Complicated position, win, lose, or draw?

I would have taken the King.

 Huh?

 I am a player of elo 2265. I have no "feeling" in this position.

My first impulse was: two pawns are two pawns, and white has no mating attack, nor can he open the e- or d - file. His immediate threat is to play c4-c5, and then to grab b7. But is this so horrible? b7 is just a pawn.

Then I started thinking about a good action for black. He can try to survive that loss on b7, and play e6, Be7 (Bd6) + 0-0, also directly Bg7 and 0-0 is possible.

This would mean, I have to calculate long variants ...  With no guarantee for the final verdict....

Analysis over some time (an hour?) could give me a hint, but the probability of errors in those double-edged position is high, at least for me.

Did you ask a computer? Fritz 9 or Fritz 10, Rybka?

Computers are "materialistic". In order to avoid an evaluation like "+1 pawn " or "+2 pawns" equivalent for black , let them think with depth of up to 15 or even 20 half moves.  And look at the main variants they spit out.

I think this is exactly the type of position they handle better than most humans! 

Mandelshtam 

grab b7??? two pawns are two pawns? you aren't looking at the final position are you? there is nothing on b7... neither black nor white has two pawns you can relate to (and the white b2 pawn is dropping off next). Open the d file? it's already open...

I have a feeling you're not looking at the final position... 

As for computers (in the final position, I've gone over this with more than one engine. No clear opinions. They differ from engine to engine, no one seems to be getting a really big advantage in the near future by force, but there is also no clear cut draw, and the position can undergo many changes. I have reached no conclusions yet. Still looking at it. 

 I apologize, when I looked over it first, IT (the move button) stopped  after move 20 or so. After checking your letter, I looked over IT again and pushing the button gave me the whole game, and the final position. I don't have a clear opinion either.

You seem to prefer white. I cannot agree with this. I believe a draw would be a justified result.

Let me explain my thoughts (without any analysis). 

Two rooks (=10 pawns) are often better and win against a queen (=9 pawns).  

This trivial materialistic judgement is true, when the player with queen has no passed pawn  (or pawn chain). Also the two kings should  be both relatively save, and there shouldn't be other pieces on the board. In all other cases "it depends".  So these criteria are not fulfilled in our position. However, more pieces on the board  often turn the advantage  to the side with the two rooks, since the Queen has then less opportunities to invade in the enemies camp. 

In the final position,  white's  pawn b2 cannot be saved directly.  White  can attack e7, with Qd7. If black decides to take on b2, then white takes on e7, and that  position is again unclear. (I believe , it is objectievely a draw, the pawn c4 is not a danger, and on the other hand, white cannot successfully attack the black king. Black has always counterplay, with two bishops, and the point f2 remains a weakness...)

If , instead, black decides to first save the pawn e7, with e6, then White can play Qa3. After Ra8 black puts the pawns on the queenside under pressure. Also a different approach: 1. Qd7 e6 , 2. Qd2 Rfd8 is unclear. I would prefer black in both approaches: he can freeze the white pawns with Bg7-f8, if necessary, and then try to win b2 (which still is not easy to defend). If this happens, and white does not manage to obtain an attack on the kingside in the meantime, black would have chances to win (he then would be 1 pawn ahead, and f2 is always a weakness). I am sure, your programs showed you already lots of those scenarios.  Anyway, the fact that black cannot obtain a passed pawn makes a draw most probable...

Mandelshtam 

Actually engines seem to prefer white a bit, but give the position as only a small advantage (during the game, I was inclined towards a draw, actually, because as I stated, I could find no way to make use of any "defects" that black may have, so I thought white has nothing special).

Since nothing special was found for white, I started digging in more deeply myself. The best I could come up with is: 26.Qc5 Rxb2 27.Qxe7 (I think this is practically the only try for white) 27...Ra8 28.Nd4 Bf8 29.Qc7 Raa2 30.Bf3 Rxf2 31.Bd5 Rfd2

32.Nc6 Rd1+ 33.Kh2 Re2 (stop the blockade of the pawn). I don't see how white or black can move anything in a constructive way here... I also don't think any other moves for white are a serious try for an advantage. This looks like a draw.

