deep.09

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Harvey Williamson, bob

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
jhellis3
Posts: 382
Joined: Fri Aug 16, 2013 10:36 pm

Re: deep.09

Post by jhellis3 » Thu Aug 25, 2016 12:10 am

Interesting stuff Sven. My own analysis agrees that Black appears to hang on after the c3 break (probably because it allows black to immediately pressure the B file), but I certainly did not exhaust all possibilities for White.

Here is an example continuation after the rook sac line on d4 though:
[pgn]
[Event ""]
[Site ""]
[Date "????.??.??"]
[Round ""]
[White ""]
[Black ""]
[Result "*"]
[FEN "r2r4/8/2p1k3/p2bPp1p/Pp1Q1PpP/1P1P2K1/2P5/4R3 b - - 0 1"]

1...Rd7 2.Qc5 Bf3 3.Re3 Bh1 4.d4 Bd5
( 4...Rb7 5.Qd6+ Kf7 6.e6+ Ke8 7.e7 Be4 8.Rxe4 fxe4 9.Qxc6+ Kf7 10.Qxb7 Re8 11.Qxe4 Rxe7 12.Qh7+ Ke6 13.Qxh5 Kd7 14.Qxa5 Re1 15.Kxg4 Kc8 )
5.Qb6 Rda7
( 5...Rc8 6.Qxa5 Rb7 7.Qc5 )
6.c4 bxc3 7.Rxc3 Kf7 8.Rxc6 Bxc6 9.Qxc6 Rb8 10.Kg2 Rab7 11.Qf6+ Kg8 12.Qxf5 Rxb3 13.Qg6+ Kh8 14.Qxh5+ Kg8 15.Qxg4+ Kf8 16.Qf5+ Kg7 17.Qg5+ Kh8 18.Qh5+ Kg8 19.e6 R3b7 20.Qxa5 Rg7+ 21.Kf3 Kh7 22.Qe5 Rb3+ 23.Ke4 Re7 24.Qf5+ Kg8 25.Qg5+ Rg7 26.Qd5 Rb8 27.f5 Rb1 28.f6 Re1+ 29.Kf3 Rf7 30.Qg5+ Kf8 31.exf7 Kxf7 32.Qg7+ Ke6 33.Qe7+ Kd5 34.Qxe1 Kc4 35.f7 Kd5 36.f8=Q Kc6 37.Qc5+ Kb7 38.Qee7+ Ka6 39.Qb5# *
[/pgn]

peter
Posts: 1760
Joined: Sat Feb 16, 2008 6:38 am
Full name: Peter Martan

Re: deep.09

Post by peter » Thu Aug 25, 2016 5:05 am

Sven Schüle wrote: [D]6r1/6r1/2p1k1pp/p1pbP2q/Pp1p1PpP/1P1P2NR/1KPQ3R/8 b - - 0 1
According to Dann the main line starts with ...

1...Qf5 2.Nxf5 gxf5 3.Rg3 h5 4.Re2 Re7 5.Re1 Rge8 6.Qe2 Kd7 7.Rgg1 Kc7 8.Qf2 Re6
Not knowing, where Dann got the position from and if this is an author's solution line at all, I still think Black doesn't have to move the King away from e6.
Sven Schüle wrote: But Black can improve Dann's main line earlier, in my opinion, by 4...Bf3! (instead of 4...Re7):

[D]6r1/6r1/2p1k3/p1p1Pp1p/Pp1p1PpP/1P1P1bR1/1KPQR3/8 w - - 0 5
5.Re4 fxe4 6.dxe4 Bxe4 7.Qe2 Bd5 8.Qa6 Kf5 9.Qxa5 Kxf4 10.Rg1 g3 11.Qxc5 g2 12.Qxb4 Kg3 (12...Kxe5 losing after 13.Re1+) 13.Qxd4 Rg4 14.Qe3+ Kh2, and now I can find no winning line for White.

4...Bf3 (?) for sure is not an improvement. Just playing 5.Rxf3 White kills the fortress at once.

