What is a backward pawn?

[Lyudmil Tsvetkov]

My take on this is that you are somewhat mixing terms.

A pawn that is defended by a pawn is not considered backward in any book I have (g6 in your first example as it is defended by h7). The fact that it can not safely advance is not so important when it is perfectly safe where it is. If you want to talk about it being immobile, that is a different issue.

the "unopposed" requirement stems from the rook, in the books I have read over the years. A pawn that is backward (can not be safely defended by a friendly pawn) is generally considered to be worse if it is (a) on a half-open file (no enemy pawn in front of it AND (b) the enemy has rooks to attack it.

IMHO, a pawn that is backward and unopposed is not any weaker or stronger than a backward pawn that is opposed, if there are no enemy rooks to contend with. Any backward pawn is weak, and inviting to enemy kings in the endgame. backward pawns on half-open files make inviting targets even in the middle game if you have rooks to attack on the file.

I've seen games decided by two types of backward pawns, repeatedly.

The first is the classic backward pawn on a half-open file, which is just as bad as an isolated pawn on a half-open file, when the enemy has rooks. Pile up on the pawn, tying up the opponent by making him defend, and either win the pawn outright, or switch to some other idea once he is tied down and not coordinated very well.

The second is a pawn that is backward, but the file is not half-open. If you have just one, you might manage to defend it with your king, depending on the position. If you have two, the enemy king can make you commit to saving one and then to eat the other pawn chain. Doesn't matter whether the file is half open or not.

[d]6k1/5p2/4p1p1/3p2P1/1p1P4/p7/P7/6K1 w - - 0 1

I tend to concur with Bob's definition. In the diagram above the e6 pawn is still defended, and thus not vulnerable to direct attack, so that's an argument against calling it backward.

However, since it cannot safely advance, there is an element of backwardness about it, and the square in front of it (e5) can serve as a good outpost for White pieces. These considerations should justify some sort of a penalty for such a pawn, but a lesser one that a fully backward [undefendable by a fellow pawn] pawn on an open file.

The Black b-pawn, even though backward, could be pushed in some cases creating a dangerous advanced passed pawn. It's not as clear whether a backward pawn that sits in the opponent's half of the board deserves any significant penalty.

CL

Carl, whether e6 is defended or not is another matter, completely unrelated to backwardness, i.e. underdevelopedness, inability to advance. Defended and undefended pawns already get their separate bonus/penalty points. We are taling here about the inherent inability of the e6 pawn to advance. That matters very much, pawns are there to advance, help in the attack, promote, etc. If your pawns do not advance to attack the enemy and promote, then it will be your opponent's pawns which will do that to win the game. So that a passive state of inability to advance should always be penalised, regardless of whether such pawns are defended or not. The defended/undefended bonus/penalty is just a small fraction of the penalty for a backward pawn unable to advance without being lost, so that backwardness is the much more important concept.

[d]6k1/5p2/4p1p1/n2p2P1/1p1P4/p7/P7/2N3K1 w - - 0 1
You see above that in the usual case with pieces present, b4 does not represent any real danger of supporting promotion, but is still unable to advance without being lost. So that it is still backward. Of course, a backward pawn on the 4th rank deserves a very small penalty, much less than a backward on the 5th. That is what we have been talking about, the introduction of a unified concept that works across ranks, i.e. a good practical eval term.

How practical and good are current backward definitions that mix backward with weak, isolated, overextended, undefended?

Again, I think the definition I specified, with some 30cps penalty for the 7th rank, 15cps for the 6th rank, 8cps for the 5th rank, and 4cps for the 4th rank, is a good, working definition, unified and applicable across ranks. When you apply a unified feature across ranks, chances are you will get the most out of it.

I ask myself again, who is to blame when modern top engines fully misunderstand backward pawns? And if they badly misplay such pawns, is it possible that the definitions used for backward are good and working ones? Why should you mix backward with isolated, overextended (meaning vertically isolated) and undefended? You already have definitions and penalties for isolated and undefended. No, the only conclusion I draw from that is that currently a complete chaos reigns in the computer chess world as to what actually a backward pawn is. And that is the main reason why engines play so weak with backward pawns.

I wonder who will be the first author who will drastically change his approach and definitions of backward pawns to easily gain (after tuning of course) another 50 elo strength?

[Lyudmil Tsvetkov]

My take on this is that you are somewhat mixing terms.

