#player1 Chris Chris Cree
#player2 David David Gibson
#title SCRABBLE ALL-STARS 2003 Round 18 Table 1: Cree-Gibson
#description SCRABBLE ALL-STARS 2003 Round 18 Table 1: Chris Cree vs. David Gibson.  Recorded by Kristen Chew and Gregg Foster.  The winner of this round advanced to the finals.  Critical analysis by John Chew.
#style sas2003
#incomplete 
>Chris: SWEOOI? 8d WOOlIES +76 76
#note The only other bingo is WOOdIES.  Cree took only 33 seconds to make an easy choice.  While the words are almost equivalent (for example, both take a -T hook), an L is probably slightly less useful to Gibson.
>David: ROFETWL k2 FELWORT +86 86
#note Gibson not only has the hooking T, but a bingo as well.  He could have played FLOWERET or FELWORTS, but FELWORT uses WOOlIES-T to score 17-19 more points.  The other three-letter prefixes for WORT are AWL-, FAN-, FIG-, MAD-, MUG-, RAG- and RIB-.
>Chris: EEHIOOQ  -EHIOOQ +0 0
#note Cree takes only ten seconds to decide to get rid of the Q and an excess of vowels.  Had he thought longer, he might have preferred 4h HOOLIE, keeping EQ, which scores 26 points with a tolerable leave.
>David: ROTELEG 7f GEL +12 98
#note Early in a close game, with no high-scoring plays available, Gibson has a difficult choice.  He can score 10-20 points in several ways, but which one will maximize his subsequent scoring chances?  He ponders for 4:46 before playing GEL and keeping EORT.  Maven strongly prefers 7d OGLE (18) keeping only the ERT.  It's tempting to keep the O, with three of the eight already on the board, but it's certainly not worth passing up on six points.
>Chris: EDEKNNR 6b KENNED +31 107
#note Cree makes another snap decision.  In 32 seconds, he plays KENNED keeping an R on his rack.  If he had played 9g KEN instead, he would have scored two fewer points but kept the much better leave DENR.
>David: ORETFOT 4g FOOTLER +22 120
#note Gibson chooses FOOTLER over 2i EFFORT, which would have scored 18 more points but opened up 1h and 1l-1o.  Maven strongly prefers EFFORT.
>Chris: RAAIINP 5c PIA +20 127
#note Cree looks for a minute and a half, but misses the bingo 3m RIPARIAN.
>David: TIMASUN j4 TAM +26 146
#note Gibson returns the favour, spending a minute and a half to miss 3a TSUNAMI or m1 NATRIUMS.
>Chris: ARNIABD 4a DRAB +33 160
#note DRAB and BRAD score the same number of points and are equivalent in simulation.
>David: UNITSE? 2g UNSIFTEd +74 220
#note Gibson's other choice is n1 NUTSIEr for three more points, or about one more point of simulated equity.  UNSIFTEd is clearly better though, as with a 60-point, Gibson is justifiably more concerned with closing the board then scoring an extra point or three.
>Chris: IANVUUN 1e VINA +29 189
#note Cree now has a tough choice.  He can also play UNAI or UNAU at 1f or 1l for 21 or 22 points, splitting his Us but still keeping INV or NUV.  With the bag consonant heavy (29-19), I-rich (5 unseen) and of course the Q still to come NUU turns out not to be seven points worth than the others, and VINA narrowly edges them out in simulation.
>David: EEGNACL a1 CLADE +24 244
#note What should Gibson do here?  He has a 31-point lead halfway through the game, and can bingo i8 ELEGANCE for 62 points at the cost of ripping the board wide open.  Or if he wants to keep things closed by playing at A1, he could play GLADE instead for three fewer points keeping CEN.  Careful analysis of simulation data shows that Cree has about a 1/200 chance of catching up to Gibson within a turn after ELEGANCE, but only a 1/300 chance of doing so after CLADE.  The EGN leave is also better in this situation than CEN, perhaps because of the shape of the board and the number of Is left.
>Chris: UUETIBC  -UUTIBC +0 189
#note As the game progresses, Cree's choices get harder.   Should he trade keeping just the E, hang onto the synergistic CEIT, or try to score a few points and make an opening for a tile he doesn't yet have with 7j ERUCT (19)?  ERUCT gives him five more points of equity than trading, but there's at least a 40% chance that Gibson has the hooking S, and only a 20% chance that Cree will draw it.  If he trades fewer tiles, he may be more likely to bingo but Gibson will accordingly be more likely to get defensive.  Against Maven, trading three tiles would let Cree catch up 1/40 of the time, compared to 1/50 after ERUCT or 1/100 after trading one tile.  It all comes down to whether or not Cree thinks Gibson has an S.
>David: AGENTNY 1m YEN +24 268
#note YEN is clearly best.  
>Chris: EEISSZP d8 WIZES +34 223
#note WIZES is a straightforward play, but it's well worth considering n4 SIZE (44).  While it does open up the O column, the equity at risk there is only about three points, and is more than offset by the ten extra points up front.  With a more volatile board, Cree would also have a better chance of catching up (1/6 compared to about 1/9).
>David: GENTMAG 11c GEM +12 280
#note GEM is best.  It shuts down the left side of the board, keeps a great leave for when Cree tries to open it up again, and maintains a healthy lead.
>Chris: PEASCQO 10d ZAP +24 247
#note ZAP is the only choice.  While Cree might want to keep the A with the Q, there are still three each of A, T and U in the bag, so the CEOQS leave is not as bad as it could be.  
>David: ANTEGIB m4 REBATING +63 343
#note Gibson's fishing pays off.  REBATING is the only playable bingo.
>Chris: CESQOUH j8 SUQ +32 279
#note Cree's fishing pays off too, with lesser results.  He draws the U, so he can dump the Q for a few points.  He cleverly passes up on 10j QUENCH (46) points, because it's not enough to close the 96-point gap, doesn't have a good enough leave, and accelerates the end of the game by turning over too many tiles.
>David: AXALDRY i7 DEX +46 389
#note Another easy choice.  The only other place to score with the X (7c) could cost him the game by opening up the A column.
>Chris: CHOSEUI n11 OUCH +21 300
#note Cree pointed out that this was a self-descriptive play.  He still has a chance, and keeping EIS while opening up a few extra bingo lanes is his best bet.
>David: TYALRAT o8 TRAY +28 417
#note There are 19 possible 7-letter bingos that Cree might have in the O column, and only one possible 8-letter bingo through the C or the H (OVERRICH).  If Gibson blocks with TRAY, Cree has almost no chance of winning.  o8 LYART would have scored a few more points, and might have been a better play, but it makes very little difference.  All Gibson needs is the win to make it into the finals.  Playing an extra tile brings the end of the game that much closer, but might also restrict his flexibility in countering Cree's next move.
>Chris: IESJURH c3 JAPE +26 326
#note The outcome of the game is certain now.  Cree makes the highest-equity play.
>David: LATDOIV 13h VOLTAIC +26 443
#note Gibson takes no chances that he might have missed a bingo through the C or the H.
>Chris: URIESHR h10 SHIVER +39 365
#note Cree makes the best play again, but it won't be enough.
>David: DIO c3 JAPED +17 460
#note Gibson misses his outplay: e10 AMIDO (10).
>Chris: RU  6m BUR +7 372
#note Cree finishes flawlessly; Gibson moves on to the finals.
>Chris:   (OI) +4 376