Sven Schüle wrote: Therefore I keep up my original idea of sacrificing a rook on c3, as follows (after 1...Qf5 2.Nxf5 gxf5 3.Rg3 h5):

4.Rf2 Rd7
(4...Re8 seems to be losing for Black according to my analysis, although I may be wrong. My main line is:
5.Rf1 Rge7
That's an impossible move.
From the starting position

[D]6r1/6r1/2p1k1pp/p1pbP2q/Pp1p1PpP/1P1P2NR/1KPQ3R/8 b - - 0 1

and this one your line
1... Qf5 2. Nxf5 gxf5 3. Rg3 h5 4. Rf2 Rd7 5. Rf1
we have this
[D]6r1/3r4/2p1k3/p1pbPp1p/Pp1p1PpP/1P1P2R1/1KPQ4/5R2 b - - 0 5
and there's no Rook on g7 to move to e7, so 4...Rge7 must be a mistyping or come from another starting position.
Probably you mean just R(d)e7?. But can you please approve that by yourself, Sven, before I go on whith your new line?

If I got you correcty, Id probably not follow that line further than to 11.Kc2 and rather play ...Rdb7 instead of ...Ra3 (?!) but let me see first, if I haven't missed anything before.
I'd really like to go on discussing, I don't think I have shown the undenieable prove for drawing the starting position myself already, I just don't want having to start from positions again and again that probably cannot be reached or are not shown how to get there, you see?
:)
Peter.

peter
Posts: 1760
Joined: Sat Feb 16, 2008 6:38 am
Full name: Peter Martan

Re: deep.09

Post by peter » Thu Aug 25, 2016 5:31 am

peter wrote:[If I got you correcty, Id probably not follow that line further than to 11.Kc2 and rather play ...Rdb7 instead of ...Ra3 (?!) but let me see first, if I haven't missed anything before.
I'd yet rather spare the moves after 7.Qf2 and continue with 7...Reb7 instead of your ...Rd7 there.
Being not sure, if the latter one even killing the fortress wouldn't be to be held still, but it's not the idea of the whole thing, to trade material down, Black doesn't have to do so.

6r1/6r1/2p1k1pp/p1pbP2q/Pp1p1PpP/1P1P2NR/1KPQ3R/8 b - - 0 1

1... Qf5 2. Nxf5 gxf5 3. Rg3 h5 4. Rf2 Rd7 5. Rf1 Re7 6. Rc1 Rb8 7. Qf2 Reb7

you cannot play your c3 at once again without coming into disadvantage with White one more time
Peter.

Lyudmil Tsvetkov
Posts: 6052
Joined: Tue Jun 12, 2012 10:41 am

Re: deep.09

Post by Lyudmil Tsvetkov » Thu Aug 25, 2016 6:17 am

This is a draw, the fortress holds everywhere.

Black king stays on e6, if white sacrifices a rook on d4, black has a rook ready on c line and plays c5, then does nothing, c3 break is impossible, black has rooks on b or d file, etc.

If I had to play one side, I would definitely play black, I do not understand why black should be worse here, material is far from everything.

peter
Posts: 1760
Joined: Sat Feb 16, 2008 6:38 am
Full name: Peter Martan

Re: deep.09

Post by peter » Thu Aug 25, 2016 6:51 am

Thanks for having looked into it too, Lyudmil.
Lyudmil Tsvetkov wrote:This is a draw, the fortress holds everywhere.

Black king stays on e6, if white sacrifices a rook on d4, black has a rook ready on c line and plays c5, then does nothing, c3 break is impossible, black has rooks on b or d file, etc.
As much as I sometimes doubted your evaluations of the one and the other one (but always then very tricky) positions in the past, here I fully agree with you.
:)

Yet I'm still open for trials to find a way to open up the fortress without losing with White neither, I just didn't see any line doing that till now.

As for your comments about White sacrificing a Rook on d4, I'm still missing concrete lines even to come to that without White disadvantage already on the way there.
Sven had such a line, but he didn't see a White win neither and I would have deviated from his variant early before letting White come to the rook- sac at all
Peter.