A pawn that is defended by a pawn is not considered backward in any book I have (g6 in your first example as it is defended by h7). The fact that it can not safely advance is not so important when it is perfectly safe where it is. If you want to talk about it being immobile, that is a different issue.

the "unopposed" requirement stems from the rook, in the books I have read over the years. A pawn that is backward (can not be safely defended by a friendly pawn) is generally considered to be worse if it is (a) on a half-open file (no enemy pawn in front of it AND (b) the enemy has rooks to attack it.

IMHO, a pawn that is backward and unopposed is not any weaker or stronger than a backward pawn that is opposed, if there are no enemy rooks to contend with. Any backward pawn is weak, and inviting to enemy kings in the endgame. backward pawns on half-open files make inviting targets even in the middle game if you have rooks to attack on the file.

I've seen games decided by two types of backward pawns, repeatedly.

The first is the classic backward pawn on a half-open file, which is just as bad as an isolated pawn on a half-open file, when the enemy has rooks. Pile up on the pawn, tying up the opponent by making him defend, and either win the pawn outright, or switch to some other idea once he is tied down and not coordinated very well.

The second is a pawn that is backward, but the file is not half-open. If you have just one, you might manage to defend it with your king, depending on the position. If you have two, the enemy king can make you commit to saving one and then to eat the other pawn chain. Doesn't matter whether the file is half open or not.

[d]6k1/5p2/4p1p1/3p2P1/1p1P4/p7/P7/6K1 w - - 0 1

I tend to concur with Bob's definition. In the diagram above the e6 pawn is still defended, and thus not vulnerable to direct attack, so that's an argument against calling it backward.

However, since it cannot safely advance, there is an element of backwardness about it, and the square in front of it (e5) can serve as a good outpost for White pieces. These considerations should justify some sort of a penalty for such a pawn, but a lesser one that a fully backward [undefendable by a fellow pawn] pawn on an open file.

The Black b-pawn, even though backward, could be pushed in some cases creating a dangerous advanced passed pawn. It's not as clear whether a backward pawn that sits in the opponent's half of the board deserves any significant penalty.

CL

I believe that Hans deeply thought that one of the major qualifiers for backward was for that pawn to be still in the lower ranks (2nd and 3rd). The ideal rank for most pawn is at 4th. If it is on 5th, and beyond, we might labelled it as over extended. It is even possible that a pawn in the 2nd and 3rd rank will get more penalty (aside from the pst) as game progresses, no need to calculate if its stop (a square in front) is attacked or not.

There is a difference, if you are on the 7th and can not advance, some limitations are set upon you, your freedom is limited. If you are on the 7th but can still advance safely, then it is just a temporary state of affairs that could change any time.

[carldaman]

My take on this is that you are somewhat mixing terms.

A pawn that is defended by a pawn is not considered backward in any book I have (g6 in your first example as it is defended by h7). The fact that it can not safely advance is not so important when it is perfectly safe where it is. If you want to talk about it being immobile, that is a different issue.

the "unopposed" requirement stems from the rook, in the books I have read over the years. A pawn that is backward (can not be safely defended by a friendly pawn) is generally considered to be worse if it is (a) on a half-open file (no enemy pawn in front of it AND (b) the enemy has rooks to attack it.

IMHO, a pawn that is backward and unopposed is not any weaker or stronger than a backward pawn that is opposed, if there are no enemy rooks to contend with. Any backward pawn is weak, and inviting to enemy kings in the endgame. backward pawns on half-open files make inviting targets even in the middle game if you have rooks to attack on the file.

I've seen games decided by two types of backward pawns, repeatedly.

The first is the classic backward pawn on a half-open file, which is just as bad as an isolated pawn on a half-open file, when the enemy has rooks. Pile up on the pawn, tying up the opponent by making him defend, and either win the pawn outright, or switch to some other idea once he is tied down and not coordinated very well.

The second is a pawn that is backward, but the file is not half-open. If you have just one, you might manage to defend it with your king, depending on the position. If you have two, the enemy king can make you commit to saving one and then to eat the other pawn chain. Doesn't matter whether the file is half open or not.

[d]6k1/5p2/4p1p1/3p2P1/1p1P4/p7/P7/6K1 w - - 0 1

I tend to concur with Bob's definition. In the diagram above the e6 pawn is still defended, and thus not vulnerable to direct attack, so that's an argument against calling it backward.

However, since it cannot safely advance, there is an element of backwardness about it, and the square in front of it (e5) can serve as a good outpost for White pieces. These considerations should justify some sort of a penalty for such a pawn, but a lesser one that a fully backward [undefendable by a fellow pawn] pawn on an open file.

The Black b-pawn, even though backward, could be pushed in some cases creating a dangerous advanced passed pawn. It's not as clear whether a backward pawn that sits in the opponent's half of the board deserves any significant penalty.

