Readless Ranges: Big Blind Play 20bb Deep Against an Unknown
When we do not know what our opponents frequencies are, we maximize our expectation by playing against the aggregate frequencies of our opponents, which I like to call “population tendencies”. Sure, sometimes we will end up being too aggressive, and sometimes we will end up not being aggressive enough, but that is inevitable no matter what strategy you use readless. You make the most money by playing against the population tendencies, and then adjusting with good Bayesian thinking once you start to see what your opponent in particular is doing.
For this article, let's work with a specific example: What to do 20bb deep readless from the big blind. Remember that we will never actually be truly readless, as we have certain information at our disposal before the first hand is even dealt. There are some jokes to be made here about knowing the nationality of our opponent and how dumb the screenname is, but I mostly mean having never seen a screenname before and thus having no information about it actually is a read that we should be taking into account. Unknown opponents are far more likely to be recreational players, and we should shade our expectations accordingly. I believe unknown screennames generally open for a raise less often than known players, probably somewhere between 50-65% of hands, limping some others. To calculate, let’s use a 55% opening range (specifically, ProProkerTools’ top 55%, which is much better for this than PokerStove’s). If we go all-in, we’ll use a calling range for our opponent of any pair, A5o+, A3s+, KTo+, K9s+, and QTs+. This includes some light hands, but nothing crazy, and encapsulates the idea that we will occasionally get some random rather light calls from this type of opponent, but not super frequently.
Based on these conditions, here is a table of the expectation from going all-in over the minraise from the big blind. For those not familiar with these type of charts, the pairs go diagonally down the middle of the table, with suited hands above the pairs and offsuited hands below the pairs. The expectation is referenced from the start of the hand – so when we go all-in with A3s with effective stacks of 20bb, this table means that we should expect, on average, to end up with 20bb at the end of the hand: 0EV. Note that if we were to fold, we would be left with 19bb, so shoving is clearly better than that. If we were to just call, it is hard to say exactly what our expectation is, but based on my historical results calling with various hands, I predict that calling would do slightly worse than 0EV from the start of the hand. A3s is also a hand that plays very poorly making a non-all-in 3-bet, allowing opponents to call in position with hands that have very good equity against it. Thus, I consider a 3-bet shove the best option, and the square is colored green.
In the traditional paradigm of looking at 3-bet shoving (is it better than folding?), 35% of hands make the cut, being equal to or better than -1.0bb from the start of the hand. However, be very careful about this: Even if J8s is a +EV all-in against this opening and calling range when compared to folding, this is a very silly comparison – we are never folding J8s to a minraise at this stack depth. That is why I think it is better to reference the expectation from the start of the hand and not to folding, as it reminds us that shoving needs to be better than all other options in order for us to maximize our expected value.
Lifetime, with [Q8s, Q7s, J9s, J8s, J7s, T9s, T8s, T7s, 98s, 97s, 87s], a cluster of hands like J8s, I am 0EV from the start of the hand from flatting. That is a whole 1bb in expectation better than folding, and much better than the expectation shown in the table for shoving, which are between -0.5bb and -1.0bb from the start of the hand for this cluster. That is why all these hands should be flatted readless against an unknown opponent 20bb deep, not jammed.
When you actually narrow it down to the hands that are worth 3-betting, it's about 19.5% of hands. Because your opponent will openfold more often when you have strong holdings due to card removal, the true percentage you will actually make a 3-bet is even lower.
Despite their mediocre equity from jamming, KQ, KJ, KTs, and QJs do extremely well in 3-bet pots, and gain a lot of value from making a potsized 3-bet and inducing calls from loads of dominated hands, or hands with two undercards, that we play very well against on the flop. I recommend 3-bet/calling these hands readless and bet/calling a ton of flops when you get flatted. Along with those hands can go your premium pairs, which also want to get flatted by a wider range.
3-bet bluffing with non-allin bets when readless is a different story, and I would be willing to consider arguments that this leads to the best expectation for hands like J3s readless. However, the math is pretty clear that it is spew to jam with these sorts of hands without evidence that your opponent is very likely opening a wide range, unless you disagree with my initial assumptions and think that the unknown opponent opens wider or calls tighter. Once you believe your opponent is opening wider, the math changes considerably.