It is not that nobody presses this button.
Plenty of people press it. It is that every single time they do, they are holding one of the same four hands: AA, KK, QQ, AK.
The 3-bet post explained why so many players 3-bet so little: in their heads that button is not a tool, it is a way of announcing “I have a big hand”. A 4-bet is the same thing one level up, and purer — at least a few people will try a suited connector as a 3-bet; almost nobody makes an exception at the 4-bet.
An announcement can only ever be honest. A tool can have a purpose. This post is about what that difference is worth at the 4-bet.
And here is the awkward part: a 4-bet range made only of AA, KK, QQ and AK looks perfectly normal in the numbers. The section on sample size does the arithmetic — the healthy skeleton and the pure-value range are 4 combinations apart, and no sample can separate them. This is not a problem you will ever spot by staring at a percentage.
The 3-bet post covered the third button when you face a single raise. The fold-to-3-bet post covered how often you simply pay up when it comes back at you. This one is about the road through that same spot where you neither folded nor called.
There is one structural difference from the rest of the series, and it is worth stating up front: this number has no target band. Not “I could not find it” — RushCraft genuinely does not give it one. No target interval, no diagnostic card, no trend rule. The second section is about what that blank actually means, and the section on sample size shows you the arithmetic that makes it the right call.
So the weight of this post is not on what your 4-bet percentage should be. It is on which hands belong in the range.
What It Counts: Facing Two Raises
In one sentence: when the action reaches you, you are facing exactly two raises, and the last of them was not yours.
That is the condition RushCraft applies when it parses your hand history (raises_faced == 2, and you are not the last raiser); the numerator is “you raised”. The definition is broader than most people assume:
| Situation | In the denominator? | Why |
|---|---|---|
| You open from CO, the big blind 3-bets you | Yes | The standard 4-bet spot, and the only one with a practice drill |
| UTG opens, CO 3-bets, action reaches you on the button | Yes | You have not put in a chip yet. This is a cold 4-bet |
| You 3-bet, he 4-bets, and it is on you again | No | You are facing three raises. That would be a 5-bet |
| Somebody opens and you are behind him | No | One raise. That is 3-bet territory |
| Somebody opens, somebody calls, action reaches you | No | A call is not a raise. You are still facing one |
The second row deserves a pause. This number does not only count the hands where you opened and got raised back; it also counts “somebody else opened, somebody else 3-bet, and here you are”. You fold almost all of those, but they are in the denominator — which is why the percentage on your screen sits lower than “how often I 4-bet after opening and facing a 3-bet”.
One more technical point, useful whenever you compare the two: 4-bet and call 3-bet share a denominator. Both are recorded off the same condition in the same place in the parser, so the two percentages plus your fold rate divide one whole into three pieces.
That is not true of fold to 3-bet, which requires that you raised and therefore runs on a narrower denominator — which is why it never adds up tidily with call 3-bet. Two denominators live inside one set of concepts. Check which one you are looking at before you compare anything.
The Blank: No Target Band, No Diagnostic Card
Every statistic covered on this site so far comes with a line: VPIP 18-25%, 3-bet 8-12%, cold call 5% or below, fold to 3-bet 65% or below.
4-bet has none. It is absent from the target-band table, no diagnostic card ever flags it high or low, and the AI coach’s leak explanation library has no entry for it.
It is computed all the same, and it surfaces in three places — all three of which show you your own play:
| Where | What you get |
|---|---|
| The detailed stats table | Your overall 4-bet % |
| Each seat on the table diagram | Your 4-bet % from that position |
| The 13x13 hand matrix tooltip | Your 4-bet % for that one hand |
And there is one place where it is conspicuously missing: the opponent profile window. That window lists nine numbers — VPIP, PFR, 3-Bet, F-Steal, F-3B, CBet, F-CBet, WTSD, Call 3B — and 4-bet is not among them. The slot is empty.
That blank is not an oversight. A number you need tens of thousands of hands to read reliably, read off the sample you accumulate on one opponent in one session, produces a random figure that looks precise. Better to withhold it than to publish something misleading. The section on sample size does the arithmetic on that “tens of thousands”.
So the right use of this statistic has to be adjusted from the start: it is not there to be compared against a target band, and it is not even there to tell you how often you press the button — the sample-size section works out that it cannot even tell whether you bluff. Its only real use is to walk you to the next question: what is actually inside the range. That is what the next three sections are for.
The Skeleton: QQ+, AKs, AKo, A5s
When RushCraft’s fold-to-3-bet drill scores your button press, the default 4-bet range it judges against is:
QQ+, AKs, AKo, A5s
Broken out, that is 38 combinations out of all 1,326, or 2.9% of starting hands:
| Hand | Combinations | Share of all starting hands |
|---|---|---|
| QQ, KK, AA | 18 | 1.4% |
| AKo | 12 | 0.9% |
| AKs | 4 | 0.3% |
| A5s | 4 | 0.3% |
| Total | 38 | 2.9% |
Three things to say about it before the interesting part.
