| Player 1 |
|
#player1 shdbwrking shdbwrking
#player2 kacattac kacattac
#description Created with Macondo
>shdbwrking: AEMRTVY 8D MARVY +32 32
>kacattac: CDGNPQS -DGNPQ +0 0
#note 20 January 2021
Club 44
woogles.io
Game 4
The rationalist says, "There must be a definitively best exchange here, even if the margin is tiny." The empiric says, "Gee, I do heart the C." After 15000k iterations, the empiric has the rationalist in a headlock and is screaming, "Why do you waste my time?!"
>shdbwrking: ?CEEOTU F1 ExECUTO. +61 93
>kacattac: AACFKST 1C FAK.ST +42 42
#note Nice find by opp, and now I have to get this turn right if I want to have any chance at all. Correctly narrowed it down to this and FECK. It's close but FECK outsims by 1.5 or so, and that seems fair. Having burned a hole in the dictionary during Covid, I'm really not too scared of duplicate AAs anymore, but I confess I got a little bit scared of them here and probably overreacted. FECK also practices tile conservation (yes, I believe in that) and exhausts the board ever so slightly less than FAKEST does. Those considerations are marginalia within ephemera wrapped in minitiae, but they do count, especially when you're already down to 20% win prob after two turns.
>shdbwrking: ADEGNNS C1 .ANG +16 109
#note SK gasps: yeesh, D(U)NNAGES.
>kacattac: AACEORU 4C .UA.O +16 58
#note Was all set to plop down 4B CURA(C)AO until this came down. Didn't like it as much at 4F. GUACO indeed sims best decisively.
>shdbwrking: DEENOSY 9C DENY +29 138
>kacattac: ACEIMOR 10B COMA +27 85
#note I'm really not far from breaking out of plateau, but I'm not quite there yet. I did look for dws-dws plays in column E, but I missed the very Nigely find of M(AN)IOC or M(AN)IOCA which scores comparably to COMA, keeps better, and opens without being so reckless about it. Sim doesn't terribly mind COMA but does pose some curious alternatives, all of which seem to me to have equal or greater downside. Holding back the C here creates a massive fork, which is entertaining but probably works against me.
>shdbwrking: AAEIOSV 9H AVO +15 153
>kacattac: DEEEIJR 8J JEREED +54 139
#note I am a player who sees JEREED but not JEERED.
>shdbwrking: AEINSUX 5E X. +36 189
>kacattac: BDIIIOR 10J BIO +29 168
#note The most interesting play of the evening, I think. Why tf did I not even see or consider BIDI? The short answer is, because I'm just not very good at this yet. With that established, is it even worth diving deeper into this? I think it is. Selon Duckbrain, BIDI sims 9 better and bingos twice as often; but "twice as often" here is 22% vs 11%, which is a very different proposition from, say, 45% vs 90%. In absolute terms, I am probably not bingoing next turn no matter what I play here; basically I just need him not to have a bingoish rack, and there's not a whole lot I can do about it if he does. This turn is really all about taking out the biggest threat, which is why nothing anywhere else on the board sims even close to the various JO(B) plays. Having thus made the block and still being down big tempo, do I really want to play max defense and extend this play by a tile? I think that a four-tile play here damages my prospects more than opp's prospects: I am the one who needs max volatility here, he is the one who would like to see the board start to exhaust itself; I am the one who has just turned my rack over; he is the one who has played short several turns in a row. What BIO does better than BIDI is that it hedges against long-term board exhaustion. Of course it scores 2 less and there's a very real risk of having triplicate Is again next turn. That is all true. BIO is objectively sub-optimal, a gamble rather than a gambit. I def would like to learn to focus better during scrabble games, to slow down and see that BIDI is an option, to weigh all of the above thoughts during the game rather after (it's tedious to try to write them down, doubly so to read them, I do realize, but really they're pretty quick thoughts in the flow of a game, no?). However, I don't want to learn never to play BIO just because the computer says it's wrong. Two more quick thoughts here: (1) perhaps BIDI is easier to underlap than BIO? (2) Opp probably thinks my rack is better than it is after BIO, whereas BIDI telegraphs a bit more. BIO is a better bluff than BIDI is a gambit. Ok I'm done now.
