| Player 1 |
|
#character-encoding UTF-8
#player1 JJB John J. Bulten
#player2 LC Lila Crotty
>JJB: ?DEGINS 8D DESIGNs +70 70
#note 1:11 [23:49] (dingeys 8d 70 +.7) JJB gets to play host LC and is off to a great start. Declining to quibble over the letters adjacent to the triples, he doesn't perservere to find dingeys or dingers 8d or swinged 8f.
>LC: ABNNRRT E4 BANN.R +16 16
#note 1:57 [23:03] LC makes the straightforward top static choice. Other plays for 16-18 have inferior leaves (brand, banter, barren, branner).
>JJB: EJOSTTV G7 J.VE +26 96
#note 0:21 [23:28] (jots 10b 36 +3.2) JJB's leave of OSTT is not so great (JST and ESTV are better), and so several options compete here. Jets/jots/vets 10b win about 80% of games, with jets counterintuitively on top; jest j6, objet 4d keep the S and have much the same ratio. Jive is holding a few percentage points back, and is even with jivest, because apparently the double T is a worse hold than the J or V.
>LC: BGRSTTZ 10F Z.ST +33 49
#note 1:19 [21:44] Not to mention that the previous play is welcome to LC's consonant-full rack, since she can use the S comfortably. Next-best play in static leave is to trade to RST, 10 points back.
>JJB: ACFOSTT 11B COTTAS +42 138
#note 1:42 [21:46] (softas h10 39 +1.1) JJB declines caf 29 and extends fas 38 in row 11 to cottas 42; the only thing missing from his process here is to review column H for sixes, which might turn up softas 39. The leave gives it a slight edge.
>LC: BEGPRTW F2 BREW +24 73
#note 0:52 [20:52] (twerp 12g 30 +8.3) LC's third rack with one vowel max might tempt her to hold three consonants leave, but there are plays that leave two: twerp/jivest 30, and powter 22. Twerp d11 26 is also worth consideration. Taking the extra time, stretching the vocabulary, and/or being willing to play wide open might improve points or leave here.
>JJB: EFIOTUU B11 .UIF +18 156
#note 4:10 [17:36] (outdesigns 8a 33 +10.6) JJB also misses a known vocab word in his search, despite spending 4 minutes on inferior alternatives like boeuf 18, footie 26, foe 23, etui/te 10. What better way to dump vowels than to use outdesigns from the extended out- list? But cuif, below the stratosphere of crazy play, is second-best at 10.6 behind with reasonable leave.
>LC: AAGGIPT A13 PIG +27 100
#note 0:39 [20:13] (apt 12e 37 +8.1) LC goes from consonant-heavy to duplicate-heavy. Since she holds the hooks for both za/p and jives/t, playing apt 37 is lots better (unless its leave were to be insane). But the easy use of column A often simulates better than apt 37 in the long haul (gig, pia, pig, gip all hold about 14% win chances). Decoupling both pairs with giga 13a 12 also rates close.
>JJB: EEIORTU D4 OREI.E +36 192
#note 2:21 [15:15] (outdesigns 8a 33 +.8) JJB writes ore 24 and euro 20, but then sees the play he's really interested in. While he still holds outdesigns 33, it only has a slight edge over his impressive parallel turnover play of oreide plus five. His leave TU -7.4 is still 4.8 behind EEIR -2.6. And either play gives the necessary psychological factor!
>LC: AAGOTXY J8 .TY +38 138
#note 1:01 [19:12] (yoga j10 41 +16.5) Facing strong score pressure, LC recognizes the synergy of her fresh Y with this board. A little more work going downward from zesty instead of upward would suggest yoga, 3 points better but 13.5 better in leave; soya, soy, and say are also better alternatives. At any rate she knows she can cash the X for at least 30 later (ax/ta/zax).
>JJB: ?EKORTU 2C OUT.REaK +73 265
#note 1:31 [13:44] JJB's out- list finally pays off with two options that score equally but don't double. Outdesigns still retains the blank and leads against tubework 65.
>LC: AAGORTX 12E AX +30 168
#note 2:14 [16:58] (gox 3a 40 +8.3; ax 3b 36 +6) LC opts for the safe points at 12e, even though there are five plays for 36-40 in row 3 and the openness works in her favor by a couple percentage points of simulation.
>JJB: AEINRVY H1 V.RY +42 307
#note 0:53 [12:51] A near-bingo rack that doesn't sacrifice too much to seal up a triple decisively.