One: the range does not vary by position. From UTG to the small blind, against a 3-bet from anyone, it is the same table. That is a first-pass approximation, stated as such in the source — the calling range in the same drill does split by seat (BTN widest at 2.7%, SB narrowest at 1.2%); the 4-bet range does not. In practice a 4-bet in position can run a little wider, for exactly the reasons position matters everywhere else, but that is not what this table currently does.
Two: it is a value skeleton, not a complete strategy. Same as with the opening and calling ranges: the table tells you which hands qualify for the action, not “press this and you are done”.
Three: 4-bet wins ties. A hand appearing in both the 4-bet table and the calling table is scored as a 4-bet. As it happens the two tables do not currently overlap, so that rule is a safety net rather than a live fork.
The question worth asking is about the last row: the other three are hands that make your pulse jump. What is A5s doing sitting with them?
That A5s: Bluffs Come Out of the Fold Pile
Start with the negative case — which is the case this post opened with: what happens when the table is nothing but QQ+ and AK.
What happens is what the 3-bet post described, moved up one level: your 4-bet becomes a face-up card. He sees you 4-bet and he knows you have QQ+ or AK — so he does two things, and both of them cost you money.
He folds JJ, TT and AQ instantly. Those hands are genuinely awkward against a 4-bet range that contains bluffs. Against a pure value range they take one second to muck. You have saved him his thinking time, and you have saved yourself your income.
He only puts money in when you are behind. You 4-bet with queens, he shoves aces or kings — are you calling? You 4-bet AK, he shoves, and you are holding a hand with a bit over 30% equity against his range. A transparent value range only ever gets called in the worst case. You think the 4-bet is protecting your big hands. What it is actually doing is guaranteeing they only get action when they should be folding.
A5s is there to stop that. It does two jobs:
First, the ace in your hand cuts down his strongest combinations. The hands he will never fold to a 4-bet are AA and AK — hold an ace and his AA drops from 6 combinations to 3, his AK from 16 to 12. You are not guessing whether he has it. You are reducing what he can have.
Second, it still has a life after being called. A5s makes flushes, makes the wheel (A-2-3-4-5), and comes with an ace kicker when it pairs. It is not a bluff that has to give up the moment somebody calls.
So why not 87s, or the more natural-feeling AJs?
87s does neither job. It blocks none of his strong hands, and once called it is a small suited connector against AA/KK/AK, where nothing it makes is quite enough.
AJs is the more interesting case, because it is the hand most people would pick. The problem is that AJs is already in the calling range (the default continuing range is JJ-TT, AQs, AJs, KQs). Move it into the 4-bet and you are mining your own assets for a bluff — your calling range gets weaker, and the blocker effect you bought is roughly what A5s gives you anyway.
A5s is different: it was in the fold pile. Using it as a 4-bet costs nothing at all.
Bluffs come out of the fold pile, not out of your assets. That sentence was about A5s in the 3-bet post, and it is still about A5s here, because the logic did not change.
One clarification so this is not misread: A5s is not there to lift your continuation frequency. The fold-to-3-bet post worked out that to make his pure bluffs unprofitable you must continue with a third of your opening range — open 46.9% on the button and that is 15% of all starting hands. A5s is 0.3%. It does not fill that hole. The hole gets filled by opening tighter or calling wider; A5s’s job is to keep your value hands from being read, and the two jobs should not be blurred together.
Play the Default Ranges and Your 4-Bet Is What, Exactly
The last section was about what goes in the range. This one converts it into the percentage you will actually see on screen.
The fold-to-3-bet post did the same exercise and concluded that the entire spread in the fold rate came from the opening range. This is the same mechanism seen from the other side — the 4-bet range never moves a single square, and your 4-bet percentage still varies by a factor of four across seats:
| Where you opened | Opening range | 4-bet hands inside it | Your 4-bet % |
|---|---|---|---|
| UTG | 11.2% | 2.6% (no A5s) | 23% |
| MP | 18.9% | 2.9% | 15% |
| CO | 29.1% | 2.9% | 10% |
| BTN | 46.9% | 2.9% | 6% |
| SB | 46.9% | 2.9% | 6% |
(The UTG row differs because the default UTG opening range stops at ATs+, so you never hold A5s there in the first place — from UTG you have no bluff 4-bet available. That is the position doing it, not a choice you made.)
