Mistake about doubled pawn theory

I was in a hurry to learn the ways of doubled paws, so I skipped an exercise in month 4 in favor of reading the pdf about doubled pawns for the next month. Something I didn't trust...



So that's what I did. I removed the right hand pawns and moved the kings to the left but still on their initial rank. And behold, the doubled-pawn-side won by creating a passed pawn! So a pawn majority which contains a doubled pawn still can force a promoted pawn. I used the self-play feature of "Arena" GUI to test this.

So what is the right analysis for the given original position? The point is... lack of king mobility! For white, in this case, the king doesn't need to assist because 4 pawns-in-a-row wins against 3-pawns-in-a-row without help.
But the doubled pawn-side needs help from its king to create a passer. Right now the central white pawn holds back the black king! And so white wins. Read on for one more detail: the role of the white king.

Now remove whites e-pawn, and blacks f-pawn. So there's no central pawn holding back the black king. But now - since whites pawns don't need the help of their king - the white king can obstruct the black king. So again white wins because black will fall many tempo's short if white is less efficient, or black can't create a passer at all!

And that, my friends, is the correct answer. [Lack of] king mobility because doubled pawns need help from their own king to create a passer.
So if you encounter such an endgame and you have the advantage,
1. use your king to obstruct the other king. Your own pawn majority can storm forward without help.
2. if the opponent decides to attack your pawns, you have your king to put behind your pawns (don't obstruct them) and then slowly push forward again. The other half can never promote without help due to doubled pawns, remember? So don't get nervous.

The instruction is to try to create a passer with only the queenside pawns. I made a couple of tries and convinced myself that a passer cannot be created.

Adding kings to the puzzle is a distraction, the objective is to learn how to advance the white pawns in a manner such that black is unable to create a passer.

The pawn structure of the initial position with seven pawns on the board is very similar to an endgame composition by Max Euwe; the difference being that black's a-pawn is at a7 and white's f-pawn is at f2 and kings are at e1 and e8. There is a very detailed analysis in Muller & Lamprecht's Secrets of Pawn Endings much too detailed to place here. Suffice to say white's winning technique is to immobilize the queenside and create the passer on the kingside; I think that is the same as what you are saying from your analysis, but to rephrase your comment black is losing because he is unable to create any counter play on the queenside.

Pawn endings are less about mobility than opposition, key squares, and tempo. To do a quick deep dive into pawn endings I highly recommend The Final Countdown by Hajenius & Van Riemsdijk, they take what is a surprisingly complex subject and really break it down into its essential elements.

The instruction is to try to create a passer with only the queenside pawns. I made a couple of tries and convinced myself that a passer cannot be created.

I think you are right, i.e. that's what the author meant, reading it again and again. And indeed, I too think no free pawn can be created in a winning fashion.

black is losing because he is unable to create any counter play on the queenside.

Maybe that was the case with Euwe's example, but here I specifically defined "counter play", which is that the black king must assist it's majority, but white doesn't.

Pawn endings are less about mobility than opposition, key squares, and tempo.

In general, sure, but these typical pawn-endgame elements are not relevant here. As long as the white king defends its center pawn, black cannot approach and if there is no center pawn, then the king simply must make sure it can reach a blocking square timely whilst defending its pawns if needed against the black king.

I think we are saying the same thing with a slightly different vocabulary. Following is the Euwe composition, it is worth a quick scroll through since the technique is representative of positions of this nature.

[Event "Deutsche Schachzeitung (te) @2"]
[Site "?"]
[Date "1940.??.??"]
[Round "?"]
[White "Euwe=M"]
[Black "(+0000.77e1e8)"]
[Result "1-0"]
[EventDate "1940.??.??"]
[FEN "4k3/1pp2ppp/p1p5/8/4P3/8/PPP2PPP/4K3 w - - 0 1"]
[SetUp "1"]