Blk need not take the b2 pawn immediately. White can't save it. If he takes the time to try, f5 frees blk's bishop to attack b2.(Qc5 e6, Qa3 f5) If white goes for the pawn on e7, advance it. The b2 pawn will still be there.Once blk takes the pawn white has to be concerned about his bishop being pinned to his king if blk gets his rook to the 1st rank. The pts. may be even, but blk can bring more pressure. He has more pieces. White can't trade his queen and it is basically locked to the c-pawn. Blk can still use his rook on f8 to apply pressure to the c,d, and e files depending on what white decides. Using the e and f pawns to dislodge the knight frees up the bishops, and the b2 pawn can be taken when blk decides to put his rook on the 1st rank. Blk only takes the b2 pawn right away if white moves his bishop to d1 or his knight to d2, planning b3. White should be happy he got the draw.

 I have to disagree. First of all, the position has undergone extensive analysis by me (~2000), numerous ~2100 players, a few masters, and numerous engines. All agree that white is never in any real danger (although some of them thought he may be at first, they changed their minds after analyzing for a while). The analysis supports it.

The idea of not taking the b2 pawn immediately was suggested, but it doesn't work that well for various reasons, first of all,

if black ignores the b2 pawn by making a move which is not related to the e7 pawn (after 26.Qc5), white always has the option, if he wants to, to take on e7 (not that it's necessarily best... but it's just an option), and never even try to do anything about b2, and black will have the same problems like the ones in the analysis I have given in making any progress. He also has to keep an eye on the c pawn (and white will just ignore the b2 pawn too). 

If black defends e7, the b2 pawn will be a part of the game... Qa3 (against some moves, b4 would be the best way... It depends...) will make it very hard for black. (Ra8 Qe3 will maintain the pawns for a while. Analysis (again - of both strong human players and later supported by engine analysis) shows this would be close to winning for white).

But if e7 drops, black probably has no way to ever make progress. He does not have enough pieces to make a breakthrough in white's weakness free king position (his pawn structure even prevents him from exploiting the slight h3 weakness).

The thing is, as one strong player showed me, that if e7 drops, then EVEN if white loses the c pawn as well, the position is STILL not dangerous for him at all. Black's king position prevents him from bringing all his pieces to use. This was (to my great surprise) again supported by analysis. The other problem is, black seems to have no way to win the c pawn at all. And without it, he's a bit tied down.

It's a surprising position.  But you know what? suggest something for black. I'd be happy to play many games from this position as white and see if he can be put in danger (if my moves would be correct, that is.....). Maybe you have a specific idea in mind.

If black's pawn structure and king position were not that fractured, indeed I think I'd prefer black, but that is not the case. (Btw, engines were used to depth ~25 to support human analysis, and both in the initial position and at the end of every suggested line white is +0.something or 0(equal)). Which just means there is nothing tactical or REALLY position changing in there... 

But us humans have not been able to find any plans for doing so as well. For either side. I'd be happy to hear concrete suggestions from you.

It seems we change arguments like politicians: nothing really convincing (general remarks about good and bad pieces don't help in such a double edged position; every "long" move with the Queen or with the two rooks can change the relative value of those pieces; human cannot evaluate this correctly, unless they analyse long variants!!!) . What convinces me (only) is a hard analysis of the two fundamental decisions

1. Qc5 e6 , 2. Qa3 Ra8 (Rb8) , or

1. Qc5 R:b2 , 2. Q:e7.

It seems that you, grolich, have done it already carefully: the computer finds in some variants a slight edge for white (never enough to win), and  no edge at all in others.  

mandelshtam 

So if I read this correctly. Unless you see 1.Qc5 e5 2.b4 Rb7 3.b5 Rf8-b8 4.g3 d4  5.Nh4 Rb6 6.Qd5 Re3 7.f3 e3 8.Nxb h7xN so you can plug it into an analyzer to pick it apart, the opinion is not wanted. So be it. What did the analyzer say of the first 25 moves? Did it tell you who was playing tentatively? cautiously? Did it analyze the personalities of the players? Because that's who was playing. Did it tell you if either player thought if the b2 pawn or the e7 pawn were important considerations? Their play did. Black thought that his fractured pawn structure required both bishops. did it? White expressed concern over the b2 pawn. What did black think about the pawn? My current rating is zero. I quit playing tournaments 35 years ago.

