| Player 1 |
|
#player1 Joel_Sherman Joel Sherman
#player2 Bradley_Robbins Bradley Robbins
>Joel_Sherman: AEEMOOO -AEEMOOO +0 0
>Bradley_Robbins: AEILSTU -U +0 0
>Joel_Sherman: DEEJMOW 8D JOWED +48 48
>Bradley_Robbins: AEILSST E6 IS.LATES +66 66
>Joel_Sherman: ?CEMPRU H1 CRUMPlE. +101 149
>Bradley_Robbins: DEINTY 13C DE.TINY +32 98
>Joel_Sherman: ADEORTT 3G O.TRATED +72 221
>Bradley_Robbins: AX F8 .AX +56 154
>Joel_Sherman: HINORTU 1H .OTHURNI +92 313
>Bradley_Robbins: CEEP H11 PE.CE +30 184
>Joel_Sherman: AIORSTV 15A TRAVOIS. +95 408
>Bradley_Robbins: OY 14H .OY +21 205
>Joel_Sherman: BGIILSZ D3 GLIB +18 426
#note 17,800 iterations at 2 plies valued BEZIL 45.8, GLIB 2nd at 45.6, and GIB D4 3rd at 44.6. [At 4 plies, GLIB beat GIB by half a point, with BEZIL 2 more points back.] GIB does well because holding the L has the extra threat of LEZ, but when that plan is successful, it still scores about 5 actual points fewer than when I draw from the FFW; and when I don't draw the E or LEZ gets randomly blocked, the L is actually in my way, and it reduces chance of drawing N for ZINE/ZAX backup, and of the possible ZAIRE 12a. I discarded BEZIL quickly because it's antithematic to keep the 1st G with less than half the bag to draw; and it also feeds opp easy points for a Q drop that are not yet on the board elsewhere. BIZ 6D just left too clunky a leave with very few remaining scoring threats for next turn.
>Bradley_Robbins: UU 4C U.U +3 208
>Joel_Sherman: FGHINSZ 7G F.Z +38 464
#note If leave utility is a difficult concept for you to grasp the value of, this is a great position to drive the point home. Sim says FEZ has an equity value 7.5 pts
higher than ZINE/ZAX. That means I recover 20.5 points for the 13 immediate points I passed up, by keeping SHING instead of FGHS.
>Bradley_Robbins: BLO 12B LOB. +23 231
>Joel_Sherman: AGHIKNS A6 SHAKING +111 575
>Bradley_Robbins: OO -OO +0 231
>Joel_Sherman: ?DEFGIN B9 IF +32 607
#note This position demonstrates the weakness of Quackle's simulation routine. Championship Player, which runs some number of sim iterations before presenting its top choices, finds DIF B8 with no difficulty and rates it massively higher than anything else, yet totally ignores IF B9. Amazingly, CP does include IF 4L, and IF 10a in its recommendations, for 13 and 19 fewer raw pts respectively. Add IF B9 to the list, and it makes a static evaluation that is way off base, and run a sim of just those 2 choices for any of 2, 4 or 6 plies, and it has DIF with about 2 pts higher equity value than IF. But any experienced player will look at the 21 unseen tiles containing 12 vowels and 9 consonants and a U-less Q and immediately know the D is far too valuable to spend for only 5 extra raw points. You don't need a sim to know this if you're any good at this game, and a sim that claims to disprove it is doing something very cockeyed.
>Bradley_Robbins: AA N1 .A.A +10 241
>Joel_Sherman: ?DEEGNV O4 aVENGED +84 691
>Bradley_Robbins: EM G7 ..ME +30 271
>Joel_Sherman: EIILNRW 6N W. +13 704
>Bradley_Robbins: AENOOQR 11H .ORNO +14 285
>Joel_Sherman: AEIILNR N9 AIRLINE +75 779
>Joel_Sherman: (AEQ) +24 803
|
