Decision Making

MIT 15.S50 Poker Theory and Analysis, IAP 2015
Instructor: Matt Hawrilenko

In this session, guest Matt Hawrilenko discusses game theory, value betting and bluffing, how to maximize the value of the entire set of hands, and exploitive strategies.

License: Creative Commons BY-NC-SA

[danielcharlton]

This was explained really well, good job hoss xD

[briannewman]

Sounds interesting, but the fact that the projector screen is rarely shown makes it hard to follow.

[chriswilson]

Or is balancing itself the exploit? Or does that just make us harder to exploit? Is it better to exploit or not be exploitable? These are some of the questions I'm interested in hearing your response to.

[albertthorval]

I don't get it at 24:16 or page 22 of the PDF.

For me they dont lose juste the valuebet (1), they lose the value bet only when you're not bluffing so they lose (1-%Bluff) . bet
So in my mind it should be b ( P + 1 ) - ( 1 - b ) = 0
And in the end b = 1 / ( P + 2 )

Could someone explain me where I got wrong?

[yoooooooooo]

What would he bluff with here, 78, 910? There's more value hands in his range, maybe he bets this much with trips or a full house to make it look like a bluff to get calls from AA, QQ or a J..

[Markus-Domanski]

At 46:00 he says that you should call with 40% of hands that beat a bluff. This is incorrect, because it would come down to a conditional probability. Given these odds you should call in 40% of the cases overall. With polarized betting on earlier streets, you should usually end up with enough strong hands on the river to do so, but it can be tricky. This one of the examples, where things look much easier in a theoretical model, like the [0, 1] game.

[brocojack]

Double like.
Youre surprising for a Game Theory Jujitsu man.

[menfoy]

they had an old guy teaching the same lecture before, anyone knows where to find him explaining this stuff ? would highly appreciate

[BobF]

Engaging with this content is enlightening. A book I read that shared these ideas shifted my entire mindset. "Game Theory and the Pursuit of Algorithmic Fairness" by Jack Frostwell

[AnteUp]

I had NO IDEA that this was on the internet! Super like! Love HOSSTBF!

[chriswilson]

What if to exploit requires us to unbalance our range? Do we not balance or do we not exploit?

[PhoenixTTD]

Why are KK and KJ part of your 40%?  These are not bluff catchers, they are the nuts.  Why not ask what is his value bet range, and then draw a line for the hands that still beat that range?  That will open up calling with your KT and K9s.

[DanielEGreen-bpff]

I respect this, but this often, but only applicable to rational players.

[johndoe-uglo]

So how can the idea of optimal game theory play alone actually turn a profit. If you are leaving your opponent with only equal ev options, then doesn't that mean that it is impossible for your opponent to make an unprofitable play, no matter how bad they are?

[shagnastyfo]

Legend has it to this day the speaker still needs a glass of water

[nikopakalen]

Not checking any Kx flop or turn makes checking range weak against overbets and also if some Kx or stronger check the river range gets slightly weaker (as was proven here Hero just has it too often as played). Also given the board ranges are not at all symmetrical on the river (due to very different preflop distributions Hero has strong Kx hands but not as many king high full houses or straight draws). When considering river bluffcatches or reraise bluffs blockers are extremely important, for example here your range is so strong villain shouldn't value overbet anything less than full house so if you want to call or raise without K blocker Jx works better than AA.

[lobsterworldwide]

Why are you reaching the MIT college fish?!?

[oletramekaf]

This lecturer talks more about himself, his favourite things in life and random stuff, than about the subject. lol

[Big-guy]

That's MIT? Seriously? Sklansky wrote about this stuff like 40 years ago. I mean some college dropout got it better than 6 MIT guys and their fancy degrees.