I searched the thread but didn't seem to find the answer.

Anyone know what these guys are being compensated for this? I imagine it would have to be a very large sum.

Either way, thanks everyone for doing it. It's been educational and enjoyable.

the sn PremiumWhey fits Cheet's 2p2 persona a bit better tbh. can't take jmo's word at face considering that he prob got grilled a good bit for outing flushie

Doug seems like a super happy person lol

Let me get this straight.

You do not even play low stakes, you play micro stakes, yet you are able to accurately determine that this bot could not beat low stakes players(that you don't beat yourself)?

Just realised I was being dumb. There is no way this bot would beat mid/lowstakes regs for a high winrate since the reason top players can crush them is because of highly exploitative play. Bot is never going to do that.

Let's talk about you instead. Who are you and why do you keep a dummy account instead of using your regular one?

naw you werent being dumb, now tho.

You dont need to adjust or exploit at all to have 10bb+ against midstakes and especially lowstakes regs

that's a fact as those who would play a winning player are either losing players(as the good ones have moved up) or will sitout after 3-4 hands

Closing in on the 1k hands played mark, Team Human up 51k.

I'm not sure if this is accurate throughout the whole competition but it seems like the bot is super polarized in spots where it raises to a 'committed' amount while still leaving like 1/3-1/4 of it's stack behind. Then the person it's playing folds or shoves for the remaining amount with what essentially a bluff catcher.

Jason is looking like the Rajon Rondo of Team Human -- time to find a "back injury" and put him on the bench for the rest of the series so that Team Human can end up a million.

haha it would suck so much being in his position getting the run of cards he has

2.5k hand mark and Team Human is +50k .

Dog Pork looks like Neo who's humanity's protagonist against singularity uprising

Gogogogo Doug show the computar machinas our only weakpoint is hungry & sleepy

i think the bot is making some huge river mistakes

first we saw it overbet shove Q2 on A2K2A vs flushie, then just now

8x 8x 5x 5x 5x

wcg open 95o otb, bot defends J8s, some flop and turn minimal action happens, then bot leads river for hundreds and calls the allin overbet shove

the logic it uses for that river call seems really broken cuz i dont think any human is bluffshoving that river

Yeah I guess on that river bot can just fold 100% unless he has quads himself vs that sizing.

Cool contest. Can't wait til the hhs are public. GL humans.

Exactly.

Edit: I misread your post ... of course you can not just fold unless you have the nuts yourself, duh

Apparently HH's will not be published. Someone would actually have to write a screen scraping script ...

just for today right? This is in addition to the 165k theyre already up?

100% is exploitable.

Lul, wat? Won't they have to publish with any paper? Sorry if this was discussed already, link is great if so.

pretty sure i heard russian bill gates saying they agreed not to release any HHs for the sake of team human. right before he asked why the bot wouldn't bet smaller with a jack.

Well then. Can't say I blame team human.

The thing is with this sort of huge river overbet situations being so rare, it's not unlikely for the bot to make this sort of mistakes. In the 88555 hand if the bot assigns himself to have very few 5x it's possible that he can call the all in with 8x to avoid being exploitable to atc shove (ofc it will be mixed strategy most likely, just not 100% fold).

I remember looking at the bot ai competition hhs from a few years ago, and they were making "huge" mistakes not raising the nuts on the river, and one hand where a bot folded the nuts ott (it was the nutflush- so not stone cold nuts as board can pair) to a huge overbet. In terms of bot development these sort of things will always be there imo, a bit similar to how there can be exploits in chess/backgammon vs ai.