Most players understand that bluffing some fraction of the time is often good strategy, but perhaps don’t have any good heuristics on when to bluff and when not to, or also on how much to bluff. Jonathan Little shared a decision tree that I thought many might find helpful.
It starts with questions designed to decide whether or not you should bluff, and then if you reach a bluff decision, questions to decide whether to make a large bluff or a small bluff.
To bluff or not to bluff?
Is the opponent capable of folding?
Do I lack showdown value?
Do I have many value hands in my range?
Do my cards block auto-calls?
Do my cards un-block auto-folds?
Do I bluff big or small? (Affirmative answers support larger bluffs)
Do I have zero showdown value?
Does my opponent have a lot of marginal value hands in their range?
I would probably expand on number three of the first part to consider what effect my current table image has on my perceived ranges.
One could argue that table image should factor into bet sizing too.
One example that comes to mind is when you have a LAG image and happen to have a top-of-range type hand. You can often bet low boards because your LAG image gives you better board coverage.
Yes, I also think that would be a great addition, though rather than perceived range, I think I might reduce it down to a simpler idea: if I’ve been betting a lot without getting called, or have had recent large bluffs called, bluff less, and if I’ve been really inactive, or if I’ve had lots of value bets recently called, bluff more.
Yeah, and we are pretty much talking about the same thing here.
Here’s something to think about…
“If I’ve been betting a lot without getting called… bluff less,”
What if I don’t want to bluff less?
I think I can selectively show my cards when they fold, so I can skew my table image to the point that I can still bluff. In fact, if I can convince them that I’m only opening very strong hands, I might be able to bluff a little more.
I would rather make them adjust to me, not the other way around.
I notice that he doesn’t mention anything about effective stacks, SPRs, or ICM.
Surely he must at least consider some of these things. For example, he’s not trying to bluff the BB if the BB has less than 1 BB behind. Maybe an extreme example, but that principal is often at work in one way or another.
Someone mentioned that your list is a lot to think about in the limited time available, and it is. But there’s more to it than the list implies too.
Some seem to think that bluffing is just a matter of shoving some chips into the pot when you miss a draw, but a profitable bluffing strategy is a lot more complicated than that.
One thing I’ve been focused on a bit more recently is how to avoid over bluffing or under bluffing, and so I try to quantify the third bullet above, “do I have many value hands in range”, and think specifically about how many combos of value I have that would make a particular sized bet, and once I have the combos and bet size, that gives me the number of combos of bluffs I should have.
For example, if I’m thinking of making a pot sized bluff on the river, then I want 2 value bets for every bluff. So if I have 18 combos of hands I’d want to value bet with, I either need to have some way of deciding what the worst 9 combos of hands I get to the river with are, or use a randomizer with all of my bluffs that lack showdown value, based on the number of combinations (so if I had 16 low card bluff combos, I’d want to bluff with those roughly 60% of the time).
Yet another approach, which fits in well here, would be to pick the 9 bluff combos you want from the hands with lowest showdown value that also block auto-calls or unblock auto-folds. Here, knowing your own ranges well can help you better decide what cards you realistically arrive at the river with given prior action.
Love the list, but keep in mind that technically you’re describing some kind of “weighted factor model” and not a decision tree. But yes, I think that’s a good list.
Keep in mind also that unless we’re talking about river play specifically, you’ve also got to factor your chance of making a nuttish hand on later streets into the equation. I think it’s worth keeping the following in mind: poker solvers tend to bluff their A-high flush draws with fairly high probabillity, and their weaker flush draws with fairly low probability. That way, you can’t easily tell if they’re value-betting or semibluffing, and in flush-over-flush situations they’re more likely to have an A-high flush whenever the pot is big, and a weaker flush when the pot is smaller.
Also, I think it’s good to emphasize “indirect blockers” when explaining these concepts to others. Most players understand that if they’re hold 88 on a 976 board, then they’re blocking the 6-9 straight and can therefore rep that hand to win the pot reasonably safely. What they don’t always understand is that if they’re holding 86 on a K77 board, they can potentially rep a 7 to win the pot fairly safely. But of course, you should only do this if there was enough raising preflop to remove most of the silly hands that might have connected with that 7. Additionally, you should only do this if the players are rational enough to response to incentives in a sensible way. Against sufficiently bad players, you basically don’t bother “repping” anything, you just value bet when you’ve got it
I’d also revise this part of my response now. I don’t think you want to select the 9 worst combos in range, but rather 9 combos without showdown value that have the best blocking or unblocking characteristics.
Well said.