Chess Genius Pro

[Scally]

It took exactly 4 minutes for my Rev II running Stockfish6 to find the mate in 6

Al.

[Yarc]

Hi Mike,

I guess your laptop is (at least!) about a factor 20 faster then the CG Pro, and Stockfish will use much more hashtable space then the poor 160 mb the cgp has. So...

I think you made my point
Regards, Paul

Hi Paul, a factor of 20 works out well with the MCGP solving time of just over 1 hour for the mate in 6 with Stockfish taking 30 seconds. Based on this result the MCGP has done well.

Regards
Ray

[Yarc]

This could be a time management problem. When that position of checkmating came up what time was left on the machine?

Checking clocks regards
Ricardo

I mentioned this possibility in a previous post. However solving time in analysis mode is one hour 9 minutes, so unfair to expect these dedicated machines to see the mate in just minutes during a tournament game.

Regards
Ray

[swarm]

Regarding the posted comment by the member of the Hiarcs Chess Forums website with the screen name ''swarm'' that he had Stockfish 7 look at the position and that it took Stockfish 7 about 9 minutes to spot the winning 6-move sequence, I had Stockfish 7 on my laptop computer examine the position in the game between MCG Pro and Novag Star Diamond and Stockfish 7 on my laptop computer spotted the winning 6-move sequence in about 30 seconds or less than 30 seconds!............Best Wishes, Mike.

Hello I was meaning stockfish 7 running at 120 mhz would solve in about 9 minutes at that speed ..judging by the time it took to solve it on my laptop ..not sure if that works out right or not so I am guessing

[Yarc]

Just completed game 4 of my 10 game match between the MCGP and Stardiamond. Both computers had an evaluation hovering around zero for most of the game, and at one point I thought it could have been another draw but the MCGP outplayed the SD in the end game. I think this is where its extra processing grunt helped it out. That said, I still think the SD is showing its worth at this time control.

In this game they followed the Sicilian, Chekover, Zaitsev variation.
Both were out of book at move 8.

[Event "40m/2h H book"]
[Round "4"]
[White "MCGP"]
[Black "SD"]
[Result "1-0"]
[ECO "B53"]

1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Qxd4 Nc6 5.Bb5 Qd7 6.Qd3 a6 7.Bxc6 Qxc6 8.Nc3 Nf6 9.Nd5 e5 10.Nxf6+ gxf6 11.O-O Be6 12.Re1 O-O-O 13.c3 d5 14.exd5 Rxd5 15.Qe4 Be7 16.Kh1 Rhd8 17.Qxh7 R5d7 18.Kg1 Bd5 19.Nh4 Bc5 20.Qf5 Kb8 21.Be3 Bxe3 22.Rxe3 Bxa2 23.Nf3 Qe6 24.Qxe6 Bxe6 25.h4 Kc7 26.g3 Bg4 27.Kg2 Rd1 28.Rxd1 Rxd1 29.Re4 Bf5 30.Rc4+ Kd6 31.h5 b5 32.Rh4 Bh7 33.h6 e4 34.Nd4 Rd2 35.Rf4 Rxb2 36.Rxf6+ Ke7 37.Rxa6 e3 38.Nc6+ Ke6 39.Kf3 exf2 40.Nb4+ Ke5 41.Ra1 Bb1 42.Kg2 Kf6 43.Ra7 Bf5 44.Nd5+ Kg5 45.Nf4 Kxh6 46.Rxf7 Kg5 47.Re7 Rc2 48.Re3 Bg4 49.Re5+ Kf6 50.Rxb5 Rd2 51.c4 Ke7 52.Nd5+ Kf7 53.Rb7+ Kf8 54.Ne3 Bd7 55.Rb6 Kf7 56.Rh6 Kg7 57.Rh4 Re2 58.Nd5 Ba4 59.Kf1 Rd2 60.Rg4+ Kf7 61.Rf4+ Ke6 62.Rxf2 Rd1+ 63.Ke2 Rg1 64.Kf3 Bc6 65.Kf4 Kd6 66.Ne3 Kc5 67.Ra2 Kd4 68.Rd2+ Kc5 69.g4 Rh1 70.g5 Rh3 71.Nf5 Rc3 72.g6 Rxc4+ 73.Ke5 Re4+ 74.Kf6 Ba4 75.g7 Bb3 76.Rc2+ Bxc2 77.g8=Q Kb4 78.Nd6 Kc5 79.Nxe4+ Bxe4 80.Ke5 1-0

Final position:

[fen]6Q1/8/8/2k1K3/4b3/8/8/8 b - - 1 80[/fen]

The SD resigned here which is just as well because it didn't stand a chance!

So for the first time in this match the MCGP takes the lead with 2.5 v 1.5 points.

I believe - having been a subscriber of the excellent Selective Search magazine (no longer published ) and from reading this - the Stardiamond has a more knowledge-based program. Maybe this is why it does better at longer time controls? Anyway, on to game 5 soon. Will the MCGP playing as black win another or will the SD level the score?

Exciting match Regards
Ray

[Steve B]

So for the first time in this match the MCGP takes the lead with 2.5 v 1.5 points.