Weighted by how often each seat opens, those five rows average about 9%. The figure on screen will be lower still, because as the first section covered, the denominator also carries a batch of cold 4-bet spots, and playing the skeleton you fold nearly all of them.
The use of this table is not to check your answer against it. It is to understand one thing: the same player, the same 4-bet table, and a change of seat moves the number by a factor of four. Which means “my 4-bet is 7%, is that normal” is unanswerable until you say which position, and how much of the denominator is cold 4-bet spots.
That is the first reason there is no target band: a total that gets dragged around by your distribution of seats would only generate false verdicts if you drew a line through it. The second reason is in the sample-size section, and it is the harder of the two.
The Bluff 4-Bet Arithmetic: 60%, Cheaper Than a 3-Bet
The 3-bet post worked out that a pure bluff 3-bet breaks even at 69%. Run the same arithmetic on a 4-bet and the answer comes out lower.
Using the sizing this site uses throughout (the table in the preflop guide: a 4-bet is about 2.2 to 2.5 times the 3-bet):
You open to 2.5bb from the cutoff, the big blind 3-bets to 10bb, you 4-bet to 24bb.
- You are putting in an extra 24 − 2.5 = 21.5bb
- The pot you are attacking is 2.5 + 10 + 0.5 = 13bb
- Break-even fold equity = 21.5 ÷ (21.5 + 13) ≈ 62%
(At 2.2x it is 60%. The full version of this “how often does it have to work” arithmetic is in how to work out pot odds.)
60% to 62%, seven points below the 69% a bluff 3-bet needs. The reason is direct: he built that pot. His 3-bet took it from 4bb to 13bb, so the prize grew while the amount you have to add did not grow in proportion. The harder somebody swings, the cheaper your bluff gets — the least-used property of this action, and the most counter-intuitive.
How often will he actually fold? Put the two ranges side by side:
| His 3-bet range | He folds roughly | Your pure bluff 4-bet |
|---|---|---|
| 3% (QQ+ and AK only) | Almost never | A pure loss. Do not do it once |
| 8% to 12% (normal) | Six to seven times in ten | Break-even to a small profit |
| 13% and up (he is stealing) | Seven in ten and up | Clearly profitable |
The middle row landing right around the break-even point is not a coincidence — a healthy 3-bet range is supposed to make an opponent’s bluff 4-bets unprofitable. The two sides price each other.
The rows above and below are what you are hunting for. This is the reverse face of the one defence in the fold-to-3-bet post: there, “when he only 3-bets 3%, folding 85% costs you nothing”; here, the same fact says bluff 4-betting a man who only raises QQ+ and AK is handing him money directly. Turn it around and a 13% 3-bettor has a range that is mostly hands which cannot stand a 4-bet — that is where the A5s starts earning.
Both judgements need one input: his 3-bet number. Note: not his 4-bet number. What you are reading is how wide he re-raises, and that statistic accumulates 35 opportunities per 100 hands, ten times faster. To judge whether your bluff 4-bet is worth it, look at his 3-bet, not his 4-bet.
(What if he 5-bets all in? That is a separate question, and the product’s default ranges stop at the 4-bet — practice mode has no fifth raise. In practice the range of people who 5-bet at low stakes is startlingly narrow, and A5s’s answer there is usually to fold, not to stand its ground.)
Sample Size: The Slowest Number on the Site, Too Slow to Deserve Tuning
Now to finish off that blank from earlier.
How the chances accumulate: for you to face two raises, somebody has to open, somebody has to 3-bet, and the action has to reach you while you are still in the hand. The fold-to-3-bet post worked out that “you open and get 3-bet” happens 2 to 3 times per 100 hands; the 4-bet denominator adds the cold 4-bet spots on top, so it runs slightly higher — about 3 to 4 opportunities per 100 hands.
Then comes the second problem, which is the deadlier one: the base rate is low. The lower the base rate, the less precision a given sample buys. Assume a true 4-bet of 9%:
| Opportunities | Roughly how many hands | Observed number lands at | What that interval is good for |
|---|---|---|---|
| 30 | ~900 | 0% – 19% | Telling zero from non-zero, nothing else |
| 100 | ~3,000 | 3% – 15% | Still cannot separate low from high |
| 400 | ~11,000 | 6% – 12% | Barely enough to bracket |
| 1,500 | ~43,000 | 7.5% – 10.5% | Only now does tuning mean anything |
(Normal approximation of the 95% interval; the orders of magnitude are right.)
The third row is the point: eleven thousand hands in, the interval still runs from 6% to 12% — from “value only” to “twice as wide as the skeleton”. Every value in that range corresponds to a completely different way of playing, and you cannot tell them apart.
