The space advantage that was non-existent

[Lyudmil Tsvetkov]

I will post another diagram (hope the mods are not going to ban me for overposting).

[D]r2q1rk1/1b1nbppp/4p3/3pP3/p1pP4/PpP2N1P/1P3PP1/R1BQRNK1 b 0 1

How would you this position?

My take on that: supposing that mobility is about equal, let's concentrate on the pawn structure. Black has enormous space advantage - if a pawn on the 5th rank gets some 10 centipawns (cps) bonus and a pawn on the 6th rank some 30cps, black should be in advantage by some 40cps, as e5 is the only white pawn that has crossed the center line. And many entities could indeed see the position in favour of black (and there are also a pair of bishops there). However, I think this is wrong. The important thing is that the queen side is fully closed, so that black can not possibly make use of its space advantage there (unless there are some tactical solutions, difficult to imagine at the moment). This space advantage is simply unimportant and it might reasonably be at least halved. At the same time, the only white pawn gaining space advantage on e5 does so on a side that is not closed, and chances are that this will help white to gradually increase its advantage on that side, which will matter, as long as the side is not fully closed. The pawn on e5 is leading a group of diagonally connected pawns (or is a peak of a larger chain of 4 pawns) that is conducive to white's attacking chances. Actually, as it is close to the enemy king, it might be worth some 15-20 cps. The decisive factor is, of course, the fact that white has advantage (in terms of space and attacking chances) on the side that is still open: and that would add another 50cps for white in the balance, assigned because by closing the entire queen side, black now simply does not have counterplay, and will have to patiently wait for the white attack to unfold. As a result, the real score scould be not some 50-80cps in favour of black, but some 50-80cps in favour of white. The best black could do is attempt blocking the entire king side (difficult to imagine as white's attack will gradually build up, enjoying the support of pieces), or, at the eraliest possible, attempt transferring the black king to the queen side (which is still uncertain if at all possible).

So, I would be very interested in your feedback - are there any engines/humans, actually seeing white in advantage?

Best, Lyudmil

[Lyudmil Tsvetkov]

I will post another diagram (hope the mods are not going to ban me for overposting).

[D]r2q1rk1/1b1nbppp/4p3/3pP3/p1pP4/PpP2N1P/1P3PP1/R1BQRNK1 b 0 1

How would you this position?

My take on that: supposing that mobility is about equal, let's concentrate on the pawn structure. Black has enormous space advantage - if a pawn on the 5th rank gets some 10 centipawns (cps) bonus and a pawn on the 6th rank some 30cps, black should be in advantage by some 40cps, as e5 is the only white pawn that has crossed the center line. And many entities could indeed see the position in favour of black (and there are also a pair of bishops there). However, I think this is wrong. The important thing is that the queen side is fully closed, so that black can not possibly make use of its space advantage there (unless there are some tactical solutions, difficult to imagine at the moment). This space advantage is simply unimportant and it might reasonably be at least halved. At the same time, the only white pawn gaining space advantage on e5 does so on a side that is not closed, and chances are that this will help white to gradually increase its advantage on that side, which will matter, as long as the side is not fully closed. The pawn on e5 is leading a group of diagonally connected pawns (or is a peak of a larger chain of 4 pawns) that is conducive to white's attacking chances. Actually, as it is close to the enemy king, it might be worth some 15-20 cps. The decisive factor is, of course, the fact that white has advantage (in terms of space and attacking chances) on the side that is still open: and that would add another 50cps for white in the balance, assigned because by closing the entire queen side, black now simply does not have counterplay, and will have to patiently wait for the white attack to unfold. As a result, the real score scould be not some 50-80cps in favour of black, but some 50-80cps in favour of white. The best black could do is attempt blocking the entire king side (difficult to imagine as white's attack will gradually build up, enjoying the support of pieces), or, at the eraliest possible, attempt transferring the black king to the queen side (which is still uncertain if at all possible).

So, I would be very interested in your feedback - are there any engines/humans, actually seeing white in advantage?

Best, Lyudmil

Just to add a bit more comments.
White has a huge advantage, probably winning, if the black king is impossible to transfer on the queen side, that is closed and safe. The only white pawn that gains space advantage on e5 is by far more important than all black pawns gaining space advantage on the queen side. As playing f5 for the moment and most probably in the future seems impossible, BLACK DOES NOT HAVE COUNTERPLAY and is doomed to patiently wait while the white attack gradually unfolds. White can play Ng3, Nh2-g4, f4 if tactically justified, etc., gradually increasing the pressure.

Lyudmil

PS. I see that no one posted here, so probably most engines think black is better.

[Eelco de Groot]

PS. I see that no one posted here, so probably most engines think black is better.