>LC: AGOOQRT 1A QAT +38 206
#note 0:29 [16:29] Simple decision over the likes of trog/rax 30, pirog/tag 20, and agorot/ka 20.
>JJB: AEEFINU J6 FU... +18 325
#note 3:05 [9:46] (ef c13 29 +7.4) Now JJB overfocuses on removing the two clunkiest tiles, without realizing the superior 29 from ef/pie/eff in an obvious hotspot; he could also play enuf/stye 20 for equivalent leave to his own.
>LC: DGLNOOR D11 .ROD +30 236
#note 0:21 [16:08] Despite the suboptimal racks, LC continues to spot optimal plays at 30+. She might even get first crack at the hook.
>JJB: AAEEILN 15D ELAIN +25 350
#note 2:16 [7:30] (realia 3h 22 +3.7) JJB seizes the hook, realizing he needs to break up his rack regardless. The better leave could be had with realia 22, making two side words, and win odds are almost equal between them and a third showy play, innate i6 19 with three side words (all around 97%). JJB sees linnet i5, but doesn't convert that into the vowel dump innate. And fishing doesn't help the leader at all here.
>LC: EEGLMNO K5 GEM +21 257
#note 1:45 [14:23] (demo 14d 27 +10.8; gem 3a 22 +1) As the game winds down LC's need mathematically is to take the highest win chances if simulation makes sense logically. By this criterion, me 32b 14 directs an E hook to herself and maintains a useful leave, winning about 2.5% of simulations; this holds even with the openest play in column K, oe/fe 9, and also demo 14d, the static value leader. Gem k5 wins less than 1%.
>JJB: AAEIOUW 1J OWIE +14 364
#note 2:42 [4:48] (owie l3 23 +9) JJB now has six out of seven "vowels" in English, and yowie has just been blocked. Given LC's contribution, the best play should have been clear, ahead of quai 14, and several points ahead of exchanging to E. JJB does not spot the double overlap placement and is several points behind trading.
>LC: EIILNOO L3 LOIN +17 274
#note 0:38 [13:45] (looie 14f 17 +11.9) LC's choice still has .5% win odds, better than Quackle's static choice of looie 17. But the win leaders appear to be direct setups: noo 2m, or oi 14g (allowing prime bingos beginning ma- or pa-, among others); she needs both a bingo and additional sterling play from middling letters. Trading all is worth considering because it gives valuable turnover and partial knowledge of opponent's rack.
>JJB: AAEHOSU 14F HAO +31 395
#note 0:50 [3:58] JJB has no trouble claiming the thirty with his 4-pointer, while leaving a bingo line open.
>LC: EHIIMNO 13H HOMINE +30 304
#note 3:29 [10:16] LC makes a gutsy phony, rating 2.4 ahead of homie 17 in the same spot. Of course the singular of homines is instead homo. If it stands, this will keep nearly 1% win odds, but hie 2m now wins closer to 4% (hoe, 2%) because of the nonuple lane and the improved rack and bag situation.
>JJB: AACEESU N10 CEASE +38 433
#note 2:17 [1:41] (challenge, sauce n1 22 +38.9) Mathematically, JJB needs to challenge and then to make a play that wins all cases and maximizes spread: that appears to be eau 13h 7, because knowing most of LC's rack he can hope to bingo in column M or N with the case S. This would be worth 30 for the challenge and about 33 spread over all cases. For simplicity, the calculation used here is off static spread, which is 30+27.9 for sauce/owies after a challenge (but in simulation only about 15). JJB's choice is the highest play without a challenge (tied with cause 38), leading to padi 39, qua 12+8, net of 19; though he can also take sauce/owies 22, padi 25, ae 15+8, net of 20. Thus available points are set at 57.9-19=38.9. But the correct play seems to be challenge, eau, and then a bingo.
>LC: ADDILLP 15L PADI +39 343
#note 0:34 [9:42] Given JJB's failure to challenge, the next two plays are obvious.
>JJB: AU A1 .UA +12 445
#note 0:42 [0:59]
>JJB: (DLL) +8 453
#note LC didn't get a chance to bingo, so after awhile her hopes were only marginal and her game technically unwinnable; but there were several places she could have improved leave, score, or both so as to get closer to that goal. JJB played well, though missed outdesigns 33 three turns in a row, and neglected to challenge. So both players had chance to affect spread more seriously in their direction, even if the blanks going to JJB at the right times was a strong factor. Known points available: LC 7, JJB 9. Overall points available: LC 63.9, JJB 75.4.
|