{ @2: Miles=A Encyclopedia of Chess Endings Vol. I#0798 1982 @2: Van_der_Heijden=H HHdbIII#49064 13-11-2004 } 1.Ke2 Ke7 2.Ke3 Ke6 3.f4 c5 4.c4 $1 c6 5.a4 b5 6.b3 $1 f6 7.a5 b4 8.g4 ( 8.Kf3 $1 { } 8...Kf7 9.e5 fxe5 10.fxe5 ) ( 8.f5+ $1 { } 8...Ke5 9.g4 g6 ( 9...h6 10.h3 ) 10.fxg6 hxg6 11.h4 ) 8...g5 9.e5 $1 ( 9.Kf3 $1 { } 9...h6 10.Kg3 Ke7 11.h4 Kf7 12.hxg5 hxg5 13.Kf3 gxf4 14.Kxf4 Ke6 15.e5 $1 ) 9...gxf4+ 10.Kxf4 fxe5+ 11.Ke4 h6 12.h4 Kf6 13.g5+ hxg5 14.hxg5+ Kxg5 15.Kxe5 Kg4 16.Kd6 Kf4 17.Kxc6 Ke4 18.Kxc5 Kd3 19.Kxb4 Kd4 20.Ka3 Kc5 21.Ka4 Kd4 22.Kb4 Ke5 23.Kc5 1-0

I set up and played out the position you described without the kingside pawns and the kings moved on file towards the queenside. The position is a dead draw - double checked with Komodo Dragon which I have configured with the endgame databases. Creation of the passed pawn is insufficient because the white king controls the key squares.

You're right! It's a dead draw... if white begins.
If black begins, black wins. I suppose yesterday late I accidentally let black begin first and assumed it's always a win for black. I can't even remember. It's a good thing you checked as well.

Whoever dominates the center d-square(s) first (standing on e4/e5 counts too) and doesn't have to give it up, blocks the other king from helping. Very instructive on the roles of the kings in these situations.

Now I will analyse your position. I'm excited! There's good things to learn.

Edit: oh, it's about the same. White chooses to give black play on the king side whist cleaning up on the other side. Stockfish chooses to deny any counterplay by closing the center and then rolling the kingside pawns up the board. Maybe I like the latter technique better as it requires far less calculation. I'm too old to think . Can you also copy Euwe's comments or something?

Valmont wrote:You're right! It's a dead draw... if white begins.
If black begins, black wins. .

I'll take your word for it, but now I am suspicious. If black can only create a single pawn, then if white can prevent the black king from occupying one of the key squares then it is a draw. More than a single pawn, then who knows...

https://en.wikipedia.org/wiki/Key_square

I'm not aware of any comments by Euwe. The technique demonstrated is fairly common - white advances the kingside pawns as far as possible creating an outside passed pawn, then rushes to the queenside for lunch and the party

Here's an example of Stockfish self-play. 11 sec per move (forced, not average).

[Event "Computer chess game"]
[Site "TORPEDO"]
[Date "2021.03.01"]
[Round "?"]
[White "Stockfish_21022022_x64_avx2"]
[Black "Stockfish_21022022_x64_avx2"]
[Result "0-1"]
[BlackElo "3600"]
[Time "20"]
[WhiteElo "3600"]
[TimeControl "0+11"]
[SetUp "1"]
[FEN "3k4/ppp5/2p5/8/8/8/PPP5/3K4 b - - 0 1"]
[Termination "normal"]
[WhiteType "program"]
[BlackType "program"]

1. ... Kd7 2. Ke2 Kd6 3. Kf2 Ke5 4. Ke3 c5 5. Kd3 Kd5 6. c3 a5 7. Ke3 a4 8.
Kd3 c4+ 9. Ke3 a3 10. bxa3 Kc5 11. a4 Kb6 12. Kd2 Ka5 13. Kc2 Kxa4 14. Kb2
Kb5 15. Kc1 Kc5 16. Kd2 Kc6 17. Ke2 Kb5 18. Ke3 Ka4 19. Kd4 b5 20. Kc5 c6
21. Kd4 Ka3 22. Ke5 Kxa2 23. Ke4 Kb3 24. Kd4 Kb2 25. Ke5 Kxc3 26. Kf4 0-1

Valmont wrote:Here's an example of Stockfish self-play. 11 sec per move (forced, not average).

OK, now I understand your comments. I assumed a different initial position for the kings plus white to move and we all know what the acronym ASSUME is...

When white is to move then it's a draw, so you got that one right, albeit accidentally. Lucky for me as I learned an extra thing, as mentioned earlier. I through for a day black is always winning, but now I have to consider whether my king can make it to the a crucial square to cut off the king.