Back to Consensus Games

From: alan d frank <alf@world.std.com>
Date: Fri, 9 Jan 1998 13:38:55 -0500 (EST)
Subject: Re: consensus game #1
Message-id: <Pine.SGI.3.95.980109125815.6294A-100000@world.std.com>

Joel (quoted below) complained about my calculation of EDIFY+ING probabilities ignoring the opponent's better chance of getting ING on future rounds and said that a sim would reveal that.

I agree that my calculation ignored future turns. If the opponent went all-out to get ING, it would take 5.5 turns on average; starting with HO, it would take us 5.75 turns. Of course, the opponent has a tempo advantage. In reality, turnover would not be perfect and it would take both players longer to get ING.

I disagree that a sim would show the truth here. My understanding of Maven is that it with a rack like AADGLMN would play MANGY as easily as MADLY, ignoring the chance of drawing an I. (Disclaimer: I haven't thought about whether there's an even better play with the above rack.)

On Fri, 9 Jan 1998, Alex Delvecchio or an impersonator wrote:

>From the desk of G.I. Joelgernaut:
Just wanted to say I liked what Alan Stern had to say about instinct for sound moves and strategy, and add that I believe those who tried to calculate probability of opponent having the -ING may have missed a key factor. Those calcs were based on opponent having -ING immediately, but we aren't holding any of those three letters when we play de/ed-ify, which means chances are good that opponent will have at least two turns to beat us to that, and opp. will also be likely to improve his chances of collecting all three first by holding any of those three in his first rack, so if your calc stops after the first move, you're missing the boat. Sims will demonstrate this effect. Even if it sits open three or four moves, 48 points is almost a bingo, not your average play, so the danger is great enough not to offer it. In some situations, I'd probably even eschew a bingo in favor of taking that 48! G.I. Joelgernaut