return to indexMike Corbett of the USA, Backgammon Giant #30, the author of the new book Backgammon Problems and has written the following article for GammOnLine readers.
Backgammon Giant Mike Corbett who can't distinguish a microwave oven from a radar range discusses backgammon problems, his field of expertise, and about the differences between a human backgammon player and backgammon software with a computer he insists on calling Hal.
Don't think less of Mike because he talks to inanimate objects. Mike insists he had this exchange with his computer. (He does not differentiate between computer & backgammon software). I can't guess why he calls it Hal.
Danielle Bastarache
Backgammon Software: Good Morning Mike
MC: Hi Hal
Backgammon Software/Hal: Were those 3024 rollouts helpful?
MC: As a matter of fact, I won a bet thanks to you.
Hal: How so?
MC: Well, I bet a friend I could find a backgammon position in which your evaluation would be contradicted by a rollout for each of the first seven possibilities.
Hal: Seems pretty unlikely, Mike. I am after all, highly sophisticated and more computationally powerful than a Holiday Inn full of humans.
MC: But aren't you really just a tool?
MC: Look Hal, I'll concede you have great potential, but you lack creativity.
Hal: What's creativity?
MC: Forget it; I don't have a conversational interface for that quality. Let's just review some of your work.
Hal: Shoot.
MC: Let's take a look at the backgammon position in this diagram.
MC: You like 6/1 5/1 because?
Hal: I made another point and retained both anchors. I love stability.
MC: The way I see it, you're overestimating the value of that anchor on the 20 point and you don't seem to be concerned about using builders for the 1 point that could be better used making the 4 point. At least the blot on the 14 point didn't alarm you.
Hal: Well, I settle that issue in plays 2 through 5.
MC: These plays sort of half accomplish a number of goals, none of which includes the 4 point. And that #5 play really stinks.
Hal: I have no olfactory capacity witling.
MC: Sorry Hal, you're a little sensitive today.
Hal: So if you're so smart, how do you explain the strength of #6 and #7 in the rollout?
MC: Three words, limited risk diversification.
Hal: What do you mean limited? Watch what happens with 4-4 in this position with only one anchor.
MC: You're not even playing that right Hal!
Hal: What happened?
MC: That double gap can be lethal pal. It's trickier to bear in with your configuration. Most numbers aren't that great Hal, and the jokers are still pretty good with or without two anchors. So can you guess, sorry compute, the reason for the superiority of 20/15 14/10 to 20/11 in the original position?
Hal: I suppose you'll tell me that a builder on the 10 point is a big damn deal.
MC: I admit to my perpetual awareness of having a convenient six to play. I can also tell you the difference between a blot on the 15 point as opposed to the 14 point.
Hal: Save it for someone who cares. I'm going to roll this all out 10,000 times, so don't spend your money yet.
MC: First, the bet was for a 432 game rollout. Second, you can roll this out until the sun goes supernova and you won't replace the best plays with any of your nonsense. Third, you don't play well enough to validate a .001 difference ranking plays by rolling them out.
Hal: I'm not your bitch, Mike.
MC: No, you're a good little worker Hal, but I'm in a mood to slap something around. I'm gonna toss a fork in the radar range.