 The shot answer to your question is: no.

That's why me and the other players I mentioned analyzed this at length before ever using engines. The engines were just used to test our conclusions or find serious errors. The general feeling from the analysis was that both sides need to make really serious errors to actually be afraid.

The long answers:

1) Any opinion is wanted, since I'm looking for new ideas that have been missed.

2) As I've hinted, the engine just strengthened the already existing opinion that as long as the pawn remains unable to advance, the position will not be dangerous, for either side.

 3) Both me and my opponent were present in the analysis and were both extremely active participants, so I can say in a less objective manner, that if the game were to continue, I have no doubt that a draw would have occured. I'm just interested in learning from this position, what's the best way to relate to that.

4) as for b2 and e7:

a) The analysis already brought the conclusion that not taking on b2 makes the position very dangerous for black, and not for white. Also, thats without the e7 pawn, black has no winning chances whatsoever. In addition that the existence of the 2 pawns is more dangerous for black than what he might be getting from keeping the e pawn. The variation you gave was already torn apart in the post game analysis.

b) Sure, engines, if used correctly, tell you if the b2 pawn is important, same as e7. Just plugging the position into an engine and pressing analyze is the wrong way to do it though. You have to ask it the right questions. Actualy try your other ideas, as well as the main line it suggests, compare both the positions and the evaluations of taking on b2 and not taking. Use the output of the analyzer to form an opinion. In this case, lets just say, not only did the engines support the post game analysis of those options, 

it also shows a very large shift in the evaluation of the position in white's favour the second you try to ignore the b2 pawn as black. Checking the lines it offers, they're quite simple.

You have to really ask the engines the important questions in order for them to be of any real help.

5) During the game, black was going to take the b pawn immediately. 

As for the line you've given, I'll open a new post

(I assume d4=e4 and Re3=Re6:) 

Rb7 is a clear mistake. No engine is required here. After the obvious b5 Rfb8 Black is going into a completely passive formation against two connected passed pawns.

In this variaiton, only black can be in danger. The fact that engine's evaluation say it looks like suicide shouldn't even matter.  It's just a position you have to avoid.

We came up with 4.Qd6(I don't like g3. Looks like a waste of time. White has nothing to fear here). The idea is to meet Bf8 with Qxf6 and Rb6 with Qd2, leaving the rook in an uncomfortable position for later. 

4...Bf8 looks like black's best defensive formation 5.Qe3 and I think R6b7 is forced. Then Nd2 and black's passive formation cannot hold for long. He may be able to switch to a more active one on the d file still... But it looks much bleaker than just staying active without giving white the connected passers.

 By the way, I've gone over the line you gave, I think it contains a few clear mistakes. I'm away from my computer for this weekend, but for now, what I can see immediately is that e4 was a huge error. by move 7 of your line, white is already probably winning:)

instead of 7.f3?! go for direct 7.c5 passed pawns need to be pushed. 7...Re5 8.Qd6 exploits the unprotected b8 rook. 8...Rbe8 is almost forced (or white gets more forcing moves), but this already looks lost for black. 9.c6 

If black tries ...e3 then just f4!

Interesting line. Instead of ...Rb7, the more natural (for blockading) Rfc8 may give black better hopes of holding on.

Still, you were claiming in the original position white should be lucky to have gotten a draw. The line you gave is only dangerous for black, without white having done anything special. It seems clear to me that if white can have two connected passers black would be in serious danger.

 Even at the end of the line you've given (With the f3 move I don't like for white), 9.f4 may still be the end of black. He his relegated to passive defense against connected passers. This is just as close to suicide as one can get without being completely sure of dying...:)

Can you think of other ideas that may put white in some trouble? that's what I've been looking for. So far, haven't found any.