|
|
#15
09-03-2007, 07:19 PM
|
|||
|
|||
Re: Sushi vs Clarkmeister
[ QUOTE ]
On the river, I don't know what you beat, other than 99 or AK without a spade. By this time, it looks like he has AK and I wouldn't think he'd call a bet on a 4 flush board with AK UI. He is somewhat agro, so I think I'd check and hope he tries to bluff since we look scared of the flush. [/ QUOTE ] I don't hope he bluffs because I don't think you have to call the river bet. 
|
|
#16
09-03-2007, 10:02 PM
|
|||
|
|||
|
Re: Sushi vs Clarkmeister
OK guys. It's time to stop relying blindly on "theorems" that aren't theorems at all. Let's instead talk about something that IS in fact, mathematically true: EV. To do that, let's start by writing out the hands in villain's range.
AA? I'm discounting this completely. KK? I'm assigning one hand's worth for the lot of them, since I expect KK to raise at least once usually. Of the 6 normal hands of KK, 3 are flushes. The non-flushes beat us too, but I'll decide later if there's a difference between the 0.5 KKs that are not flushes and the 0.5 that are. QQ? I'm assigning half a hand's worth. It's a flush, and it beats us. JJ? It's not too uncommon for JJ to just call the flop, scared of a better overpair if you're betting into him if he capped, and it's not uncommon for someone to slowplay the turned nuts to the river. I'll give it one hand. I'm leaving TT out of his range. TT is seldom capped, and usually it would raise the flop or the turn. I guess it might be slowplayed to the river, but that's pretty rare. 99 and below are out. So, then we have the AKs. 7 are flushes, 9 are bupkis. Do we discount the bupkis since it might fold the turn? Well, AK has a gutshot here, and the pot is pretty big. It would have 3.5 effective outs against AA, 5 against KK, 7.5 against QQ, and against JJ or TT, and it might be chopping with another AK sometimes. Getting 10.5:1 here, I think AK will call the turn w/o a flush or a flush draw, not that all players will do all this analysis. In total, villain has 9.5 hands that beat us, and 9 that we beat. I think AK never calls a river bet from us, and I'll do one calc where the non flush KKs always fold and another where they always call. I'll also assume AK never bluff raises, and we always fold to a raise. EV of betting: (11.5 BB) * 0.486 + (-1 BB) * 0.514 = +5.075 BB If non flush KK folds to our bet, then we have: 11.5 BB * 0.514 -0.486 BB = 5.425 BB And of course, if it folds a % of the time, the extrapolation between these two EVs is linear. This doesn't mean that betting is correct, however. It means that we shouldn't fold, but we're not sure about the EV of checking. And that's the tricky case. Sushi, if you want to really understand this situation, I challenge you to work out the EV of checking, assuming you know your opponent's strategy. What's your EV if he bets/bluffs every time? What if he only bets hands he thinks are best (how often does he think he's best, btw)? What if he bluffs with game-theory-optimal frequency (what is that, btw)?
|
|
#17
09-03-2007, 10:31 PM
|
|||
|
|||
|
Re: Sushi vs Clarkmeister
I play it exactly the way you did, Sushi. Nice hand. [img]/images/graemlins/smile.gif[/img]
And, FWIW, I am betting that river as Clarkmeister. Pretty much with any two cards, if the fourth flush card comes down HU and I'm OOP, I'm b/f'ing the river.
|
|
#18
09-03-2007, 11:35 PM
|
|||
|
|||
|
Re: Sushi vs Clarkmeister
[ QUOTE ]
And, FWIW, I am betting that river as Clarkmeister. Pretty much with any two cards, if the fourth flush card comes down HU and I'm OOP, I'm b/f'ing the river. [/ QUOTE ] I think I'm fighting a hopeless battle here. This is such horrible thinking that is costing you money.
