GM Kaufman v.Novag Constellation 3.6 QR Odds (Moves 1-40)

[Steve B]

I've been analyzing the game with a computer - and I still think Larry has his work cut out for him. In fact, I think the game may even be a technical win for black if black plays really well, or at least a very difficult draw for white.

However, I don't think Connie will play the rest of the game that well - another mistake on Connie's part and Larry will win.

I give Larry about 95% chance of winning here since the Connies pawns look so dangerous - but I expect Connie to make another game changing mistake in the next few moves as I doubt she will play the endgame as well as she played the middlegame.

If I remember correctly Connie has no hash tables - and this is going to be a real killer as more pieces eventually come off the board.

Meanwhile I'm playing this out using stockfish 1.8, I'm letting stockfish play each move with a 1 billion node search - just to see if black has enough advantage to win.

not only is there no hash table RAM.. but there is also no CMOS memory ..so she also will not retain the position if i shut her down and the Connie 3.6 has no option to disable pondering so i have been setting up the position as a new position after every move and letting Connie ponder for 3 minutes
i wrote about this in one of the earlier posts in the thread


for the next game against Movei i will choose a 1700 dedicated that either allows me to disable pondering or will retain the position in memory if i shut the computer down

here are Connie's hardware specs:

http://www.schach-computer.info/wiki/in ... lation_3.6

Steve

[Steve B]

I'm able to log in again now, thanks to Sam. I play 34.hxg6. My only chance is to weaken the Black king. ..Re5 was a bit strange looking, but perhaps not a bad move.

Well Connie finally goes haywire with all of the possible Pins,Skewers,Discovered Checks ,Forks and Captures
She decides to recapture the P in the worst possible way with the disastrous ...

34..Qxg6

[d]6k1/p1pr1p1p/1p4q1/3nr1B1/2Bp2R1/1Q6/P4P1K/8 w - - 0 34

Excellently played by GM Kaufman

Hopefully she can still cling on to the draw

Sigh Regards
Steve

It will be interesting to see GM Kaufman's reply. Black's ...Qg6 does not look as bad to me as maybe it should.

I think that after this move Larry is likely to win the game, even if Connie still technically has the win (which I'm not at all sure of.)

Unless I am blinded, after the obvious Qg3 white gets most of the material back (netither h5 or f5 seems to help, because of Qxe5 and later Qe8+ Qxd7 or Qxe5-Bxd5+, respectively), which means Connie is toasted. The black pawns will start falling like flies.

Miguel

That was my first-glance opinion too. I said a while back that the h4 stuff could be a problem for a non-deep-searcher here as one mistake and the roof falls in.

shingles-in-hair regards,

Bob

(typical Steve sign-off)

More Like Shingles -in -Grey -Hair ..after that Howler

how many Class A players would actually allow the Q to step into a pin like that?

anyway.. Larry Set this up beautifully and deserves full credit
every last piece in his diminished arsenal aimed at the King..
Pins traps and pitfalls...

If Connie does manage to loss this game ..she will be

Ebay Bound Regards
Steve

Don't be so cruel! Haven't you seen the movie Toy Story?

Miguel

you have a kind heart Miguel
i am not so kind
if Connie essays forth another howler.. she goes on Ebay AFTER a short spell as a dinner tray

Qg6 Regards
Steve

[Don]

I'm able to log in again now, thanks to Sam. I play 34.hxg6. My only chance is to weaken the Black king. ..Re5 was a bit strange looking, but perhaps not a bad move.

Well Connie finally goes haywire with all of the possible Pins,Skewers,Discovered Checks ,Forks and Captures
She decides to recapture the P in the worst possible way with the disastrous ...

34..Qxg6

[d]6k1/p1pr1p1p/1p4q1/3nr1B1/2Bp2R1/1Q6/P4P1K/8 w - - 0 34

Excellently played by GM Kaufman

Hopefully she can still cling on to the draw

Sigh Regards
Steve

It will be interesting to see GM Kaufman's reply. Black's ...Qg6 does not look as bad to me as maybe it should.

I think that after this move Larry is likely to win the game, even if Connie still technically has the win (which I'm not at all sure of.)

Unless I am blinded, after the obvious Qg3 white gets most of the material back (netither h5 or f5 seems to help, because of Qxe5 and later Qe8+ Qxd7 or Qxe5-Bxd5+, respectively), which means Connie is toasted. The black pawns will start falling like flies.

