Why is betting small on river IP not a thing?
Sure, we risk getting raised but couldn’t we just add some value to balance the range and be prepared to call or 3bet some of the raises like solvers do OOP when block-betting?
11 Replies
Sure, we risk getting raised but couldn't we just add some value to balance the range and be prepared to call or 3bet some of the raises like solvers do OOP when block-betting
Others might correct me, but likely it’s because when we bet the river IP we should do so with a very polarized range rather than a merged range. If we have a marginal hand with some showdown value, we typically should be checking rather than betting. We should only be betting the river for two reasons - to get called by worse or to make better hands fold. This implies that we should be betting our thick value and our air, I.e. a polarized range. We should typically use large sizings when we are polarized, both to increase our wins on value bets and increase our success rate when bluffing.
Sent from my iPhone using Tapatalk
Could also split the ranges in two?
Not sure what you mean by that. We only have one betting range. We can construct it in different ways, but OTR we generally want that range to include only thick value hands that can be called by worse, balanced by some bluffs, which should be air, not marginal holdings. Our marginal holdings perform better as check backs since we have SDV.
As an example consider a particular combo - 87hh- on two different boards. The first board is something like Kh 9h 6c 2c Td. The second board is Kh 8c 6h 3d 2c. If we are IP OTR and our opponent checks, we should be playing these in totally different fashion.
Board 1 is an easy bet as a bluff. We have no SDV. Assuming we raised PF and barreled at least one street we can credibly represent hands like Kx and QJ that would force our opponent to fold hands like Axcc or maybe even hands like Tx.
Board 2 is a clear check back. Villain is not going to call a bet here without a hand that beats our second pair. Any hand he folds will almost certainly be one that we beat. There is very little to gain with a bet on this board.
I didn’t include a board like Kh 9h Ts 3d Jh here. It should be obvious, though that we should be betting this type of board with our thick value. Since we are either betting with thick value or total air, though, we choose a large size in general.
Sent from my iPhone using Tapatalk
I meant that we could split our betting range into a large and a small size
With a perfectly polarized range, we maximize our EV against a rational opponent (solver) by betting as big as possible in position on the river. Betting bigger allows us to leverage more bluffs to maximize our share of the existing pot. As we bet bigger we can just add additional bluffs to force our opponent to defend and keep us from stealing the existing pot.
So if we put nutted hands into a smaller bet range we lose value with them, and we also can't bluff as much.
Okay, and why does that argument not apply OOP?
OOP you have an additional problem, in that whenever you check your opponent can then bet with a polarized range to capture additional pot share. So you also have defensive considerations, whereas in position you can immediately showdown your hand by checking.
In position you're always going to want to bet the nuts but OOP there is some incentive to trap. By betting smaller OOP with some non-nut made hands along with a sliver of traps you can change the distribution of hands so that your IP opponent can capture less pot share. Just a relatively small amount of traps in OOP's range can be enough to prevent IP from going all in with all of the hands at the top of their range, especially at deeper stack depths, as IP is now no longer perfectly polarized and has to worry about betting into the nuts.
It does get quite complex and confusing in real poker where ranges are often not perfectly polarized, and this is totally ignoring blocker effects. You might try reading the chapter on river play towards the end of Modern Poker Theory by Michael Acevedo for a more thorough and technical explanation than what I'm giving.
If you model the river with an (OOP = A/Q) vs (IP = K/J) toy game, it turns out that the optimal bet size IP = (SPR + 1)/(SPR*2), which equals 50% in the limit. This holds true for any SPR and even if you vary the amount of A in OOP's range.
If you model the river with an (OOP = A/Q) vs (IP = K/J) toy game, it turns out that the optimal bet size IP = (SPR + 1)/(SPR*2), which equals 50% in the limit. This holds true for any SPR and even if you vary the amount of A in OOP's range.
I must be missing something. Why would IP bet river at all?
OOP can lead. And you can have variations where OOP has the Ace less often.
It's an interesting toy game because you need to make IP's bluffs AND value bets indifferent. Most poker toy games stop at the former
Finally got it, thanks 🙏
