After winning two or three rounds in a row, players feel "hot", a sense that their judgment is connected to the game. It isn't. This article explains why winning streaks trigger overconfidence, why crash games make the hot-hand feel more real than in other games, and what distinguishes a statistically unusual streak from a normal one.
What is the hot hand, and where does the term come from?
The "hot hand" is the belief that a person who has experienced success in a random event has a greater chance of success in further attempts. The term comes from basketball, the idea that a player who has made several consecutive shots is "hot" and more likely to make the next one.
In 1985, Gilovich, Vallone and Tversky published a famous study arguing that the hot hand in basketball was largely an illusion: fans and players saw streaks in shooting records that looked consistent with independent shots. In 2018, Miller and Sanjurjo showed that the original analysis contained a subtle statistical bias, and that corrected, the data suggest a modest real hot hand in some shooting settings. The debate about basketball is therefore open again. It does not carry over to crash games, where there is no player skill in the outcome at all.
The same illusion operates in crash games, but with a crucial difference: in basketball, there is at least a plausible mechanism for a hot hand (confidence, muscle memory, defensive adjustments). In a crash game, there is no mechanism at all. The crash point is determined by a cryptographic hash or RNG before you place your bet. Your previous wins have exactly zero influence on the next round's outcome.
Yet the feeling persists. After cashing out at 5×, then 3×, then 8× in consecutive rounds, players can feel "in sync" with the game, a feeling that their timing is calibrated and their next decision will also be correct. This feeling is the hot-hand fallacy in its purest form.
Why do short winning streaks feel like skill even when they aren't?
Three cognitive mechanisms converge:
Confirmation bias. When you are winning, you attribute successes to your decisions ("I knew when to cash out") and dismiss failures as bad luck ("the game crashed too early"). This selective accounting creates a narrative of competence that is not supported by the data. If you tracked every decision objectively, wins and losses, correct calls and missed calls, the pattern would look random. But you are not tracking objectively. You are remembering the highlights.
The illusion of control. Crash games feel skill-based because you make a continuous decision during each round. Unlike a slot machine where you press spin and watch, you are actively choosing when to cash out. This active involvement creates a sense of agency, and agency creates a sense of control. When outcomes happen to align with your decisions, the control feels real. It is not: the crash point was fixed before you saw it. But the feeling of control is powerful and automatic.
Causal reasoning. Human brains are causal machines. We see event A followed by event B and infer that A caused B. Three wins in a row triggers the inference: something I am doing is causing me to win. In reality, the only causal factor is the game's RNG, which operates independently of your actions. But "randomness" is an unsatisfying cause. The brain prefers a narrative, momentum, timing, intuition, and manufactures one.
Are long winning streaks ever real in crash games?
Yes, in the trivial sense that they actually occur. No, in the sense that they carry predictive information.
Winning streaks are a natural property of independent random events. If you cash out at 2.0× on every round in a game designed at 97%, you win about 48.5% of the time (p = 0.485). Two different numbers are easy to confuse:
- The chance that a given window is a streak: p^k. For five particular rounds, 0.485^5, about 2.7%.
- How many streaks a session contains: much more than the window chance suggests, because a 100-round session contains many windows.
For a 100-round session at p = 0.485:
| Run of at least | Chance a given window is all wins | Expected runs per 100 rounds | Chance of at least one |
|---|---|---|---|
| 2 wins | 23.5% | about 12 | nearly certain |
| 3 wins | 11.4% | about 5.8 | nearly certain |
| 5 wins | 2.7% | about 1.3 | about 77% |
| 7 wins | 0.63% | about 0.3 | about 27% |
| 10 wins | 0.07% | about 0.03 | about 3% |
The expected number of runs of at least k wins in n rounds is exactly p^k + (n − k) × (1 − p) × p^k. For five or more in 100 rounds at p = 0.485, that is about 1.34. Seeing one or two in a session is ordinary; seeing none happens in about 1 in 4 sessions. Neither is a signal. It is what randomness looks like.
