| Player 1 |
|
#player1 JJB John J. Bulten
#player2 EL Ed Liebfried
>JJB: AEEMOOT 8G MOO +10 10
#note 2:00 [23:00] EL is one game behind JJB and can win the tournament with two really good showings; meanwhile, Pat Krohn is tied with EL and is currently playing the lowest seed, so all three top seeds are still competitive. JJB draws too many vowels for a start but perversely does not want EL to open the game, so does not pass (AEMT is a fine +9.2); further, he fails to locate an obvious play, emote 8d 16, and the next-best word is only moo, evaluated at 1.9 behind (AEET is -2.7).
>EL: ?AAERTT J6 TARTArE +66 66
#note 1:43 [23:17] EL is overwhelmed with bingos given the wide opening and chalks up the first one. Best is to play to the O, if one can spot amaretto 74 (8 points ahead). Tartare is subtle because it takes no S, although it scores more at 9c.
>JJB: AAEEOOT K4 TOEA +12 22
#note 1:30 [21:30] JJB quickly declines to hold and prepares the business of overcoming a deficit. Again he reasons that trading will not help, although ET (+1.9) is slightly ahead of his play leaving AEO (-10.6). Dumping one more vowel with a 6 would be even better, though, and he neglects a second well-known word, aortae 10 (2.3 better valuation).
>EL: ADGILNS L7 LADINGS +73 139
#note 1:21 [21:56] EL has a pretty clear scoring location, and the obscurer anagram, ligands, would score the same without significant benefit. However, both words actually create 4-overlaps if played at l1, scoring 10 higher each; potential parallels should always be noted.
>JJB: AEEHOOW L1 HOWE +30 52
#note 1:12 [20:18] JJB holds for a minute but finds no justification for a challenge despite his immediate need to make up 117 points. Relieved to have better pointers with his high-voweled racks, he contrives the highest score (though miscalculating by 2; whoa also gets 32, but with even worse leave), and he is certainly happy to hope to use the nonuple. Wahoo 21 is just not worth it. Honorable mention: erenow 16.
>EL: ACEELRS 7B RELACES +68 207
#note 3:25 [18:31] EL looks a long time at the 4 words on this rack and puts together the combination "leachers", but finally admits on turn that he does not believe in the triple-triple enough and lays down relaces/em/so. JJB holds the play (three consecutive bingos are a pelican) but knows after reflection that relaces is not invented. He also sees leachers 167 immediately, and like EL doubts that it's any good: to use an apposite expression, it's dubious. However, EL actually has opportunity cost of 99 points by not firing the gun. His risk is low: even if JJB overcomes his vowel problems and plays the nonuple himself at minimal odds, EL still has plenty of bingo replies in the 70 range and has good chance of holding onto the lead. That means the reward is virtually sealing up the game, the risk is a couple of percentage points' chance that the game is tied, a few more points' chance that the lead is about 90, and a lead of about 200 in most other cases; on the other hand, the sure thing is a lead of 155. As a very rough calculation, the mathematical expectation justifies taking the risk: it probably adds about 50 points about 80% of the time and backfires by about 50 points only about 15% of the time. (And EL could theoretically have much more: if he plays leachers and draws the same as in the real game while JJB challenges as in accord with his snap judgment, EL supplements his 306 score by bingoing a fourth time with cuboids 76, and the score is a whopping 382-52!) Yet one cannot seriously fault the playing style that ignores expectation and takes the safe and defensible lead.
>JJB: AAEEEOR M1 AREAE +30 82
#note 0:56 [19:22] JJB releases the hold quickly and spots the parallel play quickly too, reasoning that time will be necessary later and that removing the bingo likelihood from the nonuple is justifiable. In theory one should take enough time to see areolae (and rule it out on low points), but with a big deficit the turn time needs to be saved for more decisive racks.
>EL: ?BDIOSU 9B DuBIOUS +69 276
#note 3:44 [14:47] EL knows early that he has the fourth bingo (an albatross), but keeps looking because of the luxury of the broadly open board and rack. It would take sharp focus indeed to find bodysuit on this rack and then slot the 3-overlap at 9c for 77. One might say by contrast his play is dubious.
>JJB: ELOSYYZ 1L ..ZY +57 139
#note 0:38 [18:44] JJB has had to keep upping his demands on his own play; the winning chances keep taking hits, and EL might be holding a zyzzyva (a fifth consecutive bingo); but an expert can still make up the 194 points at this stage given above-average tiles. He finally has a rack that plays into this storyline, yielding an easily indicated bingo equivalent with convertible positive leave, so he again plays quickly to put more time pressure on EL (who is making good work of recovering from his last tournament).
>EL: IKNR D9 .RINK +22 298
#note 2:50 [11:57] EL comes back from the brink of scoring the fifth bingo. On partial leave, ink 10d 39 rates 18.1 better due to the prime positioning of ok, which still takes some getting used to after a year of availability.
>JJB: EELOSTY 13B YO.EL +24 163
#note 3:46 [14:58] So far JJB has done almost all that could be done with his racks, dropping only 4.2 points behind optimal play (not counting a scoring error). Now the decisions get tougher. He correctly rules out oystered 28 and the more competitive ye 19. He sees the best play, zee 36 (zees is behind), but does not evaluate the gap between LOSTY (+8.7) and EST (+14.6) as being lower than 12 (a diff of 6.1). Yoke and yok are also better because adding more tiles to a synergy often improves the leave (here the L does so but not the second E).
>EL: IMV 8A VIM +32 330
#note 0:41 [11:16] EL punches back with tiles that perfectly fit the triple slot he created.
