Imagine an opponent that makes a pot sized bet 100% of the time on every street, but that will respond normally to a raise (folding or continuing based on the strength of their holding). How should you deal with that opponent? We’ll assume that you know their strategy.
Let’s start as simple as we can, where we are heads up against that opponent. On each street we need to put in x of our own chips for a chance to win 2x chips (the existing pot plus the pot sized bet), and so against bets from a completely random hand, anything with over 33% equity seems like it should continue (or a 33% chance to be ahead if we are talking about the river).
It can be hard to continue with two thirds of your range, even when you know the bets are coming from completely random hands, so what hands do you pick, and what do you do with those hands? Let’s start by saying that I don’t know the answer here, and suspect a strong AI would choose different hands than I would, and would also choose different responses with those hands, but I’d guess the broad principles might look something like this:
Raise with some of your worst hands that block value. Note that against perfectly random hands the concept of blocking value combinations gets significantly diluted, as it is no longer true that a hand like KQ blocks jacks on a J high flop, and so this might be as trivial as blocking straights and flushes, or having combinations that dilute the overall equity of a random hand.
Call with some of your worst hands with an intent to raise on a later street if the equity of the hand improves.
Fold a bunch of your lower middle hands that are less likely to be able to hold up to bets over multiple streets.
Call a bunch with good hands, mixing in raises only to balance the bluffs above, assuming you think your opponent is paying attention to what you are doing, and where your hand benefits a lot from protection.
Why not raise more: your opponent’s range is dense with really weak holdings, and if we assume a normal response to a raise, you get more value by just calling and receiving the next pot sized bet.
Polarizing seems like the intuitive solution. Def expect solver to pick some weird hands that make humans go “huh?” Will reply in greater detail later when I’ve got time to chew on this.
Btw unsure if this was your intent, but there are definitely some Replay regs who more or less play like this! I made a forum post previously asking how to deal with villains who cbet full pot on every flop as the PFR.
3-bet or fold pre-flop. I guess the exceptions to that would be if we’re in the big blind and they’re raising 100% of buttons, or we’re opening and they’re calling OoP and then donking on the flop.
Those play a bit differently, but either way we get to massively overfold. We make so much money by just calling down with our very good hands that we never need to continue with the mediocre stuff.
The hands I’d be raising are pair + flush draw, combo draws, draws with over cards. There’s probably some vunerable value in that range too, like bottom 2 pair on a connected board. I don’t think there’s any need to find low equity bluff raises.
I just gave it a spin as the aggressor on a 6 max table. Something like this might be a little more viable heads up… I didn’t play enough hands for it to be too statistically significant, but against multiple opponents it felt pretty spectacularly minus EV. It didn’t help that several opponents just called with strong hands to leave me committed to stacking myself, and thus managed to get to 10 times max buy-in quite quickly.
Things that seemed like problems:
being committed to a pot sized bet unless someone else raises when you’ve had multiple callers on a prior street just bleeds too many chips; you just can’t bluff as wildly with multiple opponents
because you’re building the pot so fast, you don’t have a lever to create an exponentially bigger pot when you do hit a hand
I’ll tweak some things and give it another spin later.
I’d agree that it’s pretty easy to show a profit without using low equity bluff raises, but if you consider how weak your opponent’s range is, I think you’ll find that these low equity bluff raises will also be very profitable.
It wasn’t clear to me if we’re assuming villain has any two cards, or if they can get to the flop with a more reasonable range.
Against a really wide range though we generally want to be fairly linear with our responses. I agree we’re going to win a decent amount of pots with low equity bluffs, but we lose a lot of chips the times they don’t get through.
The problem is not many of our good hands want to raise, so if we start bluffing our low equity stuff, even bad opponents are going to adjust pretty quickly. I’m pretty sure this is losing in the long run.
Even if they adjust, with the assumption that their range is every single hand, that range just won’t be able to defend at an optimal frequency, and so the bottom of your range can make pure bluffs that should be profitable. I’m also not saying you should bluff will all of your lowest equity hands; you want some features that make it seem somewhat more likely that a bluff will succeed. Normal blockers will not be as meaningful against a completely random range, but there will still be some blockers or board characteristics that change the odds.
I also think you over estimate how quickly or easily even an intelligent opponent will be able to adjust. The plan I outline above doesn’t involve raising with a high frequency, and should be mixed with some weak but higher equity holdings, and some strong but more vulnerable holdings. If they do adjust, then you can simply add a higher frequency of strong hands.
In practice, do you need to do something like this? As a thought experiment, I propose this will increase your overall EV, but it will also increase the volatility you experience, and some won’t really find this appealing, even if they are sufficiently bankrolled for a higher volatility style at the level they are playing at.
I’m trying to work backwards from @lihiue’s proposal by finding a reason it wouldn’t be max EV to just play a narrow range, call down with strong, less vulnerable hands, and raise with strong but vulnerable hands. So far I can’t think of a counter. Not to say this means there isn’t one.
wouldn’t something like pair + flush draw be more profitable as a call? if we raise he might fold rags. if we call, he keeps betting with his entire range.
All fair points. I guess my view is a bit tainted because I play quite a bit on a site where the pool is mostly unknown and you have to make reads fairly quickly. I’m often making reads, rightly or wrongly, on just one hand. Not many people are capable of raising with air against pot, so I’d be tagging you immediately. Then again, I’m not blindly betting pot on every street with any two cards either, so I agree such an opponent probably isn’t going to adjust.
I think as the aggressor, that’s the only way it’s really going to work here. You make the most money by people chasing their draws and then folding on the river when they miss, which will happen, but if you have 2-3 other people in the pot as well, you’ll just be bleeding chips to the weak value hands that should fold but won’t.
Depends on the pair I think. Top pair is definitely strong enough just to call down with, middle or bottom pair could go either way I guess. Can’t really go wrong with those hands against this type of player either way though I think.
If Hero is gonna play this strat I would start by using a high-postflop-yield range preflop. I am opening 66+, not sure where the suited connectors should start but maybe 78s+, AQ+ and maybe literally nothing else.
My thinking is that we need to be able to crush when we hit, and we need hands that can continue pot-pot-pot and have a reasonable chance of making the nuts against the field.
Still gonna be brutal when we play the connectors and the flop is not favorable, though. Hmm. This sucks!
The more I think about it, I’d almost rather always check every street, assuming we’re allowed to c/r.
Hey, how’s that for a wrinkle? Would you rather play the above strat, or its opposite?
