Small Pocket Pairs: 100BB vs 20BB Strategy Differs

Set mining is the most overrated idea in low-stakes poker, and the reason isn’t that the math is wrong. The math is fine. It’s that the math only describes one of the two games a small pocket pair plays, and most players keep running the deep-stack version long after the stacks have stopped being deep.

Pocket deuces through sixes are the purest example of a hand whose value is almost entirely structural. They have no kicker, no suit story, no ability to improve into something disguised except by hitting one of two remaining cards. Whatever they’re worth comes from the stack sizes, the position, and who’s left to act. Change those and you haven’t adjusted the hand, you’ve swapped it for a different one.

The clean way to see this is to price the hand twice: once with 100 big blinds behind, once with 20. Same cards, same seat, same opener. Two completely different decisions, and the correct play at one depth is a leak at the other.

What the set-mining math actually demands from your stack

You flop a set about once in every eight and a half tries. Call it 12 percent. That number never changes. What changes is what the set is worth when it arrives.

Run the loop. Somebody opens to 2.5 big blinds and you call with pocket fives. Over 8.5 attempts you’ve put in 21.25 big blinds total. Seven and a half of those hands miss and get folded, which costs you 18.75. So the one time you hit, you need to clear roughly 19 big blinds in profit just to break even before anything else goes right.

At 100 big blinds that’s comfortable. You have 97.5 behind, and you only need about a fifth of it to come across. Sets get paid. Someone with an overpair, a top pair they like, or a flush draw will often deliver far more than 19 blinds.

At 20 blinds it falls apart immediately. You call 2.5 and have 17.5 behind. If your opponent stacks off every single time you hit, the absolute ceiling is 20 blinds. You need 19 of that 20 on average, which means you need them to go broke essentially always, on every texture, against a player who just called and then woke up. They won’t. Add the hands where your set loses to a bigger set, a made straight, or a runout that kills the action, and the flat call at 20 blinds is a slow bleed.

The old “call if you have 15 to 20 times the bet behind” heuristic was never precise, but it was pointing at exactly this. It’s a stack-depth rule wearing the costume of a preflop rule.

Below 25 blinds the pair stops being a lottery ticket and becomes a favorite

Once flatting dies, the hand doesn’t get worse. It gets reassigned.

A short stack isn’t trying to flop a set. It’s trying to get all the chips in while holding a hand that’s ahead of, or close to, whatever calls. Pocket fours against two random overcards is roughly a 52-48 favorite. Against a suited connector it’s better than that. Against a bigger pair it’s a disaster, but bigger pairs are rare and calling ranges are wide.

That’s why a hand that can’t profitably call 2.5 blinds at 20 deep can profitably shove 20 blinds. You’re no longer buying a 12 percent chance at a big pot. You’re collecting the dead money most of the time, and holding a coinflip or better the rest of the time. The hand’s job switched from implied odds to raw equity plus fold equity.

The practical version: from late position at 15 blinds or less, every pocket pair is a jam or a raise-fold-never open. At 20 to 25 blinds from early position, a small open with a plan to fold to a shove is defensible, since you fold out plenty and you’re not committed. From the blinds facing a late open, calling gets more reasonable again, because you’re getting a discount and you close the action. Kill Everyone by Lee Nelson is old now, but the short-stack framework in it still describes this transition better than most modern material.

Being outnumbered helps at 100 blinds and hurts at 20

Deep, more callers is usually good news for a small pair. You’re paying the same 2.5 blinds for the same 12 percent, but now there are three stacks that might pay you off instead of one. Multiway pots are where set mining actually earns its reputation.

Short, more players is pure downside. You have no implied odds to collect, so extra opponents only add ways for your hand to be dominated when the chips go in. Shoving 18 blinds into two callers with pocket threes is a considerably worse spot than shoving into one.

This is also where the reverse risk shows up deep. Flopping a set and losing everything to a bigger set is a real cost, and it happens more often multiway. So does flopping your set on a board that hands somebody a straight. Deep stacks give the money back on those boards, and the bigger the stacks, the more they give back.

The ante changed what “20 big blinds” costs you

A stack isn’t just a number of blinds. It’s a number of orbits.

Big blind antes made that arithmetic harsher. When one player posts an extra full blind for the table, the cost of folding a lap goes up sharply, and a 20-blind stack decays faster than the same 20 blinds would have in a structure without them. The practical effect on small pairs is that the window where you can afford to see flops closes earlier than it looks like it should. If you’re waiting to find a clean 100-blind spot to set mine, the antes are eating the stack you’d need to do it with. How big antes reshaped opening ranges is a bigger subject than pocket pairs, but small pairs are where the pressure lands first, because they’re the hands with the least equity when called and the most need for chips behind.

Where money actually matters more than chips

Late in a tournament, chips stop converting to dollars at a fixed rate. Winning chips helps you less than losing them hurts, because the pay jumps are real and your stack isn’t.

That gap has a name people throw around: risk premium, meaning the extra equity you need before a call or a shove is worth making, once you account for the ladder. On the money bubble with a medium stack, that premium can be enormous. A jam that’s clearly profitable in chips becomes an obvious fold in money.

Small pairs feel this more than most hands, because so much of their short-stack value comes from marginal spots. Pocket fives at 18 blinds with three short stacks about to bust is not a shove, even though a chip-count solver would tell you it is. Meanwhile the same hand becomes a wider jam than usual when you’re the short stack, since the ladder is worth less to someone with nothing to protect.

When you miss, which is most of the time

The unimproved pair plays differently by depth too, and this gets overlooked.

At 20 blinds, pocket sevens on a 9-4-2 board is a real hand. The pot is a large fraction of your stack, one bet commits you, and there simply isn’t enough room for your opponent to represent a long story. Check-call, or bet and go with it. You’re not folding a hand this strong when there’s a single street of pressure available.

At 100 blinds, that same pocket sevens is a bluff-catcher that will face three bets. You can call one. Calling all three is how people light money on fire. Unless you have a specific read, plan on one street of curiosity and then release. The hand didn’t get weaker; the number of bets you have to survive got larger.

Watch how the hand behaves at final tables and you’ll see both versions in one tournament. When Justin Arnwine closed out the RunGood Maryland main event at MGM National Harbor this month, the winning hand was pocket eights. At 20 blinds, that’s a shove and a prayer. At 100, it’s a flat call and a long afternoon.