I haven't tried this yet, I don't know. Maybe it would depend also on King Safety. This has been on my "To Do" list in a way, because I think it could be implemented for the computer somehow. Here is another example of such a 'Sniper Wall' as you could call it; it shields attackers from the defensive side, in one of Jon Dart's testpositions. The wall is not so long but a pawn threat with f3 is still present also:

[D]r6k/N1Rb2bp/p2p1nr1/3Pp2q/1P2Pp1P/5N2/P3QBP1/4R1K1 b - - bm Bh3; id "arasan16.206"; c0 "Persson-Gerhardt, Simon Webb mem email 2007";
1...Bh3 2. Ng5 Bg4 3.Qxa6 and in spite of two passed pawns on the the other side of the "wall", after a while engines can see White is lost. But it can take a while because the engines have to calculate this through. Seeing the "wall" would help...


'Sniper wall' between Israeli settlement of Gilo and an adjacent Palestinian village. Strange to see this described as a work of art with an "elegant composition"

Eelco

[Lyudmil Tsvetkov]

Hi Eelco.
Yes, seeing the wall, seeing Gilo, will certainly help.
I think Jon's position is more tactical in nature, as white has a number of weak squares (g3, g4), rooks directly attacking the king, pinning possibilities, etc.

Thanks

[lucasart]

Well here's what my engine DiscoCheck thinks of this:
[d]r2q1rk1/1b1nbppp/4p3/3pP3/p1pP4/PpP2N1P/1P3PP1/R1BQRNK1 b - - 0 1
eval = 77 cp, that is without searching, just "looking" at the position. this is typically some stupid piece on square + mobility "positional advantage". Clearly DiscoCheck's chess knowledge is very minimalistic, so it has not clue what is going on here...

Now let's do some searching:

info score cp 85 depth 1 nodes 35 time 0 pv d8b6
info score cp 58 depth 2 nodes 206 time 0 pv d8b6 c1g5
info score cp 59 depth 3 nodes 585 time 0 pv a8c8 c1f4 d8b6
info score cp 52 depth 4 nodes 962 time 1 pv a8c8 c1f4 d8b6 f1e3
info score cp 52 depth 5 nodes 2226 time 3 pv d7b8 c1f4 b8c6 f1e3 d8c7
info score cp 47 depth 6 nodes 3494 time 5 pv d7b8 c1f4 b8c6 f1e3 d8c7 d1d2
info score cp 45 depth 7 nodes 6979 time 10 pv d7b8 c1f4 b8c6
info score cp 40 depth 8 nodes 14982 time 18 pv d7b8 c1f4 b8c6 f1e3 d8c7 d1d2 a8d8 a1d1
info score cp 35 depth 9 nodes 34049 time 30 pv d7b8 c1f4 b8c6
info score cp 34 depth 10 nodes 87848 time 63 pv a8c8 c1f4 h7h6
info score cp 33 depth 11 nodes 199704 time 133 pv f8e8 d1e2 h7h6
info score cp 42 depth 12 nodes 250105 time 162 pv f8e8 d1e2 h7h6 f1g3 d7f8
info score cp 33 depth 13 nodes 359217 time 226 pv f8e8 d1e2 h7h6 f1g3 d7f8 g3h5 f8g6 h5f4 g6f4 c1f4 b7c6 a1d1 d8c7
info score cp 41 depth 14 nodes 515716 time 314 pv f8e8 d1e2
info score cp 39 depth 15 nodes 846077 time 498 pv f8e8 d1e2
info score cp 44 depth 16 nodes 1446701 time 837 pv f8e8 d1e2
info score cp 41 depth 17 nodes 2819315 time 1600 pv f8e8 d1e2
info score cp 41 depth 18 nodes 4062805 time 2279 pv f8e8
info score cp 42 depth 19 nodes 6955478 time 3852 pv f8e8 d1e2
info score cp 36 depth 20 nodes 14737290 time 8245 pv f8e8 d1e2 d7f8
info score cp 29 depth 21 nodes 33020554 time 18534 pv f8e8 d1e2
info score cp 29 depth 22 nodes 48459014 time 26996 pv f8e8
info score cp 29 depth 23 nodes 64300989 time 35570 pv f8e8 d1e2
info score cp 30 depth 24 nodes 111029827 time 61283 pv f8e8 d1e2
info score cp 22 depth 25 nodes 247440955 time 139313 pv f8e8
info score cp 18 depth 26 nodes 512427309 time 293323 pv f8e8 d1e2 d7f8
...

And slowly DiscoCheck is going deeper and deeper: who knows ? He might find some oil eventually! In the meantime, it appears that the deeper the search the more evident it becomes that black's advantage is not real. I don't know if this is winning for white, but at least we can say that black's advantage is not real.