[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums

Bower's Law

Posted By: Matt Cohn-Geier
Date: Monday, 20 October 2008, at 10:14 p.m.

In Response To: Bower's Law (Bob Koca)

Well, I guess the bots are "pseudo" neural nets, in that they are programmed with some weights and METs and the like and then derive their evaluation function from training games, rather than do everything from scratch. But surely -2, -2 doesn't need to be programmed in as a special case.

Suppose that you don't double and your opponent adopts the strategy of always double. In this case it takes very few games to realize that your equity is the same whether you double or not without market losers, so therefore if market losers appear you should double to increase your equity (or, in other words, the doubling window opens at 0%).

It is more difficult to prove/understand that the opponent's strategy of "always double" is at least as good as not doubling, but if opp adopts the strategy of doubling, a neural net should be able to work this out, right?

Messages In This Thread

 

Post Response
Your Name:
Your E-Mail Address:
Subject:
Message:

If necessary, enter your password below:

Password:
 

 

[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums is maintained by Stick with WebBBS 5.12.