>JJB: EGJNQST N1 .EN +36 199
#note 4:10 [10:48] JJB spends more analysis time here and decides it's time to use column N even though it has triple risk and (he remembers after a lapse) zens* will not help. Quackle suggests, in order, using J (jig 11c 22, which stays on JJB's mind but just doesn't look right here, even as it rates 2.7 ahead of zen), using S (gens 14a 34), or using QS (qis 11c 24); but using none of them really cramps the leave. In this case, hoarding leads to too much imbalance, so something on the rack must be cashed for less than one might hope in other situations. While opponent might always book the Z himself, given zen he might always book the 6-letter triple, so there's not as much gain as there seems.
>EL: DEGORU O3 DROGUE +35 365
#note 1:36 [9:40] EL does indeed hold the 6 to the triple, keeps his lead even, and is even recovering from time deficit.
>JJB: GIJQRST 6A QI +20 219
#note 0:28 [10:20] Now comes JJB's most crucial omission. He should have already analyzed that the existence of aji permits a big column K score, and he now holds it, namely jig k10 41 (22.9 better than his play), well ahead of jig 11c 22 in second place. The leave is not perfect but very Q-conducive. But JJB is recoiling from the punching and misapplying fear to the Q, and also declining qis 24 or qis/yokels 27. It turns out he does later draw the V hook as faintly hoped, but a weak bingo lane is no valid goal here.
>EL: AFFN I10 NAFF +22 387
#note 3:19 [6:21] EL considers several options here, but invalidating any column K score while balancing his rack gets a good return of 22.
>JJB: GJRSTTU 11C J.G +22 241
#note 0:33 [9:47] Now that JJB's attention is focused by consonantal necessity upon jig, it is no longer optimal: juga e4 24 rates .9 higher, and ajuga is a much more hopeful hook than virid.
>EL: IPX N8 PIX +32 419
#note 1:08 [5:13] EL's task is made easier by cashing yet another well-placed premium, making up for JJB's use of the other high tiles.
>JJB: ERSTTTU O10 UTTERS +30 271
#note 1:08 [8:39] Static evaluation suggests fishing with ut o10 11 (4.4 diff) or tore 2k 4, but this neglects the need to open lines. Amazingly, JJB still has a win chance of about .1% with a few simulated plays, and his intuition is correct that opening row 15 (with row 5 still a long shot) might theoretically work, though easily blockable. There are also 10s available, like tetraspore after ut. Simulation actually vindicates JJB's choice, putting utters on top most of the time; sutta e3 allows almost comparable openings, and exchanging to ERS allows the V hook or the 10 (e.g., underscore) to materialize on occasion. Utter/xi is a nice directed hook but is too easily blocked to win much.
>EL: CDU 15L CUD. +10 429
#note 2:14 [2:59] Blocking is indicated; cur 14m 10 rates 2.4 higher, and seems to put a little more pressure on JJB by leaving one extra bag tile and seizing 14n.
>JJB: ABEHPTV 5B VAPE +27 298
#note 6:46 [1:53] JJB now takes most all the analysis time on leftover points, having drawn one of the many sure losses. After some time he notes that there is a new V four on the rack despite using all his vowels, then that if he picks O then both/be plays at e4, then that both/ba/op/te plays even better. Despite being at a loss if a consonant is drawn and not being too happy if the likeliest (I) comes out, sometimes having one chance in eight of doing really well is enough justification. Certainly cashing 14a, as many top static plays do, leaves opponent way too easy a response, and JJB correctly avoids this trap that he has fallen for often before, so there's that additional justification that no safe plays are anywhere close to vape and nothing else reaches 5e.
>EL: IIILNNW B12 W.NN +20 449
#note ~0:59 [~2:00] EL notes immediately that JJB has one great out and that (whatever JJB drew) he got it at 1-in-8 odds, and jokes, "Well, that worked out very well for you!" He opts for the best points for himself (20) and allows the elegant out (32+8, net -20). Blocks that do not allow other outs start with wile 7, leading to holt 27, inn/if 14, bi/bop 11+2, net of -19 to him. But nip d3 10 keeps the W for higher score, leading to bit/ba/te 11 (threatening oh 25), wyn 18, hod 23+6, net of -12. And nil/ne 5 forces even less synergy on opponent, leading to yob b13 16 (threatening eth), win 17, oh 13, ef 5+2, net of 0: nil nets nil. (To plays besides yob, EL can play off IN for points and threaten to go out with IW.) So it appears that nil has the best endgame continuation, which can be arrived at by using as rules of thumb (1) block completely first and cash W later, (2) don't allow a 1-voweled opponent an easy double word or consonant dump either, and (3) cast the remaining leave as two more bites the most easily (IINW).
>JJB: BHOT 4C BOTH +32 330
#note 0:22 [1:31] With one tile in the bag JJB has netted 47, an excellent endgame ploy that would win a closer game. EL did have the option and the time to block, but he almost prefers to allow the elegant outplay, and says to JJB's agreement that at least the tiles went in JJB's favor at the end.
>JJB: (IIIL) +8 338
#note The game was dominated by EL's four opening bingos, a feat which rapidly made news around the tourney room, while JJB drew only 4 consonants in his first 18 tiles. However, EL doubted a nonuple and had strong bingos available at other times, while JJB made good use of power tiles and got a bingo equivalent, his only significant miss being overlapping the J into aji. JJB was compelled both by bad racks and high deficit to play more accurately, and caught up 83 of the 194 deficit without a bingo, even as EL kept scoring premiums himself; JJB even "swindled" another 20 points out of the endgame with sheer grit and a 1/8 draw. The top three seeds are now at 4-1 with one game to go. Known points available: JJB 2, EL 145. Overall points available: JJB 41.2, EL 165.5+.
|