CL

Carl, whether e6 is defended or not is another matter, completely unrelated to backwardness, i.e. underdevelopedness, inability to advance. Defended and undefended pawns already get their separate bonus/penalty points. We are taling here about the inherent inability of the e6 pawn to advance. That matters very much, pawns are there to advance, help in the attack, promote, etc. If your pawns do not advance to attack the enemy and promote, then it will be your opponent's pawns which will do that to win the game. So that a passive state of inability to advance should always be penalised, regardless of whether such pawns are defended or not. The defended/undefended bonus/penalty is just a small fraction of the penalty for a backward pawn unable to advance without being lost, so that backwardness is the much more important concept.

[d]6k1/5p2/4p1p1/n2p2P1/1p1P4/p7/P7/2N3K1 w - - 0 1
You see above that in the usual case with pieces present, b4 does not represent any real danger of supporting promotion, but is still unable to advance without being lost. So that it is still backward. Of course, a backward pawn on the 4th rank deserves a very small penalty, much less than a backward on the 5th. That is what we have been talking about, the introduction of a unified concept that works across ranks, i.e. a good practical eval term.

How practical and good are current backward definitions that mix backward with weak, isolated, overextended, undefended?

Again, I think the definition I specified, with some 30cps penalty for the 7th rank, 15cps for the 6th rank, 8cps for the 5th rank, and 4cps for the 4th rank, is a good, working definition, unified and applicable across ranks. When you apply a unified feature across ranks, chances are you will get the most out of it.

I ask myself again, who is to blame when modern top engines fully misunderstand backward pawns? And if they badly misplay such pawns, is it possible that the definitions used for backward are good and working ones? Why should you mix backward with isolated, overextended (meaning vertically isolated) and undefended? You already have definitions and penalties for isolated and undefended. No, the only conclusion I draw from that is that currently a complete chaos reigns in the computer chess world as to what actually a backward pawn is. And that is the main reason why engines play so weak with backward pawns.

I wonder who will be the first author who will drastically change his approach and definitions of backward pawns to easily gain (after tuning of course) another 50 elo strength?

Certainly, the definition of a 'backward pawn' can be refined to something better suited for the computer age. Hans Kmoch dates back to an era when chess engines could not even be dreamed of.

Yes Lyudmil, I agree that a backward pawn as you envision it should be penalized much as you suggest. No arguments there, but it is easy to get lost in the semantics, where even knowledgeable folks adhere to somewhat different definitions, rightly or wrongly.

[Lyudmil Tsvetkov]

My take on this is that you are somewhat mixing terms.

A pawn that is defended by a pawn is not considered backward in any book I have (g6 in your first example as it is defended by h7). The fact that it can not safely advance is not so important when it is perfectly safe where it is. If you want to talk about it being immobile, that is a different issue.

the "unopposed" requirement stems from the rook, in the books I have read over the years. A pawn that is backward (can not be safely defended by a friendly pawn) is generally considered to be worse if it is (a) on a half-open file (no enemy pawn in front of it AND (b) the enemy has rooks to attack it.

IMHO, a pawn that is backward and unopposed is not any weaker or stronger than a backward pawn that is opposed, if there are no enemy rooks to contend with. Any backward pawn is weak, and inviting to enemy kings in the endgame. backward pawns on half-open files make inviting targets even in the middle game if you have rooks to attack on the file.

I've seen games decided by two types of backward pawns, repeatedly.

The first is the classic backward pawn on a half-open file, which is just as bad as an isolated pawn on a half-open file, when the enemy has rooks. Pile up on the pawn, tying up the opponent by making him defend, and either win the pawn outright, or switch to some other idea once he is tied down and not coordinated very well.

The second is a pawn that is backward, but the file is not half-open. If you have just one, you might manage to defend it with your king, depending on the position. If you have two, the enemy king can make you commit to saving one and then to eat the other pawn chain. Doesn't matter whether the file is half open or not.

[d]6k1/5p2/4p1p1/3p2P1/1p1P4/p7/P7/6K1 w - - 0 1

I tend to concur with Bob's definition. In the diagram above the e6 pawn is still defended, and thus not vulnerable to direct attack, so that's an argument against calling it backward.

However, since it cannot safely advance, there is an element of backwardness about it, and the square in front of it (e5) can serve as a good outpost for White pieces. These considerations should justify some sort of a penalty for such a pawn, but a lesser one that a fully backward [undefendable by a fellow pawn] pawn on an open file.

The Black b-pawn, even though backward, could be pushed in some cases creating a dangerous advanced passed pawn. It's not as clear whether a backward pawn that sits in the opponent's half of the board deserves any significant penalty.

