Dylan Sharp Vs. Harvey Williamson (G4)

[Raphexon]

0.00 is only for positions that you are sure that they are drawn or you think both sides have equal probability to win.

I'm sure those positions are drawn even though I can't prove it. Also, Harvey's moves up to this point have been pretty much forced, because other alternatives were 0.00. Even if I can't prove it.

I think you and other people on the forum have this idea that you have to prove something to know it, but sometimes proof is something unreachable, and you just need to reach a level of certainty to know something.

You don't use 7 piece TBs, do you Ovy?

I don't even have 6 piece TBs

That's kinda impressive then.

[Ovyron]

And I'll claim that in the position after the move I'm making, 25...Be6 is the only black's move that isn't 0.00, so I know Harvey will make it, though I can't prove it.

The idea is that if I'm able to predict all of Harvey's moves, then I can draw the game, but it takes a while...

1. g4 d5 2. g5 e5 3. d4 exd4 4. Nf3 c5 5. Bg2 Ne7 6. c3 dxc3 7. Nxc3 Nbc6 8. O-O d4 9. Ne4 Ng6 10. a3 Bg4 11. Bd2 Bd6 12. Rc1 O-O 13. Nxc5 Bxc5 14. Rxc5 Qd7 15. Qb3 Rad8 16. Re1 Be6 17. Qd3 Rfe8 18. e4 b6 19. Rcc1 Bg4 20. Rxc6 Qxc6 21. Nxd4 Qc5 22. Be3 Ne5 23. Qc3 Qxc3 24. bxc3 Rc8 25. f3

[d]2r1r1k1/p4ppp/1p6/4n1P1/3NP1b1/P1P1BP2/6BP/4R1K1 b - -

[Harvey Williamson]

[pgn]1. g4 d5 2. g5 e5 3. d4 exd4 4. Nf3 c5 5. Bg2 Ne7 6. c3 dxc3 7. Nxc3 Nbc6 8. O-O d4 9. Ne4 Ng6 10. a3 Bg4 11. Bd2 Bd6 12. Rc1 O-O 13. Nxc5 Bxc5 14. Rxc5 Qd7 15. Qb3 Rad8 16. Re1 Be6 17. Qd3 Rfe8 18. e4 b6 19. Rcc1 Bg4 20. Rxc6 Qxc6 21. Nxd4 Qc5 22. Be3 Ne5 23. Qc3 Qxc3 24. bxc3 Rc8 25. f3 Be6[/pgn]

Predictable!

[jp]

Move?!

[Ovyron]

No way.

I'm going ahead and admit that if this was a "one week per move" (or faster) game Harvey would have probably beaten me by force (...unless I draw and I'd have played the same moves anyway).

I guess I can give something for the audience to look at, huh... what about positions analysed for my Candidate Moves?

1. Nb5 (2872 nodes analyzed)
2. Rc1 (605 nodes analyzed)
3. Bd2 (182 nodes analyzed)
4. Ne2 (1 node analyzed - deemed lost )

So I'll probably end up playing Nb5 (lest I lose +78% of my analysis), as Rc1 would need to overcome it and then stay ahead until I analyze as many nodes for it.

But I don't think anybody would bet the house for what move I'll play

[Ovyron]

Update (to save the thread from page 2)

1. Nb5 (4187 nodes analyzed)
2. Rc1 (605 nodes analyzed)
3. Bd2 (182 nodes analyzed)
4. Ne2 (1 node analyzed - deemed lost )

Apparently I haven't analyzed anything else but Nb5, though I've rechecked positions to look for alternatives in the other moves.

[jp]

4. Ne2 (1 node analyzed - deemed lost )

Explain this "1 node analyzed". For that to be even vaguely true, I guess you have to think about that move for a split-second and immediately judge it lost. Otherwise it would be more than 1 node... Or maybe I should step back and ask what you mean by "n nodes analyzed".

[Uri Blass]

4. Ne2 (1 node analyzed - deemed lost )

Explain this "1 node analyzed". For that to be even vaguely true, I guess you have to think about that move for a split-second and immediately judge it lost. Otherwise it would be more than 1 node... Or maybe I should step back and ask what you mean by "n nodes analyzed".

I guess he meant something like after
1.Ne2 xx when xx is black's move the score of stockfish after 999999999999 nodes is more than +4 pawns for black so he considers it as a win for black.

xx is the only node in his tree.
He does not count the 999999999999 nodes of stockfish.

I do not say the number are 4 pawns and 999999999999 nodes and his numbers of course may be different.

[Ovyron]

Whoa, long post lost.

But anyway, the point is that, yeah, Ne2 is below some threshold (that is deemed lost) and this is reached by only examining its node (the positions where Ne2 is played) so my engine might have analyzed more than a hundred million nodes of it but I only count the time I'm on the position and ask the engine for analysis, as Uri said.

[jp]

Okay, but if that number is the number of nodes you "personally dwell on" (to make up a description), does that mean you personally dwelt upon 4187 positions starting from Nb5 (with SF calculating "999999999999 nodes" at each)?