>shdbwrking: AEINSUW K7 N.W.E +26 215
>kacattac: DDEGIIR M3 GIDDI.R +24 192
#note This does sim best by a sizeable margin. Didn't want to do it, for all the reasons expounded upon above, but the upsides are obvious.
>shdbwrking: AEIIRSU 2E I.IA +13 228
>kacattac: EGHILNT 11E LIGHT +26 218
#note Wow, Duckbrain's static algorithm really undervalues this play. It sims very close to best. Sim wants to fish the H, but that's not really defensible here in the flow of the game. 11C TEL is an alternative to LIGHT which scores same and doesn't give back H11 plays. But really, I'd like nothing more than for opp to play at H11 and open the board for me; also noticing now that I can score+open there myself if he bingos down the upper right, whereas with TEL I have no such hedges and a counterbingo basically seals the game.
>shdbwrking: BERSSTU 12K SUB +19 247
>kacattac: EEFNPRZ N7 F.EZE +44 262
>shdbwrking: AAEHRST O11 HEART +32 279
#note SK butts in: 'nother pretty super interesting position here. Sim likes playing short with HA any number of places, or O10 ARHATS for big points but floating a dangerous S. Those are hard options to swallow. Sim loves AH at L3, perhaps because it blocks QI plays, but it seems to me that this also could give back big parallels and/or add value to a parallel play that I am already lining up in column N. 13L HA does not do nearly as well in the sim, but it seems to me like a very strong non-standard option, because it is virtually neutral vis-a-vis board dynamics and opp is actually in a very strong position here. Since he has the last AAs and is very synchronous, my options at O11 are very unlikely to be game-breaking and more likely to give back a bingo lane he can use; hence there's not as pressing a need as it might seem to take out this spot.
>kacattac: ?ENOPRR 14J PRONE. +28 290
#note No really, I did look for bingos up top, and I have drilled on PARERGON quite recently. A future project will be special lists of duplicate L, N and R words. I really have trouble seeing them. At least it's not quite a fatal error in this case as PRONER sims next best. This is not as dangerous an opening as it looks, and opp has just burned five tiles.
>shdbwrking: ALLOPSS H11 .OLPS +33 312
#note SK equivocates: FMFL, I realized this was phoney as soon as I stepped away from the computer after the game. I really needed to realize it in the moment.
>kacattac: ?IIIRTW N2 WIT +17 307
#note Does it count as a meltdown when you've previously guiled your way back from near-hopeless circumstances? This should have been a triumphant moment, but instead it's just another gaffefest. I am almost out of time at this point. Sim loves 3A WI(N), which besides the dreaded irony (why does this keep happening?!) doesn't seem to me like the kind of thing that actually works very often against live human beings. There are many similar alternatives, though. I did see that WIT potentially gives back QATS. I was thinking that I need to keep open the option to outrun without bingoing; sim essentially disagrees, finding that my best chances are to go all-in for a walk-off.
>shdbwrking: ALLNNQS C9 ..LL +5 317
>kacattac: ?EIIORU O1 ROEs +20 327
#note Now over time, so actually down by at least 20. Plays forming (WIT)E block QATS! Sneakily, this makes somewhat WIT less bad than it first seemed, no? But that is results-based thinking, which has no place in Our Game. My process here was FUBAR and now there's literally no time to recover.
>shdbwrking: ANNQST 15E QAT. +13 330
>kacattac: IIU - +0 327
#note About to flag, so just passed.
>shdbwrking: NNS 4L N... +6 336
>kacattac: IIU J14 .I +4 331
#note I flag anyway.
>shdbwrking: NS C9 ....S +12 348
#rack2 IU
|
