The near-miss effect: why almost winning feels like information

What a near miss is

Stop a fruit machine one symbol short of the jackpot and something odd happens: you lost, exactly as completely as if the reels had shown nothing, yet it does not feel like nothing. It feels like a signal — the win is *around here somewhere*. That feeling has a name in the research literature, the near-miss effect, and it is one of the oldest documented quirks in the psychology of gambling.

B.F. Skinner remarked on it in the 1950s: a machine that stops just short of paying is, in his terms, reinforcing behavior without paying for it. The first systematic treatment came from R.L. Reid in 1986, who defined the near miss as a failure that is structurally close to a success — and observed that closeness of display is doing the psychological work, because in outcome terms there is no such thing as close. A lottery ticket one digit off pays the same as one that shares no digits: zero.

The definition matters because it separates two things the mind merges. In games of skill, nearly succeeding genuinely is information — a dart just off the bullseye says your aim is almost calibrated. The near-miss effect is that machinery firing in a domain where it has nothing to grip: a game of independent chance, where the previous spin’s geometry predicts the next one’s no better than a roulette streak predicts red.

What the research shows

Three findings recur across the literature, and together they explain why the effect is worth an article rather than a footnote.

• Near misses extend play. In controlled slot-machine experiments, players dealt losses dressed as near misses persist longer than players dealt plain losses. Kassinove and Schare’s 2001 study found the effect strongest when about 30% of losses were near misses — enough to keep the signal alive, not so many that it wore out.
• The brain treats them partly as wins. Clark and colleagues’ 2009 imaging study found near misses activating parts of the same reward circuitry that genuine wins do — while being rated, by the same participants, as *more unpleasant* than full misses. Losing by a hair feels worse and motivates more, at the same time.
• The effect needs agency to bite. The same research line found near misses most potent when the player had some involvement — choosing when to stop, picking the numbers. Illusory control and the near miss compound each other: it almost worked, *and I almost did it right*.

None of this means a person who feels the pull is broken. The effect appears in ordinary players in laboratory settings; it is the standard machinery of learning from near-success, operating on inputs it was never built for. The finding is not that some people misread near misses. It is that people do.

The anatomy of a tease, measured on our own reels

Our slot Hanabi launches its free-spin round when three or more 祭 scatter symbols land across fifteen cells. The near miss here has a precise shape: two 祭 on screen, the round over, the third conspicuous by absence. How often does that happen? On these reels, that is not a rhetorical question.

The point of the arithmetic is what it rules out. Nothing in the engine watches your spin, notices a loss and arranges two scatters to console you: each cell is drawn independently, and the whole grid replays from the committed seed pair — you can recompute any spin yourself. The tease happens six times as often as the trigger for the same reason two heads in three coin flips happen more often than three: it is the shape of independent chance, not choreography.

That distinction — near misses that fall out of the math versus near misses that are put there — is where the ethics of the subject actually live. Regulators in several jurisdictions have acted against machines engineered to *display* near misses more often than their odds produce them: reels weighted so the jackpot symbol crowds the payline’s neighbors, stopping routines that steer just short. A game whose teases are exactly as frequent as independence implies is playing fair with your dopamine; a game that manufactures extra ones is hiding a price signal in a feeling.

Living with a brain that keeps score this way

The near-miss effect does not switch off once named — imaging studies suggest the response persists even in players who know the odds. What knowledge changes is what you do with the feeling. Three habits follow directly from the research:

1. Treat “that was close” as a description of the picture, not the probability. The display was close; the outcome was a loss identical to any other. If a session budget was going to end the session, a near miss is not a reason it should not.
2. Notice the urge to continue *specifically after* a near miss. That timing is the effect’s signature. The stopping rules in our responsible-play toolkit work precisely because they are set before the feeling arrives.
3. Prefer games that show you their mathematics. A published paytable and a verifiable draw do not blunt the feeling, but they let you check that no one is manufacturing it.

And the honest limit, stated plainly: this article is science journalism, not clinical advice. If gambling has stopped feeling like entertainment priced in advance — if the near misses are doing the deciding — the resources below are free, confidential and staffed by people whose job this is.