Sven
Posts: 3819
Joined: Thu May 15, 2008 7:57 pm
Location: Berlin, Germany
Full name: Sven Schüle
Contact:

Re: deep.09

Post by Sven » Thu Aug 25, 2016 10:13 am

peter wrote:
Sven Schüle wrote:But Black can improve Dann's main line earlier, in my opinion, by 4...Bf3! (instead of 4...Re7):

[D]6r1/6r1/2p1k3/p1p1Pp1p/Pp1p1PpP/1P1P1bR1/1KPQR3/8 w - - 0 5
5.Re4 fxe4 6.dxe4 Bxe4 7.Qe2 Bd5 8.Qa6 Kf5 9.Qxa5 Kxf4 10.Rg1 g3 11.Qxc5 g2 12.Qxb4 Kg3 (12...Kxe5 losing after 13.Re1+) 13.Qxd4 Rg4 14.Qe3+ Kh2, and now I can find no winning line for White.

4...Bf3 (?) for sure is not an improvement. Just playing 5.Rxf3 White kills the fortress at once.
Agreed. Note that this would mean, if there would be no other improvements for Black, that 7.Qf2 instead of 7.Rgg1 would be winning for White.
peter wrote:
Sven Schüle wrote:Therefore I keep up my original idea of sacrificing a rook on c3, as follows (after 1...Qf5 2.Nxf5 gxf5 3.Rg3 h5):

4.Rf2 Rd7
(4...Re8 seems to be losing for Black according to my analysis, although I may be wrong. My main line is:
5.Rf1 Rge7
That's an impossible move.
No, it isn't. We are within the bracket opened for 4...Re8, so it is 4.Rf2 Re8 5.Rf1 Rge7. Unfortunately serious chess analysis is hardly possible without using brackets and even nested brackets, sorry for that. I could have posted the whole pgn of my analysis but that would even have been much harder to read for everyone.

All conclusions that have been drawn from the assumption that my line has an error are therefore not valid, it seems (unless there are other actual errors, of course) ...

Sven
Posts: 3819
Joined: Thu May 15, 2008 7:57 pm
Location: Berlin, Germany
Full name: Sven Schüle
Contact:

Re: deep.09

Post by Sven » Thu Aug 25, 2016 10:57 am

peter wrote:1... Qf5 2. Nxf5 gxf5 3. Rg3 h5 4. Rf2 Rd7 5. Rf1 Re7 6. Rc1 Rb8 7. Qf2 Reb7

you cannot play your c3 at once again without coming into disadvantage with White one more time
Again, I see no disadvantage for White (see below) but at least doubling rooks on b-file seems to hold the draw for Black if Black finds some "singular" moves:

8.Rgg1 Rb6 9.Ka1 (9.c3 bxc3+ 10.Rxc3 dxc3+ 11.Kc2 c4 (or 11...Lxb3+) 12.Qc5 cxd3+ 13.Kxd3 Be4+ 14.Ke3 Rxb3 15.Qd6+ Kf7 16.e6+ Kg8 17.Kd4 c5+ 18.Ke5 R3b6 19.Qc7 R6b7 20.Qxa5 c2 21.Qxc5 Ra8 22.Re1 Rba7 23.Kf6 Rf8+ 24.Ke5 Rfa8 =) 9...R6b7 10.c4 bxc3 11.Rxc3 dxc3 12.Qxc5 Rxb3 13.Qd6+ Kf7 14.e6+ Bxe6 15.Re1 Re8 16.Rxe6 Rxe6 17.Qd7+ Kf6 18.Qd8+ =

Sven
Posts: 3819
Joined: Thu May 15, 2008 7:57 pm
Location: Berlin, Germany
Full name: Sven Schüle
Contact:

Re: deep.09

Post by Sven » Thu Aug 25, 2016 11:25 am

After 1...Qf5 2.Nxf5 gxf5 3.Rg3 h5 4.Re2 Re7 5.Re1 Rge8 6.Qe2 Kd7 (main line given by Dann), I still believe that 7.Qf2 would now be winning for White, see my analysis above. But 6...Kd7 may be flawed. For 6.Rge3 (instead of 6.Qe2) I previously wrote that after 6...Rg8 7.Re4 g3 "[...] the original idea of Re4 to take d4 with the rook and then with the queen is no longer available for White". But I continued that analysis and found the following promising line:

