Back to Consensus Games

From: Alex Delvecchio <icenine@panix.com>
Date: Sat, 17 Jan 1998 08:59:07 -0500 (EST)
Subject: Re: Consensus Game #1, Turn #2
Message-id: <Pine.SUN.3.94.980117065245.19995C-100000@panix2.panix.com>

From the desk of G.I. Joelgernaut:

We inadvertently discarded about a month's worth of incoming cgp and private posts with an errant keystroke Thursday night, so I'm grateful that Tracy posted his reply on this topic to refresh my memory of the position. And I hope my response is not too late to be counted by John. [Please, do NOT anyone of you faithful archivists try to forward us the month we lost. I had it up to here with that month!]

I shall list the moves I considered in this position with brief (for me ;-) ) reasons for dismissal in most cases, in order from least* preferred to most* preferred. * = by me, and only me! ;-) 

FEODS    9c  (33)  -LT:  The best score available and not a bad move
in a lot of cases, but here the S is worth a lot more to me than the 8 points this adds over the singular. *I expect* to bingo with that S in the next two moves. I am not capable nor willing to attempt to prove mathematically that this expectation is reasonable, and I even expect that those who are capable of working out the math will demonstrate that two moves is too soon, but (sorry to sound a little confusingly like Marlon here), my instinct for this game tells me it is a reasonable expectation. Even if the math says otherwise, I would balance that argument with the idea that the chance of having a playable bingo is so much greater with the S in hand than without it, that I will regain the 8 points with compound interest simply by comparing the times that I draw a playable non-S bingo or have to wait several non-bingo turns to draw a new S with the frequency of getting a quick bingo using the S. None of the other moves to be discussed use the S, so this point won't be mentioned again. 

DELF     9g  (23)  -OST: The optimal consonant leave in most positions,
and the optimal vowel leave in this slightly unusual specific case, but it's twice as easy for opponent to hook as DELFT, scores 1 less point, and gives up 3 points to my top choice. Minor consideration is we may recoup hook points if opp. plays the T and we pile on our S later. 

DELFT    9g  (24)   -OS:  Not as strong leave as DELF, but with 5 more
T's unseen, the additional try at other positive tiles may return the T or offset its loss with something even better. However, my instinct is toward less randomization; my experience tells me my rack is 25% more likely to get annoyingly screwed up drawing 5 than 4 even when I have a power tile in hand, again I don't care what the statisticians have to say about that, call me paranoid if you must. 

DEFTLY   e3  (26)   -OS: Same tiles as DELFT but the 2 extra points
compensates my paranoia enough to risk the extra draw. When I first glanced at this position, I was inclined to choose this move, but I took another, longer look and decided I had reasons to prefer others. 

FEOD     9c  (25)  -LTS: A very nice play, it's always hard to argue
with any move that leaves my brother's initials in rack. Score is good, opp's comebacks are limited, and openings do favor our leave. However, it would be nicer if LTS were accompanied by at least one vowel. My inclination is to avoid the unbalanced leave, and with 3 other plays scoring 1 point more while retaining a vowel, I have to relegate this to "good" third choice.

Here I have to interject a reminder about the depleted pool vs. generic optimal leave idea. In most cases, it would be a clear choice to keep an E before an O, and in many cases I have a personal aversion to O's that would cause me to hold any other vowel ahead of it, but not here. One should NEVER make holding the FIRST E a priority consideration in choosing a play unless it is in company of 3 or 4 of the consonants in CANISTER.

The first law of Scrabble Entropy is that the more there are of any given letter in the pool, the more likely you are to draw an unfair number of that given letter in one rack, and that goes double for all vowels. The second law is that whenever there are several of more than one letter in the pool, you will draw an unfair number of the letter of which fewer are available; and the corollary to the second law is that the second law maximizes the chance that upon playing off a bunch of bad tiles, you will draw an unfair number of whatever multiple tiles you left in pool with your previous rotten draw.

In this particular position, I would prioritize in the following manner with respect to leave: 1st choice, any move that plays the E and holds the O; 2nd choice, any move that uses both vowels and lots of the consonants; absolutely last choice, any move that plays the O and holds the E. And to anyone who wonders if I actually follow these ideas in tourney play, the answer is YES, just watch me. 

DEFT     9g  (26)  -LOS: Gatos and Angeles love this leave, ordinarily
I would hate it. The first L is usually a good thing to get rid of fast, and we already discussed my aversion to O's. But with the Y left open, the L is suddenly not that bad to hold, and the fear of duplicating it is also well compensated by 9 unseen A's. The S-hook is now spread from BOG to BID, and the E we expect to redraw provides a potential dual overlap in same place. DEFT opens the o8 tw, but doesn't give away a hook or play-thru for the k5 ddw, and if opp hits o8, we get 3 more tiles to play 8's thru. Overall this is an excellent move and I'd expect it to rate a tossup or even slightly higher in simulation with the next one, but I have personal style reasons to express a very slight preference, and therefore vote for: 

FLYTED   e6  (26)  -OS: I like an open board and the extension on
both sides of row 8 aids expansion in every direction -- DEFTLY with the same tiles and score doesn't provide the opening of the southwest quadrant so is not as useful to my style of play, and also yields too many points to parallel plays which quickly close off the open DEF. The score and leave are equal or nearly equal to most of the other useful choices. The opp replies which may score slightly more against this move than the others will likely be vertical on columns d or f and provide even more play-thru tiles and access to more lucrative premium square lines for our anticipated quick bingo. Worst case scenario, opp plays XENIC f10 giving only two play-thrus and a bad 8-starting N. But if you consider this an important factor, you are too paranoid to play this game well.

In closing I'd like to note that I'm glad opp chose BOG over GOB 7c (20), removing the temptation to play the Widergren-and-Zaxlike FELTOIDS g3 against any opp not from the Bay Area or on cgp. Good luck to all in Reno this weekend.

G.I. Joelgeranalyst