CL

Carl, whether e6 is defended or not is another matter, completely unrelated to backwardness, i.e. underdevelopedness, inability to advance. Defended and undefended pawns already get their separate bonus/penalty points. We are taling here about the inherent inability of the e6 pawn to advance. That matters very much, pawns are there to advance, help in the attack, promote, etc. If your pawns do not advance to attack the enemy and promote, then it will be your opponent's pawns which will do that to win the game. So that a passive state of inability to advance should always be penalised, regardless of whether such pawns are defended or not. The defended/undefended bonus/penalty is just a small fraction of the penalty for a backward pawn unable to advance without being lost, so that backwardness is the much more important concept.

[d]6k1/5p2/4p1p1/n2p2P1/1p1P4/p7/P7/2N3K1 w - - 0 1
You see above that in the usual case with pieces present, b4 does not represent any real danger of supporting promotion, but is still unable to advance without being lost. So that it is still backward. Of course, a backward pawn on the 4th rank deserves a very small penalty, much less than a backward on the 5th. That is what we have been talking about, the introduction of a unified concept that works across ranks, i.e. a good practical eval term.

How practical and good are current backward definitions that mix backward with weak, isolated, overextended, undefended?

Again, I think the definition I specified, with some 30cps penalty for the 7th rank, 15cps for the 6th rank, 8cps for the 5th rank, and 4cps for the 4th rank, is a good, working definition, unified and applicable across ranks. When you apply a unified feature across ranks, chances are you will get the most out of it.

I ask myself again, who is to blame when modern top engines fully misunderstand backward pawns? And if they badly misplay such pawns, is it possible that the definitions used for backward are good and working ones? Why should you mix backward with isolated, overextended (meaning vertically isolated) and undefended? You already have definitions and penalties for isolated and undefended. No, the only conclusion I draw from that is that currently a complete chaos reigns in the computer chess world as to what actually a backward pawn is. And that is the main reason why engines play so weak with backward pawns.

I wonder who will be the first author who will drastically change his approach and definitions of backward pawns to easily gain (after tuning of course) another 50 elo strength?

Certainly, the definition of a 'backward pawn' can be refined to something better suited for the computer age. Hans Kmoch dates back to an era when chess engines could not even be dreamed of.

Yes Lyudmil, I agree that a backward pawn as you envision it should be penalized much as you suggest. No arguments there, but it is easy to get lost in the semantics, where even knowledgeable folks adhere to somewhat different definitions, rightly or wrongly.

That is the point, instead of sticking to the routine, better try some basic things to pinpoint backward pawn definitions and efficiency:

- does it make sense to consider backward pawns when opposed at all, is giving different values to opposed and not opposed backwards better than skipping opposed completely
- should the enemy pawn attacking the stop square and making you backward necessarily be blocked or not; try to measure the impact of both versions
- could chain/connected pawns also be backward, or only undefended pawns; try measuring the impact of both versions
- do backwards exist only on the 7th and 6th rank, or other ranks are also possible; try to measure the impact of both
- what is the influence of minor pieces attacking the stop square, etc.

Again, the only reason why I am reasing this issue is that backward pawns are the worst pawn element in engine play, and by far. It is quite possible that engines play better even closed positions, the KID and opposite-coloured bishops, elements they are renowned for playing badly. Engine authors are beguiled on their engine's performance with backward pawns for the simple reason that it is impossible for them to measure any deficiencies in this field, as engines are pitted against other engines that also do not understand backward pawns.

Have you noticed me sometime saying: 'Engines play very bad in attacking, or engines play very bad in open positions' ? No, I am raising only issues that validly exist.

[BeyondCritics]

Both of you miss this point: The definition of "backward pawn" in a chess engine is an extremly bold definition designed solely to be awarded a single, hefty penalty later

.

So what is not extreme? Why it is bold?

Statical chess engine knowledge concentrates around a few meaningful concepts. If you match this 1:1 with human evaluation, the chess engine evaluation will look bold and extreme. That was my point. This is expected, since there is an "impedance mismatch" as engineers say, between "TheBrain" and "VonNeumann".
For concreteness an example: A problem of the "backward" pawn is, that it is "weak". Why is it weak? Because it can be attacked. So obviously, if the pawn can _not_ be attacked, it is _not_ weak.
A decent chess player can analyze a position and estimate meaningful attack vectors on his pawns. If there is no attack vector to a backward pawn, it is _not_ considered as weakness, trust me on that.
Compare this with fruit, it will apply boldly the penalty. Even with a very deep search, there is no way out, you will still get the penalty.

