#### Only Chess Forum

1. 07 Jul '15 02:19
Hello,

In one of my games on chess.com I came across the following position
. When running it through on my computer, my computer said that 20 ...d4 was unequivocally the best move. I went through a few lines but there were too many reasonable looking variations to me to see why d4 was best there. I was wondering if a human being could help me understand that move, and explain how one might find that move in an OTB game.

Thanks.
2.  BigDoggProblem
07 Jul '15 03:07
Originally posted by amolv06
Hello,

In one of my games on chess.com I came across the following position [fen]1k5r/ppp3p1/3r2Q1/3pq3/1P6/P1b1P2P/R1P1KP2/2B4R w - - 0 21[/fen]. When running it through on my computer, my computer said that 20 ...d4 was unequivocally the best move. I went through a few lines but there were too many reasonable looking variations to me to see why d4 was b ...[text shortened]... elp me understand that move, and explain how one might find that move in an OTB game.

Thanks.
White's King is in the center, but hiding behind a pawn shield. White's Rook on a2 is badly out of play as well. Thus, there is a strong motivation for Black to act quickly to open as many lines around the white K as possible, not giving white time to coordinate his pieces defensively.
3.  JerryH
Hyperbole Happy
07 Jul '15 03:41
Unless one can see forced mate and even then when OTB, I'd just Rxg6. Too many chances to misplay, letting the queen go. For the computer a timely Qb5+ is probably part of the calculation but you're right too many lines.
4. 07 Jul '15 03:44
That certainly makes sense, and while d4 seems natural I don't see how it leads to a concrete advantage over other moves. Fritz puts d4 at -4.72 and Rhf8 at -1.47 so I assume there is a piece to be gained or something from d4. However, that's not clear to me. I ended up playing Rd6 with the idea of eventually doubling up the rooks on the c-file. How the ideas behind d4 end up actually playing out is just not clear to me, and I don't think I'd be able to make the distinction of d4 as distinctly better than other moves OTB, even though it is something I would probably consider.
5. 07 Jul '15 03:45
I'm sorry, I just noticed that my original image is wrong. This is what should be posted:

6.  moonbus
Uber-Nerd
07 Jul '15 07:00
Originally posted by amolv06
I'm sorry, I just noticed that my original image is wrong. This is what should be posted:

[fen]1k1r3r/ppp3p1/6Q1/3pq3/1P6/P1b1P2P/R1P1KP2/2B4R b - - 0 20[/fen]
20. ... d4 looks like a typical machine-move to me. There may be a long variation trailing behind it leading to the win of a pawn (typical machine-play). I would not have played ...d4 as it weakens the white squares in the center for Black. 20... Rd6 is certainly not bad, with the idea of swinging the rook over to double on the h-file and pound White's isolated h-pawn. This threat pretty much forces White to defend the h-pawn with Q & R, giving Black a free hand to penetrate the position with his Q. White's position has so many holes in it that Black has an embarrassment of riches: one could also put the Black Rs on the d and e files and then play ... d4 to crack open the White K's bastion.
7. 07 Jul '15 07:24 / 1 edit
The high computer score might be caused by it placing value on creating the way for a more powerful diagonal check line with ....Qb5+. Also the pawn to d4 defends c3 while attacking e3. The weaker white rook on a2 will also fall under more pressure with Qb5. While there are many other possible patterns, this one might indeed have he most benefits overall. The black queen is currently not being maximized, somewhat blocked by e3 and its very own d5. Of course to be sure, all the alternative lines will have to be followed which is what the computer might have done but I think in general it's a pretty solid move to upgrade the maneuverability of the Queen in the middle of a major attack.
8.  Ragwort
Ex Duris Gloria
07 Jul '15 07:27
Originally posted by amolv06
I'm sorry, I just noticed that my original image is wrong. This is what should be posted:

[fen]1k1r3r/ppp3p1/6Q1/3q4/1P1p4/P1b1P2P/R1P1KP2/2B4R b - - 0 20[/fen]
I think the biggest threat of 1...d4 is Qd5 forking the two white rooks.

The rook on a2 looks particularly vulnerable with no escape squares. Moves like Bb2 intending to answer Qd5 with Rha1 fail to Bxb2 Rxb2 d3 check picking up the rook.

Because these are "simple" two move tactics your engine probably isn't showing them to you because it is trying to defend these threats at a much more complicated level teaching you very little about "what's going on in the position". Even if you push the engine along the line I've put in above it will promote more complex lines because it can "see" this easily. There will be a number of these two and three movers to tease out before you will be able to understand why the computer favours the lines it does.
9.  moonbus
Uber-Nerd
07 Jul '15 10:50 / 2 edits
Well seen, Ragwort. Thanks.