Miguel

That was my first-glance opinion too. I said a while back that the h4 stuff could be a problem for a non-deep-searcher here as one mistake and the roof falls in.

shingles-in-hair regards,

Bob

(typical Steve sign-off)

More Like Shingles -in -Grey -Hair ..after that Howler

how many Class A players would actually allow the Q to step into a pin like that?

anyway.. Larry Set this up beautifully and deserves full credit
every last piece in his diminished arsenal aimed at the King..
Pins traps and pitfalls...

If Connie does manage to loss this game ..she will be

Ebay Bound Regards
Steve

Don't be so cruel! Haven't you seen the movie Toy Story?

Miguel

you have a kind heart Miguel
i am not so kind
if Connie essays forth another howler.. she goes on Ebay AFTER a short spell as a dinner tray

Qg6 Regards
Steve

What is your buy it now price?

[Steve B]

What is your buy it now price?

actually there is a Super Connie(with clock attachment) on Ebay even as we speak..

http://cgi.ebay.com/Schachcomputer-Nova ... 0435035004

and a Connie 2Mhz sold only a few days ago
Steve

[lkaufman]

[quote="Steve B

Thanks Don for relaying Larry's move once again
i dont know why he is having all of this trouble given that the TCADMIN helped him to reset his account

anyway.....

Connie i guess shows us why she moved the R to e5 on the move before last
here.. she unleashes ..in Judit Polgar like fashion ..the shocker...

35..Rf5

[d] 6k1/p1pr1p1p/1p4q1/3n1rB1/2Bp2R1/6Q1/P4P1K/8 w - - 0 35

removes the R from attack,threatens the f2 pawn with check and has a "rattling effect "on the psyche of the opponent

Whom i kidding? regards
Steve[/quote]

I play 36.Bh4. So I'll have queen vs. two rooks, and a bishop vs. a whole lot of pawns. Now I understand why Connie played ...Qg6?? As I suspected, she found a delaying move (...Rf5) which means that the mistake required a six ply search to detect, and I believe Connie usually does no more than 5 ply in 3 minutes. So no big mystery here. It just shows that computers sometimes require more ply to see things than would be the case for humans. We see that after we pin the queen it doesn't matter how many delaying moves like checks the other side has (unless they lead to something concrete). So sometimes "obvious" tactics take the engines several ply to see.

At the start of the game, I had to decide whether to play on Connie's primitive eval or on her short search. I judged that her short search was by far the bigger weakness and so chose an open game rather than an anti-computer strategy. If I were playing some deep searching program with minimal eval (for example Crafty at lowest level which Bob calls 1800 or so) at rook odds I would play closed openings. It looks now like I made the right call.

Larry

[Don]

Thanks Don for relaying Larry's move once again
i dont know why he is having all of this trouble given that the TCADMIN helped him to reset his account

anyway.....

Connie i guess shows us why she moved the R to e5 on the move before last
here.. she unleashes ..in Judit Polgar like fashion ..the shocker...

35..Rf5

[d] 6k1/p1pr1p1p/1p4q1/3n1rB1/2Bp2R1/6Q1/P4P1K/8 w - - 0 35

removes the R from attack,threatens the f2 pawn with check and has a "rattling effect "on the psyche of the opponent

Whom i kidding? regards
Steve

I play 36.Bh4. So I'll have queen vs. two rooks, and a bishop vs. a whole lot of pawns. Now I understand why Connie played ...Qg6?? As I suspected, she found a delaying move (...Rf5) which means that the mistake required a six ply search to detect, and I believe Connie usually does no more than 5 ply in 3 minutes. So no big mystery here. It just shows that computers sometimes require more ply to see things than would be the case for humans. We see that after we pin the queen it doesn't matter how many delaying moves like checks the other side has (unless they lead to something concrete). So sometimes "obvious" tactics take the engines several ply to see.

At the start of the game, I had to decide whether to play on Connie's primitive eval or on her short search. I judged that her short search was by far the bigger weakness and so chose an open game rather than an anti-computer strategy. If I were playing some deep searching program with minimal eval (for example Crafty at lowest level which Bob calls 1800 or so) at rook odds I would play closed openings. It looks now like I made the right call.

Larry

I'm playing the game out with stockfish at 1 billion nodes per move and it is looking drawish - unfortunately I don't see a way to alter the contempt factor in stockfish so it's tough to see what would happen if both sides were to fight.

But the game is not over yet - the position is not yet repeating and there are still non-reversable moves in the PV and being made.

[Terry McCracken]

[quote="Steve B

Thanks Don for relaying Larry's move once again
i dont know why he is having all of this trouble given that the TCADMIN helped him to reset his account

anyway.....

Connie i guess shows us why she moved the R to e5 on the move before last
here.. she unleashes ..in Judit Polgar like fashion ..the shocker...