Dont get me wrong here: I don't say the programmers got it wrong. I just point to the difference in attacking chess.

Distinguish carefully:
"Backward pawn" for a chess player <-> "Backward pawn" for a chess engine.

Backward pawn from a human generally works fine when applied to engine.

I know that. My point was, that concepts are not identical. Chess players do have a far more advanced understanding of backward pawns than chess engines (statically!) have.

Thanks for raising some interesting points.

Thank you for for reading me.

[Lyudmil Tsvetkov]

Both of you miss this point: The definition of "backward pawn" in a chess engine is an extremly bold definition designed solely to be awarded a single, hefty penalty later

.

So what is not extreme? Why it is bold?

Statical chess engine knowledge concentrates around a few meaningful concepts. If you match this 1:1 with human evaluation, the chess engine evaluation will look bold and extreme. That was my point. This is expected, since there is an "impedance mismatch" as engineers say, between "TheBrain" and "VonNeumann".
For concreteness an example: A problem of the "backward" pawn is, that it is "weak". Why is it weak? Because it can be attacked. So obviously, if the pawn can _not_ be attacked, it is _not_ weak.
A decent chess player can analyze a position and estimate meaningful attack vectors on his pawns. If there is no attack vector to a backward pawn, it is _not_ considered as weakness, trust me on that.
Compare this with fruit, it will apply boldly the penalty. Even with a very deep search, there is no way out, you will still get the penalty.

Dont get me wrong here: I don't say the programmers got it wrong. I just point to the difference in attacking chess.

Distinguish carefully:
"Backward pawn" for a chess player <-> "Backward pawn" for a chess engine.

Backward pawn from a human generally works fine when applied to engine.

I know that. My point was, that concepts are not identical. Chess players do have a far more advanced understanding of backward pawns than chess engines (statically!) have.

Thanks for raising some interesting points.

Thank you for for reading me.

I think both chess engines and humans compute and evaluate in precisely the same way. Humans are stronger where they have more knowledge, and engines where their knowledge is superior to human knowledge. A good example is king attack. Engines are on average, at least the top, much better than humans in attacking the enemy king with pieces. Why? Because they have more knowledge than humans about attacking; while humans just follow general patterns, engines calculate the attacking weight for each piece and each specific square. This is what I call knowledge, humans do not have this knowledge. Engines are very good on this because they started very early elaborating on king attack. This is an area they worked hard on.

Same goes true for mobility: engines are better what concerns overall piece mobility, as they calculate precise numbers there, something humans can not do. But humans are better in assessing the mobility for each particular piece. Why so? Simply because, as I understand, most engines do not give individual bonus for each piece, which I think is important. Because this lack of knowledge, you see engines like Komodo, Stockfish, Houdini in TCEC playing sometimes ugly games, losing because of extremely low mobility for a particular piece.

Same is true of pawn features like passers and isolated pawns: engines treat them relatively well, as the definitions and knowledge they have about those features are more or less precise and sufficient. However, when you come to backward pawns, here already engines are very bad and know almost nothing, I think in the first place because their definitions of backward pawns are, if not completely wrong, then very much imprecise.

An example: what you mention about backward pawns being weak, undefended pawns, is simply not true. Backward pawns are not about being weak or undefended, but about the inability to advance.

Below a position from TCEC, I am not certain this is exactly the same position, but an approximation. It is from round 26 of current stage 4, in the game Komodo-SF. I watched the game on the site, but when I now try to download games, I do not have access at the moment. So it is an approximation, but the pattern is the same.

[d]4r3/3q3k/pBn1r1pp/Pp1p1p2/2pP1P1P/2P1PP2/3Q1K2/4R2R w - - 0 1

Komodo thinks here white is 40cps worse, SF even sees 60cps black advantage. However, the position is equal. Both engines fail to take something into consideration. I guess this is the b5 black pawn, which is objectively a backward pawn, but I am almost certain none of the engines thinks so.

Do you agree b5 is a backward pawn due a nice penalty?
Well, b5 is a chain/connected pawn, so it is defended/not weak, but in spite of this it is backward because of its inability to advance. One white pawn, c3, stops 2 black pawns at the same time, it blocks c4 and attacks the stop square of the b5 pawn. As b5 is not opposed, it is simply a backward pawn. Checking for friendly pawns that could support b5 from behind on adjacent files: well, there is such a pawn, a6, but it is blocked by white a5, so realistically it can never support b5 in its advance to the b4 square.

So that b5 is simply backward, and the definition that a chain pawn could be a backward pawn if there is a friendly pawn behind on adjacent files, but that friendly pawn is blocked or also backward, is correct. Do you agree now that chain pawns could be backward and that b5 is backward?

