Quote:
Originally Posted by Taper_Mike
His concept is particularly useful when you want to analyze a single exchange. That’s because it is not overly difficult to set up all 21 rolls an opponent might have after you make a given play, and roll out each one. Usually this exercise will reveal that certain rolls for the opponent are more likely to generate misplays than others. You can then form a weighted sum of the equity your opponent will lose if makes those misplays.
Now you have a useful number. By comparing the equity you sacrifice when you make a deliberate misplay with the equity you gain via your opponent’s response errors, you can determine whether “creating complexity” is a good idea.
I think it will not differ much from a 4-ply evaluation. But whatever, the argument is again that the bot is no human representative. In one way or the other the sumtotal of the difficulty degree of all next moves of the two players should be compared, and this outcome should be included in the evaluation, which looks not so feasible to me. And then still it could be that you could take on a much greater degree of difficulty than your opponent. I am less a theoritician, and more a practicalist, and I know what I am talking about. It's not often that I play live games, but the times that I do I am pretty successful. Ask a dutch topplayer and the chance is great that they will know me.
Quote:
Originally Posted by Taper_Mike
Against the competition Rockwell usually faces, David estimates that he will break even or gain when his sacrifice is around 0.01 or so. He won’t try one of these “complexity” gambits when his initial sacrifice is much more than that. 0.015 is definitely too high for him. If your opponent’s are of a lesser caliber than the open-level tournament players David competes against, you may well observe that your threshold is higher. Whether it reaches 0.03 is something each of us will have to decide for ourselves.
It could be that he is a little bit apprehensive about his PR (pun intended). The performance rating reminds me of the IQ discussion of years ago, in which it was contended that IQ also shows to what extent someone is adapted (read: slave) to the culture or the system. If you ask me, performance rating has some connotations with a
drycleaner.
Quote:
Originally Posted by Taper_Mike
My own practice is very conservative. I am probably ranked among the lower half of open-level players, so you can imagine how difficult it is for me just to find the right play in a given position.
You have written a backgammon ebook. Could you give the link, please?
Quote:
Originally Posted by Taper_Mike
Usually that is hard enough for me. Since I don’t even know what the best play is, I don’t try to find the next best play. Even if I wanted to, how could I be sure I was right? And if I found the second-best play, how good would I have to be to correctly estimate that it only sacrificed a small amount of equity?
That's because you are f*cking ants (dutch colloquial). Again, the certainty offered by the bot is an illusion at this stage. Backgammon is about risk taking, and your approach is like a chess player. It is the difference between the street fight and the ring fight. I don't advocate anarchy, but when you vary your approach it will be an enrichment to your play, and maybe also to your personal psychological evolution. And I am not even American!