M ANSARI wrote:Ofcourse black back rank is to weak to go pawn hunting, but if rooks are not exchanged then black gets even better chances to win than before and at least has an easy draw. Black is totally out of danger because bishop on a6 controls important squares that would allow the white Knight to be dangerous.
This all was seen by Rybka 3 Monte Carlo analysis. It is perfect for these type of positions because it generates thousands of very quick games trying different move orders and going back to where one side won and seeing if a different move would change results. After I put in Bc8 and let it sit on it for a while it became pretty clear that Bc8! draws. This is the feature in engines that will remove the last weakness engines have ... and that is the horizon effect.
When engines can do this in _Real Time_ in a 40/2 game OTB then chess will be effectively dead. This will happen within the next ten years I imagine.
Capablanca announced this about 100 years prematurely.
I am not sure about dead, but it would definetely change how engine chess is played. Hardware will continue to improve at a breakneck speed, and with more and more cores becoming available, you are correct in assuming that some cores will be dedicated to exactly such Monte Carlo analysis in real time. I know Vas has it in mind for Rybka. This feature is amazingly useful in positions that need a lot of manoeuvering to avoid a blocked fortress ... something that engines used to play horrendously.
This feature is also very powerful when looking for a forced win in positions with few pieces remaining on the board where you do not need to have say an 8 piece EGTB ... you create your own relevant database very quickly to find the forced win.
M ANSARI wrote:I am not sure about dead, but it would definetely change how engine chess is played. Hardware will continue to improve at a breakneck speed, and with more and more cores becoming available, you are correct in assuming that some cores will be dedicated to exactly such Monte Carlo analysis in real time. I know Vas has it in mind for Rybka. This feature is amazingly useful in positions that need a lot of manoeuvering to avoid a blocked fortress ... something that engines used to play horrendously.
This feature is also very powerful when looking for a forced win in positions with few pieces remaining on the board where you do not need to have say an 8 piece EGTB ... you create your own relevant database very quickly to find the forced win.
Agreed Majd,my first impression of the Monte Carlo feature was negative,but with time I realized that it's actualy a powerful tool....
I can only imagine a future Rybka running on a 16 cores system and 4 of them dedicated for the OTB Monte Carlo calculations....wow amazing even when you think about it....
BTW,I find the Monte Carlo very useful also when analayzing gambit opening lines and finding novelties....
_No one can hit as hard as life.But it ain’t about how hard you can hit.It’s about how hard you can get hit and keep moving forward.How much you can take and keep moving forward….
Yes Wael it is very powerful ... especially on Octa. With 8 processors you have 8 different engines running simultaneously checking every little nuance in the position. Ofcourse you can also play with the depth, but usually I leave it at the default depth of 5, which results in thousands of games very quickly. I still don't know how to use it to its maximum effect, but already I have tested in positions which have long been problematic for engines ... such as blocked positions with long term manouevering ... and it is extremely effective in those positions.
As for using it real time, this is probably another 3 or 5 years away before it can be used effectively. But already today with top hardware it can be used in a few instances even in real time.
M ANSARI wrote:Ofcourse black back rank is to weak to go pawn hunting, but if rooks are not exchanged then black gets even better chances to win than before and at least has an easy draw. Black is totally out of danger because bishop on a6 controls important squares that would allow the white Knight to be dangerous.
Blacks may draw this, but they are in the defensive.
This all was seen by Rybka 3 Monte Carlo analysis. It is perfect for these type of positions because it generates thousands of very quick games trying different move orders and going back to where one side won and seeing if a different move would change results. After I put in Bc8 and let it sit on it for a while it became pretty clear that Bc8! draws. This is the feature in engines that will remove the last weakness engines have ... and that is the horizon effect.
Why is this called Montecarlo? Where is the random part?
This would be a good experiment (a scientific positive control) to test its efficacy:
Try it in this position After Rb2 Qd7/ Qxd7 Bxd7/ Rb7 Be8 (not Bc8). White wins after this. It would be interesting to know whether this Montecarlo method can pick up the winning method.
M ANSARI wrote:Here is a good explanation to how Monte Carlo method works. I also don't understand why it is called Monte Carlo ... why not Las Vegas ???
I know what the Monte Carlo method is, but I do not know why this Rybka method fall into the category of a Monte Carlo simulation. What you describe does not sound completely random.