The 2 strongest engines on the Earth, Komodo and SF, both do not see that. Why? Because they do not have the right definition of a backward pawn. So that knowledge always helps, both humans and engines. It is about correct knowledge, and not about incompatible paradigms.

[bob]

[d]6k1/3p4/1p6/p3p1p1/P3P2p/1P1P3P/6P1/6K1 w - - 0 1
I consider b3, d3 and g2 as backward for white. The engine will strive to remove this bacward pawn by supporting the push by pieces. For example white will play b4, if axb4 and a white piece takes b4 too, what we have now is a black backward pawn in b6, the white a4 pawn is not backward because it is already in the fourth rank. So black will prepare something if ever white will achieve the push, because if black will not take the white's b4 pawn push, white has the posibility now of either moving b4b5 grabbing space or b4xa5.

[d]6k1/3p4/1p6/4p1p1/P3P2p/3P3P/6P1/6K1 w - - 0 1
So why on Earth b6 should be backward and a4 not? Those pawns already get different psqt values for 4th and 3rd rank respectively, but apart from that, what is the difference between them? They both can not advance, it is 1 vs 1, some bonus for pawn control only for a4 is OK, but that would be 5-8cps, we are talking here of giving 20-30cps and more penalty for a real backward pawn. Is it possible that b6 gets penalised so much?

And also why d3 is backward? It can not advance, but black's d pawn also can not advance past the 6th rank, and it is an equal distribution of resources, 2 vs 2, neither side gets anything special from that situation. So that I think, although d3 could be penalised, this is wrong in practical terms.

To ask you one question, Ferdinand, how is Deuterium able to play so strong when it understands backward pawns so poorly?

Think like a computer. d6 is a "weakly backward pawn", As is d3. We call that a "wash". So no side has any advantage there. If one side has more of those weak pawns than the other, they then become a real disadvantage.

What is the purpose of considering terms that cancel each other and have no real contribution to evaluation?

They don't always cancel. If I have 3 of those backward pawns and you have two, my position is worse. Chess IS a "zero sum" game, don't forget.

[bob]

My take on this is that you are somewhat mixing terms.

A pawn that is defended by a pawn is not considered backward in any book I have (g6 in your first example as it is defended by h7). The fact that it can not safely advance is not so important when it is perfectly safe where it is. If you want to talk about it being immobile, that is a different issue.

the "unopposed" requirement stems from the rook, in the books I have read over the years. A pawn that is backward (can not be safely defended by a friendly pawn) is generally considered to be worse if it is (a) on a half-open file (no enemy pawn in front of it AND (b) the enemy has rooks to attack it.

IMHO, a pawn that is backward and unopposed is not any weaker or stronger than a backward pawn that is opposed, if there are no enemy rooks to contend with. Any backward pawn is weak, and inviting to enemy kings in the endgame. backward pawns on half-open files make inviting targets even in the middle game if you have rooks to attack on the file.

I've seen games decided by two types of backward pawns, repeatedly.

The first is the classic backward pawn on a half-open file, which is just as bad as an isolated pawn on a half-open file, when the enemy has rooks. Pile up on the pawn, tying up the opponent by making him defend, and either win the pawn outright, or switch to some other idea once he is tied down and not coordinated very well.

The second is a pawn that is backward, but the file is not half-open. If you have just one, you might manage to defend it with your king, depending on the position. If you have two, the enemy king can make you commit to saving one and then to eat the other pawn chain. Doesn't matter whether the file is half open or not.

[d]6k1/5p2/4p1p1/3p2P1/1p1P4/p7/P7/6K1 w - - 0 1

I tend to concur with Bob's definition. In the diagram above the e6 pawn is still defended, and thus not vulnerable to direct attack, so that's an argument against calling it backward.

However, since it cannot safely advance, there is an element of backwardness about it, and the square in front of it (e5) can serve as a good outpost for White pieces. These considerations should justify some sort of a penalty for such a pawn, but a lesser one that a fully backward [undefendable by a fellow pawn] pawn on an open file.

The Black b-pawn, even though backward, could be pushed in some cases creating a dangerous advanced passed pawn. It's not as clear whether a backward pawn that sits in the opponent's half of the board deserves any significant penalty.

CL

Crafty calls that a3 pawn a "hidden passed pawn". GM Dzhindi used to call this a protected passed pawn himself, because a3 is certainly passed here when you think about it and it is defended by b4. after b3 black ends up with a passed pawn for certain.

[bob]

My take on this is that you are somewhat mixing terms.

