The scientific approach to the game of backgammon. Part 4.

But let's say, to finish this topic, about the difference between Soviet and European backgammon. I correspond with different players, including successful in international competitions. And I had the impression that in recent years the situation in the same programs, exposed on the Internet, a little different than it was 3-4 years ago. There are basically exposed the program, where the game of backgammon for money is well developed. Well and, accordingly, sites that played the smartest people in the world (I am not kidding, so you can really define intelligence) - these sites are gradually closed, and there are sites where you can play a game of backgammon for money. There, too, there is the doubling cube, then there is the game of backgammon, backgammon related to the European rather than the Soviet backgammon ... AG Still, returning to your program where you have the confidence that it can become a champion? BM Yes, this idea should be to develop a more precise. We are in our Soviet version of backgammon beat everyone who was with us. Moreover, I would send to everyone and continue to send a demo version of the game of backgammon, which is played by the so-called "fish-Geli." And we beat, and beat a couple more programs. That is, we take a consistently greater than 55% points. If the same method is applicable to the European Backgammon, that is, add the cube, you win and everyone left, we just have not had time to do it. AG So. Now - what are the methods? BM Yes. But let's start from the other end: what is built on absolutely all the other gaming program, with very few exceptions, except maybe the very first program for the game? There was a Berliner ... Maybe you can tell about a Berliner? Alex Radionov: In any programs featured such a thing as an assessment of the position in the game, some estimators. What it is? At the end of the party already clearly see who won, who lost - on board, we can say that yes, indeed, a certain number of points won one player, and another, respectively, more. This can be seen at the end of the game. How to evaluate the position where we are still left in the game did not get? Here, as a rule, the program simulates the moves of opponents with a view to one side sought to increase their winnings, and the other side has sought to reduce the enemy win. That's minimax method, minimizing and maximizing goes from here. BM But this standard. It has no relation to nedeterminizmu.A.R. Yes. Here on the calculation of such alternations minimum and maximum score is obtained positions that have not yet finite, where it is not clear who and what won positions at some intermediate level, somewhere in the middle of the game. Thus the program can evaluate your situation and make the move, which guarantees her a prize or guarantee some minimum loss, ie, does not worsen the situation.