So, White must somehow stop the rook-fork on d5:

This looks perilous for White though.
10.  Ragwort
Ex Duris Gloria
07 Jul '15 18:37 / 2 edits
Originally posted by moonbus
Well seen, Ragwort. Thanks ...
This looks perilous for White though.
Good work moonbus.

I think White is now losing the a2 rook to the move suggested above Qb5+

for example:

and that is no doubt the basis for the 4.7 score of the computer for the d4 move. Once again computers will promote a more "accurate" line than the one I have given even though it will mask some of the basic threats going on in this position.

As for how to spot these moves you just have to ask yourself if you would see it as a candidate move once you realised that white had two hanging pieces which you might then have tried to exploit. Some people advocate working through lots of tactical problems to hone their vision in the hope that when something is in a position it will be subliminally recognised. Others advocate carefully unpicking positions as we have begun to do so here in the hope that the lessons learned stick more permanently. Although both have their place I favour the latter as I believe learning to look for things you don't recognise and learning to ask questions builds your problem solving skills at the board for when there is no-one to ask.
11. 07 Jul '15 19:47
Thank you for all your replies. I now understand the threat. At this point, I think some of these variations are too detailed to analyze OTB. I just don't have that kind of vision, but now that I've seen it played out I do see how to implement the plan. Thanks again.
12.  moonbus
Uber-Nerd
08 Jul '15 18:15
Originally posted by amolv06
Thank you for all your replies. I now understand the threat. At this point, I think some of these variations are too detailed to analyze OTB. I just don't have that kind of vision, but now that I've seen it played out I do see how to implement the plan. Thanks again.
Let me recommend a book to you. It explains how to analyze a position into its strategic elements; once you can do that, plans just jump out at you. Moreover, tactical shots (like the one Ragwort discovered above with the Black Q penetrating through White's holes) become obvious.

http://www.amazon.co.uk/Chess-Strategy-Club-Players-Positional-ebook/dp/B00IXOLLH6/ref=sr_1_fkmr0_1?ie=UTF8&qid=1436379130&sr=8-1-fkmr0&keywords=hermann+grooten
13.  Paul Leggett
Chess Librarian
10 Jul '15 23:05 / 1 edit
Originally posted by moonbus
20. ... d4 looks like a typical machine-move to me. There may be a long variation trailing behind it leading to the win of a pawn (typical machine-play). I would not have played ...d4 as it weakens the white squares in the center for Black. 20... Rd6 is certainly not bad, with the idea of swinging the rook over to double on the h-file and pound White's isola ...[text shortened]... put the Black Rs on the d and e files and then play ... d4 to crack open the White K's bastion.
Wow, ...d4 was one of the first moves that came to mind for me, simply because of the reasons BiggDoggProblem pointed out. Opening lines and exposing the white king just has to lead to favorable tactics.

It's one of those priyome-type moves that just pops out, and then you have to start calculating to see where it goes and if it actually works. I'd play it immediately in a blitz game, but calculate it in a slow game.

At a minimum, I think it is a basic line-clearing sacrifice, and I am not sure it is even a sac.

I guess this turns out to be an excellent position for the forum, because different players will see different things. That said, I don't see it as a machine-move as much as a human intuitive move that a machine confirms by calculation.

It's like a Bxh7 greek gift sacrifice- you can easily see when it might be possible, but you have to verify the conditions and calculate it to make sure it works.
14.  Paul Leggett
Chess Librarian
10 Jul '15 23:09 / 1 edit
Originally posted by Ragwort
I think the biggest threat of 1...d4 is Qd5 forking the two white rooks.

[fen]1k1r3r/ppp3p1/6Q1/3q4/1P1p4/P1b1P2P/R1P1KP2/2B4R b - - 0 20[/fen]

The rook on a2 looks particularly vulnerable with no escape squares. Moves like Bb2 intending to answer Qd5 with Rha1 fail to Bxb2 Rxb2 d3 check picking up the rook.

[pgn]
[FEN "1k1r3r/ppp3p1/6Q1/3pq3/1P6/ ...[text shortened]... s to tease out before you will be able to understand why the computer favours the lines it does.
Nice post- both white rooks are loose pieces, and I am annoyed at myself for not seeing that immediately.

I am one of those "1) Check all checks; and 2) Look for loose pieces written-on-a-Post It Note" guys inspired by Greenpawn34's advice, and this is another example of we chessplayers looking for sophisticated answers when the simple stuff is right in front of us.