| Player 1 |
|
#character-encoding UTF-8
#player1 David_Weisberg David Weisberg
#player2 Randy_Hersom Randy Hersom
>David_Weisberg: EEIKLNR 8G RELINK +30 30
#note Sims best. ELK and LEK are a few points behind
>Randy_Hersom: AEHIV K4 VAHI.E +24 24
>David_Weisberg: ABEEOUU G6 BU.EAU +10 40
#note Sims best, with exchanging to E about 2 points behind
>Randy_Hersom: EHILO H11 HELIO +32 56
>David_Weisberg: AEEGORR 4H GRO.E +22 62
>Randy_Hersom: ?ABDOOS 13C DiABO.OS +74 130
#note It played at 14G for 82, but Quackle also found 6G BABYHOODS 70!
>David_Weisberg: AEELNPR 15E PER.NEAL +61 123
#note The only bingo
>Randy_Hersom: AA F5 AA +10 140
>David_Weisberg: CEIRUXZ E3 CUZ +50 173
>Randy_Hersom: GIN 8G ......ING +42 182
#note Scored as 39
>David_Weisberg: EEIMRSX 12C EX +37 210
#note Sims 3.5 ahead of 12B MIX 43
>Randy_Hersom: AAEI 3C AE.IA +16 198
>David_Weisberg: EIMORST C3 .MORTISE +62 272
#note I thought about this for a long time. I missed ATOMISER which I would have preferred, and it sims about a point ahead of AMORTISE. I was nervous about what B6 MOISTER 77 gave back, with the Q unseen and it opening a big new bingo line, but 15 points is a ton to give up to keep it more closed when I was only slghtly ahead before the bingo, and other lines are still open. MOISTER sims nearly 10 points ahead of AMORTISE and 9 ahead of ATOMISER.
>Randy_Hersom: TTW 8A TW.T +24 222
>David_Weisberg: DEFJPTW 10B J.FE +30 302
#note I thought about B10 FED or WED 39, but the J looked really bad on this board and I wanted to shed it quickly. FED sims best at half a point better than WED. JEFE is 9 points behind FED. I'm not sure how much I thought about 10B JET 26, setting up E9 FEW 29 with only one more E unseen, but it sims a point behind JEFE. I'm really puzzled by this result because I just don't see much I can do with the J after FED or WED, but the 9 points must be making the difference.
>Randy_Hersom: MY 2F MY +28 250
>David_Weisberg: DIPTVWY 1G PITY +35 337
#note The computer prefers 6J WHITY 30, but I don't know if Randy would have played MY if he weren't holding an R or blank. WHITY has the merit of blocking a lot more bingos in general, with Randy only bingoing about 16% after WHITY vs. 30% after PITY. PITY sims 4 points behind WHITY, and about evenly with another defensive option, N6 WINDY 28. Because of possible inference considerations I'll take this one with a grain of salt, but I can see the case for WHITY.
>Randy_Hersom: FRU L12 FUR. +14 264
#note It seems I might have been right about the R, at least.
>David_Weisberg: ADGNNVW M5 WAV.NG +18 355
#note I jumped on this but missed M4 DAWNING 25. It has the advantage of scoring more and turning over more to go for the blank and hasten the end of the game, but keeping V isn't good. The Q may or may not be a concern since it plays by itself at 7J for 22, but if Randy gets close to a bingo, it could make it more difficult, forcing me to possibly use a turn to dump it. DAWNING sims 6 ahead of WAVING, and wins 2% more. It also has the possible benefit of using the D when two more are unseen.
Another interesting option I hadn't considered was N7 GNAWN 17, killing the M, N, and O columns. It sims well behind though, perhaps because it only scores 17 and keeps DNV, and there are still row 2 and column A to contend with.
>Randy_Hersom: CDO I7 C.OD +11 275
>David_Weisberg: DELNNOT N5 OD +19 374
#note Fascinating situation. With only three tiles in the bag and both Q and blank unseen, and columns A and O still open, my options were very limited. Randy just blocked the easiest spot for the Q, which didn't require having an I or anything else. Therefore drawing it could well mean I'd get stuck with it, since I have to assume he has the blank for these Q-stick scenarios to be a threat. The caveat with simulating this position is the known issue with Quackle's finite search depth, which isn't the best at handling situations where a player ends up stuck with a tile. However, after a bingo, there at most only one or two tiles he draws, so he can't slow-play me here.
I'm also not sure how to approach this since Quackle thinks Randy has 3 more points than he did during the game itself, due to undercounting RELINKING, and I'm not sure whether that makes any difference. Per the correct scores though (355-275, and not 355-272), and running a 4-ply sim to 25k iterations, (to allow for him fishing), emptying the bag with N2 NOTED 32 sims best at over 99% wins. N1 TENDON 27 wins over 98.5%, and N2 TONED wins nearly 98%. I'm not sure why NOTED and TONED perform any differently. My play looks to be the best one that doesn't empty the bag, and wins nearly 94.5% at the correct scores. Just behind it is N5 ON 13 at 94.25%.
Blocking the A column instead with B5 ON 8 wins just over 94%. B5 NO 10, oddly, comes out 1% lower. Another consideration is playing only one tile, to better avoid the Q. I looked at 6M AN 4 for this during the game, which wins just under 90% (it still allows -ING bingos on column O). The much better option for this, which I didn't think of, is 5B NO 2, which wins just about 94%. This would all be very tough if not impossible for a human to calculate well in real-time, especially with my usual low time likely remaining by this point. In a close enough game, we'd find those 3 points Randy missed during a recount, but it's possible thinking I was further ahead than I in fact was influenced me to try less to simply outrun bingos. The idea scenario for Randy if I draw the Q is for him to get an S as the last tile after a bingo, for K3 VAHINES 13, and I think this is part of most of the wins fo him.
>Randy_Hersom: ?DIIRST A1 DISTRIc. +77 352
>David_Weisberg: ELNNOST O1 NONET +24 398
#note I should slow-play him here, but I was about to go overtime if I didn't go out, so no time to calculate whether I could make up those 10 points I'd lose.
>Randy_Hersom: Q - +0 352
>David_Weisberg: LS K4 ......S +13 411
>Randy_Hersom: Q - +0 352
>David_Weisberg: L 12B L.. +10 421
#note Final score recorded as 441-349.
>David_Weisberg: (Q) +20 441
|