1...Qf5 2.Nxf5 gxf5 3.Rg3 h5 4.Re2 Re7 5.Re1 Rge8 6.Rge3 Rg8 7.Re4 g3 8.R4e2 Rg4 9.Rg2 Reg7 10.Reg1 R7g6 11.Qd1

and now I only see winning lines for White, e.g.:
11...Rg7 12.Rxg3 Rxg3 13.Rxg3 Rxg3 14.Qxh5 +-
11...Bxg2 12.Rxg2 Ke7 13.Qf3 Kf7 14.Rxg3 Rxg3 15.e6+ Kxe6 16.Qxc6+ +-

The reason is 7...g3 which opens the fortress. But without g3, White will play the Joseph idea, e.g.:

7...Rge8 8.Qf2 Rg8 9.Rxd4 cxd4 (9...g3 10.Qg1) 10.Qxd4 which looks very promising for White.

Any improvements over that?

Sven
Posts: 3819
Joined: Thu May 15, 2008 7:57 pm
Location: Berlin, Germany
Full name: Sven Schüle
Contact:

Re: deep.09

Post by Sven » Thu Aug 25, 2016 12:00 pm

jhellis3 wrote:Interesting stuff Sven. My own analysis agrees that Black appears to hang on after the c3 break (probably because it allows black to immediately pressure the B file), but I certainly did not exhaust all possibilities for White.

Here is an example continuation after the rook sac line on d4 though:
[pgn]
[Event ""]
[Site ""]
[Date "????.??.??"]
[Round ""]
[White ""]
[Black ""]
[Result "*"]
[FEN "r2r4/8/2p1k3/p2bPp1p/Pp1Q1PpP/1P1P2K1/2P5/4R3 b - - 0 1"]

1...Rd7 2.Qc5 Bf3 3.Re3 Bh1 4.d4 Bd5
( 4...Rb7 5.Qd6+ Kf7 6.e6+ Ke8 7.e7 Be4 8.Rxe4 fxe4 9.Qxc6+ Kf7 10.Qxb7 Re8 11.Qxe4 Rxe7 12.Qh7+ Ke6 13.Qxh5 Kd7 14.Qxa5 Re1 15.Kxg4 Kc8 )
5.Qb6 Rda7
( 5...Rc8 6.Qxa5 Rb7 7.Qc5 )
6.c4 bxc3 7.Rxc3 Kf7 8.Rxc6 Bxc6 9.Qxc6 Rb8 10.Kg2 Rab7 11.Qf6+ Kg8 12.Qxf5 Rxb3 13.Qg6+ Kh8 14.Qxh5+ Kg8 15.Qxg4+ Kf8 16.Qf5+ Kg7 17.Qg5+ Kh8 18.Qh5+ Kg8 19.e6 R3b7 20.Qxa5 Rg7+ 21.Kf3 Kh7 22.Qe5 Rb3+ 23.Ke4 Re7 24.Qf5+ Kg8 25.Qg5+ Rg7 26.Qd5 Rb8 27.f5 Rb1 28.f6 Re1+ 29.Kf3 Rf7 30.Qg5+ Kf8 31.exf7 Kxf7 32.Qg7+ Ke6 33.Qe7+ Kd5 34.Qxe1 Kc4 35.f7 Kd5 36.f8=Q Kc6 37.Qc5+ Kb7 38.Qee7+ Ka6 39.Qb5# *
[/pgn]
Seems correct. I would have stopped that analysis after 7.Rxc3 since this is obviously a simple win for White, considering that the queen will razor almost all black pawns.

But I also don't think this line is relevant for the original problem since I don't see why it would be necessary to move the white king to g3 anymore, see my previous posts of today in another subthread. Your idea of opening the fortress through the center seems to work well even without Kg3.

jhellis3
Posts: 382
Joined: Fri Aug 16, 2013 10:36 pm

Re: deep.09

Post by jhellis3 » Thu Aug 25, 2016 1:11 pm

But I also don't think this line is relevant for the original problem since I don't see why it would be necessary to move the white king to g3 anymore. ... Your idea of opening the fortress through the center seems to work well even without Kg3.
Yeah, the king walk is likely optional. I just included it because it really illustrates Black's helplessness. And it is certainly still relevant, as it serves as a lower bound for White.

Post Reply