lines were found/uncorked in Febrary 2019 and were polished up to May 2019 - since then the Italian has become rare and black scores better when it is tried - again, because of unprepared opponents with old books.)

However, we're discussing apples and oranges as Larry Kaufman is talking about the Italian in OTB human games where engines are forbidden, in such cases he's probably right, and it's possible the engine-engine lines useful in corr chess have no use in practice of human games (black needs engine-like tactical awareness to succeed.) I still expect that a black win on the Italian will be played in an important human game, and to win one at corr chess, but black's advantage in it is so small I won't hold my breath.

I understand now that you do not claim that the italien is better for black but only that black can get an advantage against old books.

It is clearly a different claim and it is not clear that engines can get a position when they show advantage for black against kaufman's new repertoire.

https://www.newinchess.com/kaufman-s-ne ... 14db0a6035

[D Sceviour]

Here is how I treat the drawish Italian Opening. For blitz chess I sometimes play the Hungarian Defense 3...Be7. This gets white out of the book and ends any ideas that white can play Ng5 as an attack. It can take time off the opponents clock as he tries to find the continuation.

For computer chess, I have also added the Hungarian Defense. Recently in an ICS match, Arasan-Schooner played to under a 20 move draw by repetition after 3...Bc5. Not once, but twice in a row! To get around this 3...Be7 has been added to my polyglot book. This offers better possibilities for a decisive result.

[Ovyron]

I understand now that you do not claim that the italien is better for black but only that black can get an advantage against old books.

But new books with those lines score better against themselves from the black side, what is the explanation for that?

It could be argued that all my claims are null, since I'm mostly talking about Stockfish (and related branches)'s games, so it wouldn't be that the Italian is better for black in general, only if white is Stockfish. But I've been able to win with the Italian from the black side against much faster hardware and against Leela in a decent GPU, that was enough to convince me (mainly because I was never able to win against such opponents from the white side, sometimes they just ran over me, so something is going on there).

It is clearly a different claim and it is not clear that engines can get a position when they show advantage for black against kaufman's new repertoire.

The question is, was Larry aware of these black lines before building his new repertoire? Because it's clear these lines are mostly secret, held by privileged individuals, and I only know about them because I was extremely lucky that they were shared to me. I'm even afraid to talk about the resulting positions because people are smart enough to use retrograde analysis to figure out the moves from the Italian required to reach those positions.

I get the impression that either Larry didn't know about them, or he looked at them and concluded they had no value in human chess (because they're lines designed for computers to beat other computers), but only he can answer such questions.

Note the Italian isn't alone in this category. What is the best move in this position in the Sicilian?

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6

[d]rnbqkb1r/1p2pppp/p2p1n2/8/3NP3/2N5/PPP2PPP/R1BQKB1R w KQkq -

Well, it turns out the best move is 6.h3!! It's so good it almost kills the Sicilian (you're better playing 1...e5 just to avoid this), but it probably kills the Najdorf (so what happened?... that in computer chess the Sicilian became very rare!), but this is only because engines from the black side have a bigger problem dealing with it than with the other moves, wouldn't expect someone building a repertoire for 6.h3 and recommending it being played in human games, because the critical lines were also designed for computers to beat other computers.

I repeat that it's discussing apples and oranges and that these things are only useful against engines, but if you're going to play an engine it'll beat you anyway so playing the Italian as black or 6.h3 as white isn't really going to improve your chances.

[Uri Blass]

I understand now that you do not claim that the italien is better for black but only that black can get an advantage against old books.

But new books with those lines score better against themselves from the black side, what is the explanation for that?

It could be argued that all my claims are null, since I'm mostly talking about Stockfish (and related branches)'s games, so it wouldn't be that the Italian is better for black in general, only if white is Stockfish. But I've been able to win with the Italian from the black side against much faster hardware and against Leela in a decent GPU, that was enough to convince me (mainly because I was never able to win against such opponents from the white side, sometimes they just ran over me, so something is going on there).