The critical question is not "did a streak happen?" but "does this streak carry information about the next round?" In a crash game that works as designed, the answer should be no: games are designed so round N+1 is independent of rounds 1 through N, and the lab will test this. The streak is a description of what happened, not a prediction of what will happen.
What do crash game streak distributions actually look like?
The distribution of streak lengths in a fair crash game follows a geometric distribution. For a win probability p per round, the probability of a streak of exactly k wins followed by a loss is:
P(streak = k) = p^k × (1 - p)
This distribution is memoryless: the probability of the streak continuing is always p, regardless of how long the streak has been. A player who has won 10 rounds in a row has the same probability of winning the 11th as a player who just lost the previous round.
What this means in practice: if you plot the streak lengths from 10,000 rounds of a fair crash game, you will see many short streaks (1-2 wins), fewer medium streaks (3-5 wins), and rare long streaks (7+ wins). This is the exact pattern that the streak illusion misinterprets as evidence of cyclicality or reversion.
Comparing the streak lengths in a large count of rounds with this geometric distribution is a direct test of independence. If long streaks turned up far more or far less often than predicted, the rounds might not be independent, which would be a serious finding. In a game working as designed, they match within the range chance allows. The patterns and hot hours check explains the same point for clusters by time of day.
How do you check whether your streak is unusual or normal?
You cannot do this reliably by feel. Human intuition about probability is unreliable here: most of us read more meaning into streaks than they carry.
But you can do it with simple math. Here is the check:
- Define your win condition. For example: "I win if the crash point is above 2.0×."
- Estimate your per-round win probability. For a 97% RTP game at 2.0× cashout: approximately 48.5%.
- Compute the expected number of runs. For runs of at least k wins in n rounds, it is p^k + (n − k) × (1 − p) × p^k. For k = 5 and n = 100 at p = 0.485: about 1.3.
- Compare to your experience, over many sessions. One or two runs of five in 100 rounds is expected, and so is none. Only a large, persistent excess across thousands of rounds would be worth investigating, not because you are "hot," but because the rounds might not be independent.
The distinction is critical: an anomalous streak frequency is a signal about the game's fairness, not about your personal luck or skill. If streaks happen more often or less often than the geometric distribution predicts, the question is not "am I on a hot streak?" but "is this game fair?"
What's the right response to a winning streak?
The mathematically correct response to a winning streak is: nothing. Do not change your bet size. Do not change your cash-out target. Do not change your session plan. The streak carries no information about future rounds.
In practice, "do nothing" is harder than it sounds because the hot-hand feeling is powerful. Two pragmatic techniques:
The bet-size rule. Before the session, set a fixed bet size. Do not deviate, up or down, regardless of outcomes. This is the same technique recommended for chasing losses, and it works for the same reason: it removes the opportunity for cognitive biases to influence your bet sizing.
The gratitude check. When you are on a winning streak, pause and ask: "Am I grateful for what I have won, or am I focused on winning more?" If the answer is the latter, the hot hand is in control. Cash out, close the game and take a break. The streak's profits are real. The feeling that the streak will continue is not.
The hot-hand fallacy is the mirror image of the streak illusion. One tells you losing must stop. The other tells you winning must continue. Both are wrong in the same way: they attribute predictive meaning to a sequence of independent events. Understanding this, really understanding it, not just knowing it intellectually, is the single most important cognitive skill a crash game player can develop.
References
- Gilovich, T., Vallone, R., & Tversky, A. (1985). The hot hand in basketball: On the misperception of random sequences. Cognitive Psychology, 17(3), 295-314.
- Miller, J. B., & Sanjurjo, A. (2018). Surprised by the hot hand fallacy? A truth in the law of small numbers. Econometrica, 86(6), 2019-2047.
Was this page useful?
Anonymous: no account and no cookie.