English Chess Forum

Joey Stewart wrote: ↑

Fri Nov 13, 2020 9:40 pm

Genuine 1300s have no place beating genuine 2100s

Joey Stewart wrote: ↑

Fri Nov 13, 2020 9:40 pm

Genuine 1300s have no place beating genuine 2100s - either they are using assistance or they are a 2000+ level player who has deliberately engineered their rating low in order to win more grade limited prizes so by hook or by crook they still end up guilty.

MJMcCready wrote: ↑

Fri Nov 13, 2020 11:53 pm

If you recall, and it was only 14 years ago or thereabouts, a player rated 1800 defeated a GM with the Smith Morra Gambit

MJMcCready wrote: ↑

Sat Nov 14, 2020 12:29 pm

I remember thinking why on earth someone of his strength couldn't see it. It was the main opening trap in that line.

MJMcCready wrote: ↑

Sat Nov 14, 2020 12:29 pm

It's a long time ago and its not archived on their site. I don't know that opening well at all but there's a move black must not play as it loses on the spot almost, which is what the GM did. He was lost within 10 moves and resigned around move 16 I think, one of the players was German but I can't really remember much more than than other than the move that cost the GM was so obviously bad, I remember thinking why on earth someone of his strength couldn't see it. It was the main opening trap in that line.

Roger de Coverly wrote: ↑

Sat Nov 14, 2020 12:38 pm

MJMcCready wrote: ↑

Sat Nov 14, 2020 12:29 pm

I remember thinking why on earth someone of his strength couldn't see it. It was the main opening trap in that line.

The Morra Gambit is usually considered so dubious that it just isn't played amongst GMs, so even a GM stumbling into the position might not be aware of the trap. There's also the effect that some amateurs can be GM standard in their pet openings, but pretty poor outside that. Also not all GMs are particular experts at sharp and sometimes irrational positions. All in all there's a sufficient tail in the expected result curves at extreme rating differences that an occasional anomalous result is to be anticipated intermittently.

Geoff Chandler wrote: ↑

Fri Nov 13, 2020 11:59 pm

Hi Roger,

"Has this test ever been done? Take GM games from the pre-computer era..."

This was done by a lad on the Fischer-Spassky match who was very good at ridding RHP of cheats
using the top three match up system.

The kit and system he used was:

Fritz 11 @ 30 seconds per move
Pentium 4 2.93GHz 1GB RAM
Hash Table 192MB
Database used www.chesslive.de

Well it's not as clear cut as people think as some of the accusations levied may be nothing more than sour grapes. Certain moves are identifiable as computer moves, we all know that. But how do you factor in form and that stronger players sometimes go wrong and hand it on a plate to their opponent. I've played people 400 points above me and took them to pieces. A friend had someone 760 ELO points on the ropes completely and should have won, the GM, who shall remain nameless, even told him that and showed him the win he missed. It's long overdue but chess really does need to remove its fixation with numbers. It's become a running joke.

"For obvious reasons a game is classed as “out of book” once I cannot find the position in the
chesslive database prior to the 1972 Fischer-Spassky World Championship."

Result: (over the 21 games)
Fischer
Top 1 Match: 385/658 (58,5% )
Top 2 Match: 509/658 (77,4% )
Top 3 Match: 563/658 (85,6% )

Spassky
Top 1 Match: 368/657 (56,0% )
Top 2 Match: 461/657 (70,2% )
Top 3 Match: 525/657 (79,9% )

Conclusion:

So, in the 1972 Fischer-Spassky match my analysis confirms that Fischer achieved results comparable to the extreme upper limit considered humanly possible/overwhelming evidence of engine use. His top 2 & top 3 matchup rates were actually a small % over this threshold.
Spassky, an incredibly strong player in his time, attained % rates around the totals that I was secretly expecting Fischer to get.
So the next time someone shouts from the rooftops that such & such a player is clearly a cheat because they exceeded some of the thresholds of
Top 1 Match: 60,0%
Top 2 Match: 75,0%
Top 3 Match: 85,0%
and no human CC player can do that in many games with non-book moves, you can point them in the direction of this thread
and say yes they can; Fischer achieved this against Spassky, over-the-board nearly 40 years ago!

The full analysis of all 21 games and further comments here:

https://www.redhotpawn.com/forum/linked ... ead.106405

I can remember commentating that Fischer taking the h-pawn in game one was the computer's top choice!

As Justin comments it's not conclusive. However if a player has a higher 1st choice match up over 20 games of say 80%
(and the games must be fairly long and v reasonable opposition) it can indicate the player needs investigating.
And from there the games and time taken for moves need going over with a fine toothed comb.

Having said that, I agree with Dr. Regan that Stealth Cheats, those that use a computer for one or two moves
when the position is critical. (good players can recognise this, the moment when only one move will do).
These players will always slip under the radar.

Hi MJM,

"If you recall, and it was only 14 years ago or thereabouts, a player rated 1800 defeated a GM with the Smith Morra Gambit."

That is why cheat hunters need to look at a large batch of games. Freak one off games can and do happen.