btw FP wrote a pretty good blog post on this, hope there's no problem cross-posting:
I did sort of promise this post, so here it is. It's inspired by a recent thread on CC (http://www.cardschat.com/f11/4bet-bluffing-125589/), and a lot of what I'm saying here is what I said there, but I feel I can perhaps expand on a few points.
What Jagsti asked is "when, why and how" one 4-bet bluffs, and also with what hands. I think this is a pretty good starting point for the reasoning, so I'll use a similar format, albeit in a different order:
The obvious answer is that it's extra money in our bankroll. 4-bet bluffing is often profitable in a vacuum against aggressive regulars in the online games; just by occasionally 4-bet bluffing we can show an immediate profit. As an added benefit we also add some deception and make sure that our opponents can't squarely put us on QQ or better when we 4-bet preflop, thus allowing us to get more action on our big hands. Trust me, if your opponent shoves after you 4-bet and you fold, you can be certain that he just made a note on you. You should probably make a note back saying that he has seen you do this and avoid making your next 4-bet against him a bluff.
Here, I'm itching to actually answer the "when" first because a lot of what I want to say about "how" depends on the "when," but I think there's pedagogical value in doing this in reverse so I'll try it. When I 4-bet bluff preflop, I do it with hands that I can't profitably call the 3-bet with, and I do it to an amount that lets me get away from the hand if my opponent shoves.
The second part of that sentence is key. I need to be able to raise a large enough amount that my opponent doesn't just call because he's in position with good pot odds
, but a small enough amount that I don't get pot odds
myself to call a shove with any-two. A min-4-bet, in other words, is out of the question. A 4-bet that puts considerably more than a third of my stack in is also a bad idea.
After doing some simulations and a tiny bit of math, it seems that 4-betting to about a third of my stack (or rather, the effective stack) does the trick and is about as high as I can go without being committed. Perhaps needless to say, I use the same raise size with my big hands as well. This brings us very quickly into the "when."
Now, since I want to 4-bet with (at most) a third of my stack, I need my opponent's 3-bet to be small enough that my 4-bet allows this without being a pesky min-raise. If my opponent 3-bets to 25BB (with 100BB stacks), for instance, 4-bet bluffing is not an option for me. But the standard open-raise is typically between 3 and 4 big blinds, and the standard 3-bet for most regulars in the games I play seems to be between 12 and 15 big blinds. A 4-bet to 35 big blinds, then, achieves what I want to achieve, and is what I aim for in these cases.
It's imperative, then, that the effective stacks are at least 100BB. Otherwise, I'll have put in more than a third of my stack and will be very close to break-even on calling with any two cards when/if my opponent shoves, which I certainly don't want.
Furthermore, the whole point in 4-bet bluffing is that we think there's a decent chance that our opponents will fold whatever they have, so we want some history between us that shows that he's capable of 3-betting light. If you use stats, the "3bet preflop" number should typically show at least 7% for me to start considering 4-bet bluffing, and it's of course also a given that our opponent must be "smart" enough to fold the worst part of his hands when we bluff. Don't bluff a calling station - and definitely not preflop.
I also said that I want to do it with hands that I can't profitably call the 3-bet with, which adds to the when; I'm more often out of position than in position when I 4-bet bluff. In position I can often take a flop and play a big pot in position with some of my weaker hands, albeit certainly not all of them.
With what hands?
I'll cut to the chase: I (almost always) 4-bet these hands before the flop: QQ, KK, AA, AKo and AKs. Out of position, I also 4-bet all suited connectors from 76s up to JTs. In position, I 4-bet bluff with JTo.
It might seem like I'm being predictable if I always use the same hands, but I don't think that that's true. The way the combinations of these hands work out, I will "have it" about 75% of the time when I 4-bet. That's decidedly enough not to make shoving over the top immediately profitable for my opponents, and it's also enough that I still add a little bit of profit. So why have I picked these precise hands?
Because I want to be able to make the decision to 4-bet bluff automatic. I have two reasons for this:
First, it frees up time in my decision making when I'm multi-tabling (which I typically always am). A few seconds saved on making a decision on one table means a few more seconds to make a better decision on another table. This is pretty important and why I'm a big fan of having default ranges on reflex. That doesn't mean that I can't adjust, but some decisions I really just prefer to have made for me.
Secondly, it takes away the risk that I'm overdoing it. I think a lot of aggressive regulars seriously overdo the LAG style of their play and simply go nuts too often. They might "know" that they should be bluffing with a certain small frequency, but simply guessing how often they've been doing it lately is borderline impossible. By using a set of pre-determined ranges, I know for a fact that I'm bluffing with a frequency that I've decided on. No guessing. No 4-betting because I'm tilting.
So I pick JTo when I'm in position and suited connectors when I'm out of position, and you may already have guessed why, but it's simply because these are hands that I typically can't play profitably when I get 3-bet. In position, I can opt to call with JTs on the button when the small blind 3-bets. But I can't play that hand out of position for 14 big blinds with effective stacks of 100BB. And I typically can't play JTo profitably even in position when I get 3-bet.
An argument could perhaps be made to pick the weakest part of my range instead of these hands that are sort of in the middle. But I don't think it matters that much since these hands were all going in the muck otherwise anyway. There is an upside to choosing hands that are no better than J-high though, and that is that in the rare cases when the other person cold-calls the 4-bet (a move that I seriously question is legitimate for any hand but AA, and probably not even then), I don't have to be in the sticky situation of flopping nothing and wondering if I'm best.
This is not a hugely important point, but it matters at least a little bit, because if I see a flop with the bluffing part of my range and my opponent open-shoves out of position, I don't want to have a difficult decision to make. Of course, sometimes this is going to happen anyway. I might flop top pair with JT, but that's a much better situation to be in than to flop nothing with A7o and be worried that I might be laying down the best hand with ace-high. Since you're bound to ask, yes, I've had opponents stop-and-go a 4-bet on me with KQs unimproved and similar from the SB. Fortunately I wasn't bluffing at the times that this has happened, so I didn't make the mistake of folding, but I mention it as a reason for why I'm unhappy 4-bet bluffing with hands that have some chance of being the best hand. "Typically don't bluff with the best hand" is a valuable lesson I learned while grinding limit hold 'em, and it applies here, too.
So: 4-bet bluff because it adds profit and deception, do it in a way that lets you get away from your hand if your opponent shoves, do it versus opponents that you have reason to believe are 3-betting light often, do it with hands that you can't otherwise profitably continue with and do it with a range of hands that you've decided on before you are even dealt the cards.
And do it because, frankly, big bluffs are fun.