Set it against the rest of the series:
| Statistic | Opportunities per 100 hands | Hands to separate high from low |
|---|---|---|
| 3-bet | about 35 | around 3,000 |
| Cold call | about 20 | around 5,000 |
| Fold to 3-bet | 2 to 3 | around 13,000 |
| 4-bet | 3 to 4 | 40,000 and up |
More opportunities than fold to 3-bet and slower convergence anyway — the difference is the base rate: something that happens 9% of the time is far harder to measure than something that happens 65% of the time.
Which makes the line nobody drew the correct decision. A target band of, say, 6-10% would do nothing but flag a crowd of correctly-playing people as leaking at ten thousand hands.
So what is the number good for? Start with what it cannot do, because that is exactly the illness this post opened with:
It cannot tell whether you bluff. Of the 38 combinations in the skeleton, A5s is 4 of them — take it out and a pure-value 4-bet range still holds 34, or 89% of the skeleton. On screen that moves the percentage from roughly 9% to roughly 8%. The entire difference disappears inside the narrowest interval in the table above with room to spare.
A player who only ever 4-bets AA, KK, QQ and AK produces the same number as a player doing it well. Which is what the title means by measuring a range and not a frequency: the thing you are looking for is not in this percentage. It was never being counted.
So what is left? Two things, both coarse:
- A 4-bet of zero: no sample required, that is already the conclusion. You are only ever calling with queens and ace-king, and facing a 3-bet you have two buttons instead of three
- Clearly high, fifteen and up: check the seats first (the table above; UTG is 23% by construction), and only once that is ruled out start suspecting you are 4-betting carelessly
The brackets stop there. Across the whole middle of the range — six percent to the middle teens — this statistic has no opinion, nor should it, and that is where most players live.
There is exactly one place those 4 combinations become visible: the per-hand tooltip in the 13x13 matrix, showing which hands your 4-bets are actually made of. Whether the A5s square is empty or has a number in it is a far more honest question than the percentage — that view shows a range, and the range is the part you can genuinely change.
If You Play Rush & Cash
On fast-fold tables folding moves you to a new table, and the effect on this number runs both ways, more sharply than for any other statistic.
The bad direction: the whole of the bluff arithmetic rests on knowing how wide he 3-bets, and on a fast-fold table you never play ten hands against the same person. The condition cannot be verified, so your bluff 4-bet has nothing to stand on. The sensible response is to lean on the skeleton — play QQ+, AKs, AKo, A5s, and save the A5s for the times you have actually seen somebody playing very wide.
The good direction: the biggest cost of a pure value 4-bet is being read, and being read requires an opponent who has accumulated data on you. On a fast-fold table nobody is building your profile. That is the same effect the 3-bet post described, only more extreme here — a 4-bet is one of the rarest actions in the game, and even at a regular table somebody has to sit beside you a very long time to see you make several.
Put plainly: “he will notice I only 4-bet aces and kings” does not happen on a fast-fold table. But do not take that as permission to ignore the range — of the two costs in the A5s section, “he only puts money in when you are behind” has nothing to do with whether he knows you. That one is a property of the range itself.
As for the sample, this is the one piece of good news: at 3 to 4 opportunities per 100 hands, fast-fold speed is the most practical way to reach a number that shows you anything. Export your hand history from PokerCraft first (six steps, see downloading your Natural8 hand history); import it into RushCraft and your 4-bet, your call 3-bet and how both distribute across positions are already computed.
A 4-bet comes down to one sentence: when somebody raises you back, it is the option left over after calling and folding.
- The denominator: facing exactly two raises, cold 4-bet spots included. It is shared with call 3-bet, so the two plus your folds make one whole
- The blank: no target band, no diagnostic card, and no place in the opponent profile window. That is not an oversight — this number cannot support a line
- The skeleton: QQ+, AKs, AKo, A5s. 38 combinations, 2.9%, and the same table from every seat (a first-pass approximation)
- That A5s: it removes half his AA combinations and a quarter of his AK, and it still makes wheels and flushes when called. Pick it over AJs because AJs is already in your calling range — bluffs come out of the fold pile
- The cost of a pure value range: he mucks JJ and AQ in a second, then only puts money in when you are behind
- That 62%: a bluff 4-bet breaks even below the 69% a bluff 3-bet needs, because he built the pot. It only holds against opponents who 3-bet wide; against a QQ+-only player it is a pure loss
- The sample: 3 to 4 opportunities per 100 hands at a 9% base rate, and forty thousand hands to tell high from low. On top of that it cannot tell whether you bluff — a pure-value range is 4 combinations short of the skeleton, under a point of difference in the percentage. Only the matrix shows it
Next time somebody 3-bets you and you are holding A5s, do not reach straight for the fold pile. That hand’s job in this spot is not to win. It is to make sure somebody pays off your aces.