|
|
#19
09-04-2007, 12:14 AM
|
|||
|
|||
|
Re: Sushi vs Clarkmeister
[ QUOTE ]
OK guys. It's time to stop relying blindly on "theorems" that aren't theorems at all. Let's instead talk about something that IS in fact, mathematically true: EV. To do that, let's start by writing out the hands in villain's range. AA? I'm discounting this completely. KK? I'm assigning one hand's worth for the lot of them, since I expect KK to raise at least once usually. Of the 6 normal hands of KK, 3 are flushes. The non-flushes beat us too, but I'll decide later if there's a difference between the 0.5 KKs that are not flushes and the 0.5 that are. QQ? I'm assigning half a hand's worth. It's a flush, and it beats us. JJ? It's not too uncommon for JJ to just call the flop, scared of a better overpair if you're betting into him if he capped, and it's not uncommon for someone to slowplay the turned nuts to the river. I'll give it one hand. I'm leaving TT out of his range. TT is seldom capped, and usually it would raise the flop or the turn. I guess it might be slowplayed to the river, but that's pretty rare. 99 and below are out. So, then we have the AKs. 7 are flushes, 9 are bupkis. Do we discount the bupkis since it might fold the turn? Well, AK has a gutshot here, and the pot is pretty big. It would have 3.5 effective outs against AA, 5 against KK, 7.5 against QQ, and against JJ or TT, and it might be chopping with another AK sometimes. Getting 10.5:1 here, I think AK will call the turn w/o a flush or a flush draw, not that all players will do all this analysis. In total, villain has 9.5 hands that beat us, and 9 that we beat. I think AK never calls a river bet from us, and I'll do one calc where the non flush KKs always fold and another where they always call. I'll also assume AK never bluff raises, and we always fold to a raise. EV of betting: (11.5 BB) * 0.486 + (-1 BB) * 0.514 = +5.075 BB If non flush KK folds to our bet, then we have: 11.5 BB * 0.514 -0.486 BB = 5.425 BB And of course, if it folds a % of the time, the extrapolation between these two EVs is linear. This doesn't mean that betting is correct, however. It means that we shouldn't fold, but we're not sure about the EV of checking. And that's the tricky case. Sushi, if you want to really understand this situation, I challenge you to work out the EV of checking, assuming you know your opponent's strategy. What's your EV if he bets/bluffs every time? What if he only bets hands he thinks are best (how often does he think he's best, btw)? What if he bluffs with game-theory-optimal frequency (what is that, btw)? [/ QUOTE ] Anyone who hasn't read this post, should. Anyone who has read it and doesn't understand it should re-read it until they do. -J
|
|
#20
09-04-2007, 12:35 AM
|
|||
|
|||
|
Re: Sushi vs Clarkmeister
I'll try.
An estimate, that (hopefully) follows from MrWookies reasoning. Assumptions: Vils hand range is not influenced by our decision to check. (OBV) If we check, Vil is valuebetting all hands that beat us 90% of the time(give or take a few % to account for passive people with KK-no spade). We would expect a bluff from hands that we beat 40-50% of the time when we check since Vil is getting 10.5:1 on that bluff b/f (this number will vary largely depending on reads.). That would amount to... =(10,5)*0,484+1*0,242-1*0,514*0,9 = 4.86 And to benchmark the result, if Vil bluffs less than 50% the EV decreses on a sliding slope to 4.62 at 0% bluff. If Vil bluffs 100% of the time the EV converges to that of bet (note that the difference in EV is only ~0.5BB). OTOH... the c-c line will protect us from folding incorrectly against a bluff raise by a tricky player. Still, against an unknown bet ftw although the difference is not large and a read will change it into a simple c/c (i.e. if Vil has has tagged hero as someone who b/f often, this is (should be) an easy 100% bluff raise for him) If my % assumptions are off, please note the low impact a change in ratios of even as much as 30% have on the overall EV (i.e. if vil bluffs 100% when behind and checks 20% when ahead EV = 5.2, conversely if Vil is bluffing 0 and valuebetting 100 EV = 4.56
|
 |
|
|
Linear Mode