A pawn that is defended by a pawn is not considered backward in any book I have (g6 in your first example as it is defended by h7). The fact that it can not safely advance is not so important when it is perfectly safe where it is. If you want to talk about it being immobile, that is a different issue.

the "unopposed" requirement stems from the rook, in the books I have read over the years. A pawn that is backward (can not be safely defended by a friendly pawn) is generally considered to be worse if it is (a) on a half-open file (no enemy pawn in front of it AND (b) the enemy has rooks to attack it.

IMHO, a pawn that is backward and unopposed is not any weaker or stronger than a backward pawn that is opposed, if there are no enemy rooks to contend with. Any backward pawn is weak, and inviting to enemy kings in the endgame. backward pawns on half-open files make inviting targets even in the middle game if you have rooks to attack on the file.

I've seen games decided by two types of backward pawns, repeatedly.

The first is the classic backward pawn on a half-open file, which is just as bad as an isolated pawn on a half-open file, when the enemy has rooks. Pile up on the pawn, tying up the opponent by making him defend, and either win the pawn outright, or switch to some other idea once he is tied down and not coordinated very well.

The second is a pawn that is backward, but the file is not half-open. If you have just one, you might manage to defend it with your king, depending on the position. If you have two, the enemy king can make you commit to saving one and then to eat the other pawn chain. Doesn't matter whether the file is half open or not.

[d]6k1/5p2/4p1p1/3p2P1/1p1P4/p7/P7/6K1 w - - 0 1

I tend to concur with Bob's definition. In the diagram above the e6 pawn is still defended, and thus not vulnerable to direct attack, so that's an argument against calling it backward.

However, since it cannot safely advance, there is an element of backwardness about it, and the square in front of it (e5) can serve as a good outpost for White pieces. These considerations should justify some sort of a penalty for such a pawn, but a lesser one that a fully backward [undefendable by a fellow pawn] pawn on an open file.

The Black b-pawn, even though backward, could be pushed in some cases creating a dangerous advanced passed pawn. It's not as clear whether a backward pawn that sits in the opponent's half of the board deserves any significant penalty.

CL

Carl, whether e6 is defended or not is another matter, completely unrelated to backwardness, i.e. underdevelopedness, inability to advance. Defended and undefended pawns already get their separate bonus/penalty points. We are taling here about the inherent inability of the e6 pawn to advance. That matters very much, pawns are there to advance, help in the attack, promote, etc. If your pawns do not advance to attack the enemy and promote, then it will be your opponent's pawns which will do that to win the game. So that a passive state of inability to advance should always be penalised, regardless of whether such pawns are defended or not. The defended/undefended bonus/penalty is just a small fraction of the penalty for a backward pawn unable to advance without being lost, so that backwardness is the much more important concept.

[d]6k1/5p2/4p1p1/n2p2P1/1p1P4/p7/P7/2N3K1 w - - 0 1
You see above that in the usual case with pieces present, b4 does not represent any real danger of supporting promotion, but is still unable to advance without being lost. So that it is still backward. Of course, a backward pawn on the 4th rank deserves a very small penalty, much less than a backward on the 5th. That is what we have been talking about, the introduction of a unified concept that works across ranks, i.e. a good practical eval term.

How practical and good are current backward definitions that mix backward with weak, isolated, overextended, undefended?

Again, I think the definition I specified, with some 30cps penalty for the 7th rank, 15cps for the 6th rank, 8cps for the 5th rank, and 4cps for the 4th rank, is a good, working definition, unified and applicable across ranks. When you apply a unified feature across ranks, chances are you will get the most out of it.

I ask myself again, who is to blame when modern top engines fully misunderstand backward pawns? And if they badly misplay such pawns, is it possible that the definitions used for backward are good and working ones? Why should you mix backward with isolated, overextended (meaning vertically isolated) and undefended? You already have definitions and penalties for isolated and undefended. No, the only conclusion I draw from that is that currently a complete chaos reigns in the computer chess world as to what actually a backward pawn is. And that is the main reason why engines play so weak with backward pawns.

I wonder who will be the first author who will drastically change his approach and definitions of backward pawns to easily gain (after tuning of course) another 50 elo strength?

I don't think the term "backward" is related to mobility directly. A backward pawn is one that can not be defended, and which is weak where it stands offering the opponent a clear target. There is, to me, a difference between being immobile, and being backward. Immobile is certainly a problem, but not nearly the same as being backward AND immobile.

[Lyudmil Tsvetkov]

My take on this is that you are somewhat mixing terms.