35..Rf5

[d] 6k1/p1pr1p1p/1p4q1/3n1rB1/2Bp2R1/6Q1/P4P1K/8 w - - 0 35

removes the R from attack,threatens the f2 pawn with check and has a "rattling effect "on the psyche of the opponent

Whom i kidding? regards
Steve

I play 36.Bh4. So I'll have queen vs. two rooks, and a bishop vs. a whole lot of pawns. Now I understand why Connie played ...Qg6?? As I suspected, she found a delaying move (...Rf5) which means that the mistake required a six ply search to detect, and I believe Connie usually does no more than 5 ply in 3 minutes. So no big mystery here. It just shows that computers sometimes require more ply to see things than would be the case for humans. We see that after we pin the queen it doesn't matter how many delaying moves like checks the other side has (unless they lead to something concrete). So sometimes "obvious" tactics take the engines several ply to see.

At the start of the game, I had to decide whether to play on Connie's primitive eval or on her short search. I judged that her short search was by far the bigger weakness and so chose an open game rather than an anti-computer strategy. If I were playing some deep searching program with minimal eval (for example Crafty at lowest level which Bob calls 1800 or so) at rook odds I would play closed openings. It looks now like I made the right call.

Larry[/quote]

Yes, you certainly did. I was wrong about how you approached the game.
Sorry for anything I may have said that could put you off.

I'm also sorry for snapping at Don. A terrible misunderstanding and my fault.

I find Rf5 even more interesting/clever then h5 and appears to give more resistance. But a poor Endgame by the Connie and you'll wrap her up for Christmas! I suspect the Connie will have a harder time with an Endgame and/or basic ending then the Middlegame like Don mentioned. It was a weak point with dedicated units and Novag may have been the weakest compared to Fidelity or Mephisto. Of course you know this best.

[Terry McCracken]

That's why, particularly in the eighties computers were so easy to smash.
For the last 10 years they've been very difficult to win against and since 05, well my days of smashing computers is long over.

Still, Qg6?? was a bit shocking at first sight. I had expected the Connie to either play something weaker than it had far earlier in the game, it hadn't, then in the last two moves it starts to really play strange stuff, Re5(?), Re8 was obvious and now one of the worst mistakes I've seen even from a dedicated from the mid eighties. Larry got what he needed all at once, amazing.

Qg6 appears to look good to computers at low depth. GarboChess JS still likes it at 5 ply (switches to hxg6 at 6 ply).

Here it it's rather optimistic view of Qxg6
Ply:5 Score:3772 Nodes:44181 NPS:68075 Qxg6 Qg3 Re4 Bf4 Nxf4

And hxg6:
Ply:7 Score:3586 Nodes:293279 NPS:67327 hxg6 Bb5 Qd6 Bxd7 Qxd7 Qd1 c5

I don't think we should be spewing out variations - this is supposed to be Larry vs Connie, right? And so we should probably not be giving computer analysis or our own analysis - just for the sake of decorum.

Is guess the move ok? Either for one or both sides without analysis? Steve was fine with it and Larry, I don't think said anything about it.

I've been analyzing the game with a computer - and I still think Larry has his work cut out for him. In fact, I think the game may even be a technical win for black if black plays really well, or at least a very difficult draw for white.

However, I don't think Connie will play the rest of the game that well - another mistake on Connie's part and Larry will win.

I give Larry about 95% chance of winning here since the Connies pawns look so dangerous - but I expect Connie to make another game changing mistake in the next few moves as I doubt she will play the endgame as well as she played the middlegame.

If I remember correctly Connie has no hash tables - and this is going to be a real killer as more pieces eventually come off the board.

Meanwhile I'm playing this out using stockfish 1.8, I'm letting stockfish play each move with a 1 billion node search - just to see if black has enough advantage to win.

Yes he does, but like you, think he has excellent chances to catch another bad error, especially in the Endgame, and the house will come down.

Demolition Man Regards,

[Don]

Yes, you certainly did. I was wrong about how you approached the game.
Sorry for anything I may have said that could put you off.

I'm also sorry for snapping at Don. A terrible misunderstanding and my fault.

This was just as much me as you - I certainly didn't cut you any slack and I should have not been so harsh. Sorry.

Don

[Terry McCracken]

Yes, you certainly did. I was wrong about how you approached the game.
Sorry for anything I may have said that could put you off.

I'm also sorry for snapping at Don. A terrible misunderstanding and my fault.

This was just as much me as you - I certainly didn't cut you any slack and I should have not been so harsh. Sorry.

Don

Thanks! I really appreciate that! I do need to keep my temper in check and as Bob said to write with greater clarity and thoughtfulness.

Best,