It is clearly a different claim and it is not clear that engines can get a position when they show advantage for black against kaufman's new repertoire.

The question is, was Larry aware of these black lines before building his new repertoire? Because it's clear these lines are mostly secret, held by privileged individuals, and I only know about them because I was extremely lucky that they were shared to me. I'm even afraid to talk about the resulting positions because people are smart enough to use retrograde analysis to figure out the moves from the Italian required to reach those positions.

I get the impression that either Larry didn't know about them, or he looked at them and concluded they had no value in human chess (because they're lines designed for computers to beat other computers), but only he can answer such questions.

Note the Italian isn't alone in this category. What is the best move in this position in the Sicilian?

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6

[d]rnbqkb1r/1p2pppp/p2p1n2/8/3NP3/2N5/PPP2PPP/R1BQKB1R w KQkq -

Well, it turns out the best move is 6.h3!! It's so good it almost kills the Sicilian (you're better playing 1...e5 just to avoid this), but it probably kills the Najdorf (so what happened?... that in computer chess the Sicilian became very rare!), but this is only because engines from the black side have a bigger problem dealing with it than with the other moves, wouldn't expect someone building a repertoire for 6.h3 and recommending it being played in human games, because the critical lines were also designed for computers to beat other computers.

I repeat that it's discussing apples and oranges and that these things are only useful against engines, but if you're going to play an engine it'll beat you anyway so playing the Italian as black or 6.h3 as white isn't really going to improve your chances.

I do not believe that a move is good only against engines.
Top humans can play many games against engines from the position after 6.h3 when they are black and simply memorize the lines that the engines beat them in order to use the same lines that the engines beat them against other humans.

There are 2 options:
1)They win the game thanks to a prepared line.
2)The opponent play something different when they get a good position that they do not memorize how to continue.

I think that even in the last case they can generalize from their experience against engines and play correctly moves that they know to be good in similiar situations.

[Master Om]

where Stockfish prefers white but it's black that has the advantage, and the critical positions contain material imbalances that engines don't seem to play well from the white side.

If white does not give chance all black will get is equality, Don;t except White plays like a noob against black. If Ruy Lopez stands out then Italian will so as both have same themes.

[Ovyron]

I think that even in the last case they can generalize from their experience against engines and play correctly moves that they know to be good in similiar situations.

Engines and humans play so completely different chess that even if you get a variation where you perform relatively well against engines they might be useless against humans.

I tried this back in Rebel Decade times, because I could beat the engine at Depth 6. So what I did was using the same openings and moves I was using to defeat Rebel Depth 6 against humans. The result was that I performed worse against humans than with my usual repertoire, because humans played nothing like Rebel, and they wouldn't play the blunders that were allowing me to beat Rebel, so it was useless.

[Ovyron]

If white does not give chance all black will get is equality, Don;t except White plays like a noob against black. If Ruy Lopez stands out then Italian will so as both have same themes.

It is weird that this comes from you Master Om, because in 2011 you were holding that the Italian was a superior defense for black than the Spanish. Back then I held the Spanish was the best defense, so we started a series of matches where I'd try to beat your Italian and you'd try to beat my Spanish.

The result should be well know by you:

{url=http://rybkaforum.net/cgi-bin/rybkaforu ... ?tid=22314]Game 1 1-0[/url] (you beat my Spanish)
{url=http://rybkaforum.net/cgi-bin/rybkaforu ... ?tid=23346]Game 2 1-0[/url] (you beat my Spanish)
{url=http://rybkaforum.net/cgi-bin/rybkaforu ... ?tid=25302]Game 3 1-0[/url] (I resigned the Spanish)
Game 4 1/2-1/2 (I drew with the Spanish)

It took me four games to be able to find a defense.

Apparently we never played the Italian part of the match but you used to be on the other side of these discussions, where it'd be more likely for white to "give chance" to black in the Italian than in the Spanish, where my best lines could never beat the opponents that I beat with the Italian.