Match Prediction holds steady at :
MCGP 6.5-3.5

4-2 To Go Regards
Steve

[Yarc]

Match Prediction holds steady at :
MCGP 6.5-3.5

4-2 To Go Regards
Steve

It does indeed.
What is your prediction for the next match with the SD and using the London book?

I'm keeping an open mind Regards
Ray

[Steve B]

All games where vs CGPro
Results active chess

CGPro -Res I Ruffian 9,5-10,5
CGPro- Star Diamond 10-0
CGPro- Atlanta 7,5-2,5
CGPro -Magellan -7,5-2,5
CGPro -V11 5,5-4,5
CGPro - CM Madrid 3.1 Aggr. 7-3
CGPro- R30 2.23 Aktiv 4,5-5,5
CGPro-Montreux Aktiv 5-5
CGPro- Risc 2 1MB 6-4

tournament
CGPro-R30 2.23 4,5-5,5

Thanks Hans
Which Book Regards?

Hello Steve,

Your favourite Book
Kr,Hans

All your games above were with the H Book ...Yes??
Do you know of any games CGPRO(London)Vs the Sparc Module?

The Spracklens Send Their Regards
Steve

[Steve B]

Match Prediction holds steady at :
MCGP 6.5-3.5

4-2 To Go Regards
Steve

It does indeed.
What is your prediction for the next match with the SD and using the London book?

I'm keeping an open mind Regards
Ray

Ill predict now that the London Book will perform worse but HOW much worse is still not know
need to develop a formula...
im thinking something like...CGPRO(London)=CGPRO(Hiarcs)-75

E=MC2 Regards
Steve

[donkeylane]

And E equals energy in joules,a joule being the amount of energy needed to move one kilogram one metre in one second,which then equals M which is mass in kilograms multiplied by C squared which is the speed of light in metres per second which is 300,000,000 times by 300,000,000.Honestly!

[lexman]

I think its hard to come to a firm prediction over such a short number of games in regard to what a particular book will score. My own hypothesis having seen a number of CG Pro games is that it tends to do less well in blocked strategic positions. This is true of a lot of the programs of the vintage late nineties early noughties and was the basis of a lot of anti computer chess that made a many of the man machine matches of that era so interesting. It is again my impression that some knowledge was stripped out to make it a faster searcher which was again fashionable in that period as the speed of cpus increased rapidly.
The London book seems designed to go into a higher percentage of these blocked maneuvering games since the London program being more knowledge based was very good at them, but the CG Pro is less so, in my view based on the games I have seen. So I guess the first variable is how many more strategic games ensue, and the second variable is how good the opposition program is at this as well. There is a random factor in this in a short series.
What the CG Pro is excellent at, given its relatively fast CPU, is tricky tactical endgames where it will regularly search 15 ply deep and selectively probably a good deal further. This is to much for many of the old dedicateds. So a third variable is how many of these kind of endgames ensue But a slower tournament time limit may even some of this out.. In my own series against the London Pro the London book scored 2117 while the Hiarcs book scored 2516. Pretty stark. But as to a precise prediction in regard to the StarDiamond
with so many variables
its beyond me regards

[Yarc]

I agree that 10 game matches are not enough. Even if the next match I perform using the London book produces a stark difference, even this is not enough to make a conclusive decision. However, add these results to others that hopefully will come along then maybe we will see a pattern. But as you say, there are many variables and as I've said before, maybe the London book might be more effective against some opponents than the Human book and vice-versa.

Regards
Ray

[lexman]

I guess the issue is that if there is a significant difference in the elo performance with each book it makes it harder to come to a reliable conclusion with the computer as a whole. The german site seems to be almost exclusively using the London book in games which I guess will feed into the final wiki rating and perhaps reflect unfairly on the CG Pro. An ideal would be to have a different rating for each book, as some computers do which have a choice between a brute force and a selective search in their capabilities.

[Steve B]

The german site seems to be almost exclusively using the London book in games which I guess will feed into the final wiki rating and perhaps reflect unfairly on the CG Pro. .

Yes...
The London book became the "automatic" choice for the first matches played... this ..Sadly.. a result of the early claims that the H book was not well suited for the Lang program in the CGPRO.

We here are trying to dispel that claim by using the H book

Correcting The Record Regards
Steve

[ricard60]

Regarding the posted comment by the member of the Hiarcs Chess Forums website with the screen name ''swarm'' that he had Stockfish 7 look at the position and that it took Stockfish 7 about 9 minutes to spot the winning 6-move sequence, I had Stockfish 7 on my laptop computer examine the position in the game between MCG Pro and Novag Star Diamond and Stockfish 7 on my laptop computer spotted the winning 6-move sequence in about 30 seconds or less than 30 seconds!............Best Wishes, Mike.

Hello I was meaning stockfish 7 running at 120 mhz would solve in about 9 minutes at that speed ..judging by the time it took to solve it on my laptop ..not sure if that works out right or not so I am guessing

I believe there is only one dedicated chess machine that you can slowdown the speed of its hardware and then load it with any engine and test. This could be a good test to see in how much time Stockfish 7 can see the mate in 6 at 120 MHz.

slowdown regards
Ricardo