Double pawn push

[Lyudmil Tsvetkov]

In an attempt to completely scandalise all fervent supporters of redundancy theories, here one more suggestion concerning the sovereign right of any pawn on the 2nd square of the board to make a double move.

Give some 5 cps bonus standard values, both for the middlegame and the endgame, to any pawn on the 2nd rank, having the squares on the 3rd and 4th rank in front of it unoccupied, either by own pawns and pieces, or by enemy pawns and pieces.

That will mean, that such a pawn is able to execute currently a double pawn push, which is always an advantage.

Why so?
The answer is easy and unambiguous:

In the middlegame,

- such pawns on the 2nd rank can easily occupy central and more advanced squares, which is a benefit
- they can start either king side, or queen side attack, very quickly and forcefully, without being annoyed by the possible existence of own or enemy pawns and pieces in front of them; that is very useful for storming pawns, but also for other pawns

In the endgame,

- such pawns, especially when passers, can advance very quickly from the home rank, and beneficially influence the game; a passer on the 2nd rank when able to make a double move is twice quicker than a passer on the 2nd rank, which is blocked by one object on the 3rd or 4th ranks, and even speedier than a passer on the 2nd rank blocked by 2 own or enemy objects; and also, equally quick to a paser on the 3rd rank that is free

- sometimes such pawns on the 2nd rank can provide a valuable reserve tempo in case of zugzwang positions, which could ensure the win

Fervent redundancy theories supporters will be appalled by myself first suggesting some strange en passant term, involving a double push, and then even more insolently coming up with a very redundant idea again involving a double push.

However, it happens so, that both terms are absolutely non-redundant.

Very simply, a pawn that could be captured en passant may not be free, i.e., on the 4th rank in front of this pawn there might be an own or enemy piece, which distinguishes this from a pawn on the 2nd rank able to make a double move.

Secondly, such a pawn may be entirely independent of any en passant enemy captures, i.e. there might not be enemy pawns on the 4th rank to the left or to the right.

Moreover, the bonus is given for the ability of the pawn to advance quickly and have reserve tempo, and also for the opportunity of such a pawn to choose between 2 possible moves, which implies an increased mobility and increased tactical resources, with both assets being particularly important.

Many authors might not consider pawn mobility, but mobile pawns definitely play a very important role in chess, just like immobile pawns.

For the time being I will leave at that, not posting any diagrams, as the idea is sufficiently simple and straightforward.

Any comments very much appreciated, especially any heavy reprimands by disgusted anti-redundancy supporters.

[Lyudmil Tsvetkov]

Usefulness of double pawn pushes in the endgame:

[d]6k1/6pp/6bb/8/8/8/PP3BB1/6K1 w - - 0 1
While material is completely even, available double moves to the a2 and b2 white passers, and unavailable such for black h7 and g7 passers give white real winning chances.
Search might see, it, sometimes not, but static eval might also help.

[d]8/3k1p2/p5p1/2Pp2P1/5P2/2K5/P7/8 w - - 0 1
This is an example for a spare/reserve tempo I took from wikipedia. This is from the game Nunn-Bischoff.

As your engine might see, white to move wins precisely because of the availability of a reserve tempo, assured by the available double push of the a2 pawn. For comparison, the black a6 pawn lacks such a reserve tempo.
So a double push is especially useful in zugzwang positions.

Well, the black f7 pawn also has a double push and as if a reserve tempo, but here g5 has en passant rights over f7, so you should tune both, nothing to do.

[Lyudmil Tsvetkov]

A middlegame position:

[d]r1bq1rk1/1pp2pp1/p1np1n1p/2b1p3/4P3/1BNP1N1P/PPP2PP1/R1BQ1RK1 b - - 0 1
a2,b2,c2,f2 and g2 pawns are still on their home, 2nd rank, with possibly available double pushes.
However, of those 5 white pawns on the 2nd rank, only 2 may currently make a double move, a2 and g2, with the other 3 being blocked by own pieces, b2 by Bb3, c2 by Nc3, and f2 by Nf3.

Similarly, black has still 4 pawns on their home 2nd ranks, b7,c7, f7 and g7, however, of those 4 pawns only 2 may currently make a double move, b7 and g7, with the remaining 2 being blocked by own pieces (but enemy pawns and pieces can also block them), f7 by Nf6, and c7 by both Nc6 and Bc5.

So equality in terms of double pawns pushes here, 2 each side, but in many positions one side will have more or much more pawns able to make a double move, and that will imply additional mobility and tactical opportunities.

So the requirement that the 3rd and 4th rank in front of the pawn on the 2nd rank are empty is an essential one.

[jorose]

I like the idea and depending on how you implement it I don't see any problem of redundancy. It's something I would definitely test before accepting it in the end but I feel this is just the stronger term compared to your backwards pawn term. Perhaps it would make sense to calculate a bonus for a pawn on the second rank with no piece in front of it on the 3rd and 4th ranks as well as no enemy pawn that could capture it en passant. As such you might be describing the additional pawn mobility correctly as if you move your pawn one square further or two squares further and the opponent captures (en-passant if two regular if one square pushed) it would be identical, hence not necessarily extra mobility involved. Implementing it like this you would have a fairly independant heuristic describing both your en-passant bonus as well as your double pawn push idea. The cost would also be slightly lower then evaluating your ideas separately.

[Lyudmil Tsvetkov]

I like the idea and depending on how you implement it I don't see any problem of redundancy. It's something I would definitely test before accepting it in the end but I feel this is just the stronger term compared to your backwards pawn term. Perhaps it would make sense to calculate a bonus for a pawn on the second rank with no piece in front of it on the 3rd and 4th ranks as well as no enemy pawn that could capture it en passant. As such you might be describing the additional pawn mobility correctly as if you move your pawn one square further or two squares further and the opponent captures (en-passant if two regular if one square pushed) it would be identical, hence not necessarily extra mobility involved. Implementing it like this you would have a fairly independant heuristic describing both your en-passant bonus as well as your double pawn push idea. The cost would also be slightly lower then evaluating your ideas separately.

Well, whatever programmers say.

The problem is I do not see how such terms could be tuned with a visible elo gain. You have to tune, and tune, and tune...
I would not like to be in the place of programmers.