A pawn that is defended by a pawn is not considered backward in any book I have (g6 in your first example as it is defended by h7). The fact that it can not safely advance is not so important when it is perfectly safe where it is. If you want to talk about it being immobile, that is a different issue.

the "unopposed" requirement stems from the rook, in the books I have read over the years. A pawn that is backward (can not be safely defended by a friendly pawn) is generally considered to be worse if it is (a) on a half-open file (no enemy pawn in front of it AND (b) the enemy has rooks to attack it.

IMHO, a pawn that is backward and unopposed is not any weaker or stronger than a backward pawn that is opposed, if there are no enemy rooks to contend with. Any backward pawn is weak, and inviting to enemy kings in the endgame. backward pawns on half-open files make inviting targets even in the middle game if you have rooks to attack on the file.

I've seen games decided by two types of backward pawns, repeatedly.

The first is the classic backward pawn on a half-open file, which is just as bad as an isolated pawn on a half-open file, when the enemy has rooks. Pile up on the pawn, tying up the opponent by making him defend, and either win the pawn outright, or switch to some other idea once he is tied down and not coordinated very well.

The second is a pawn that is backward, but the file is not half-open. If you have just one, you might manage to defend it with your king, depending on the position. If you have two, the enemy king can make you commit to saving one and then to eat the other pawn chain. Doesn't matter whether the file is half open or not.

[d]6k1/5p2/4p1p1/3p2P1/1p1P4/p7/P7/6K1 w - - 0 1

I tend to concur with Bob's definition. In the diagram above the e6 pawn is still defended, and thus not vulnerable to direct attack, so that's an argument against calling it backward.

However, since it cannot safely advance, there is an element of backwardness about it, and the square in front of it (e5) can serve as a good outpost for White pieces. These considerations should justify some sort of a penalty for such a pawn, but a lesser one that a fully backward [undefendable by a fellow pawn] pawn on an open file.

The Black b-pawn, even though backward, could be pushed in some cases creating a dangerous advanced passed pawn. It's not as clear whether a backward pawn that sits in the opponent's half of the board deserves any significant penalty.

CL

Carl, whether e6 is defended or not is another matter, completely unrelated to backwardness, i.e. underdevelopedness, inability to advance. Defended and undefended pawns already get their separate bonus/penalty points. We are taling here about the inherent inability of the e6 pawn to advance. That matters very much, pawns are there to advance, help in the attack, promote, etc. If your pawns do not advance to attack the enemy and promote, then it will be your opponent's pawns which will do that to win the game. So that a passive state of inability to advance should always be penalised, regardless of whether such pawns are defended or not. The defended/undefended bonus/penalty is just a small fraction of the penalty for a backward pawn unable to advance without being lost, so that backwardness is the much more important concept.

[d]6k1/5p2/4p1p1/n2p2P1/1p1P4/p7/P7/2N3K1 w - - 0 1
You see above that in the usual case with pieces present, b4 does not represent any real danger of supporting promotion, but is still unable to advance without being lost. So that it is still backward. Of course, a backward pawn on the 4th rank deserves a very small penalty, much less than a backward on the 5th. That is what we have been talking about, the introduction of a unified concept that works across ranks, i.e. a good practical eval term.

How practical and good are current backward definitions that mix backward with weak, isolated, overextended, undefended?

Again, I think the definition I specified, with some 30cps penalty for the 7th rank, 15cps for the 6th rank, 8cps for the 5th rank, and 4cps for the 4th rank, is a good, working definition, unified and applicable across ranks. When you apply a unified feature across ranks, chances are you will get the most out of it.

I ask myself again, who is to blame when modern top engines fully misunderstand backward pawns? And if they badly misplay such pawns, is it possible that the definitions used for backward are good and working ones? Why should you mix backward with isolated, overextended (meaning vertically isolated) and undefended? You already have definitions and penalties for isolated and undefended. No, the only conclusion I draw from that is that currently a complete chaos reigns in the computer chess world as to what actually a backward pawn is. And that is the main reason why engines play so weak with backward pawns.

I wonder who will be the first author who will drastically change his approach and definitions of backward pawns to easily gain (after tuning of course) another 50 elo strength?

I don't think the term "backward" is related to mobility directly. A backward pawn is one that can not be defended, and which is weak where it stands offering the opponent a clear target. There is, to me, a difference between being immobile, and being backward. Immobile is certainly a problem, but not nearly the same as being backward AND immobile.

See this: http://talkchess.com/forum/viewtopic.ph ... 44&t=52300

[d]4r3/3q3k/pBn1r1pp/Pp1p1p2/2pP1P1P/2P1PP2/3Q1K2/4R2R w - - 0 1
Is not b5 a backward pawn?