The Maths Behind Spaceman: Probability, Expected Value and Variance
Spaceman looks like a game of timing, but the numbers underneath are simple. This article walks through the standard crash-game model: how likely a round is to reach a given multiplier, what a bet is worth on average, and why results swing so widely around that average.
An important caveat first. Pragmatic Play does not publish the exact crash-point distribution of Spaceman. What it states is the RTP of 96.5% and the [5,000x](/spaceman/spaceman-max-win) maximum. Everything below uses the standard textbook model for crash games, in which the chance of reaching a multiplier is the RTP divided by that multiplier. It is the model this site uses for its own calculations. The real curve may differ in detail, so treat the figures as a clear approximation, not as a statement from the developer. The RTP shown in the rules screen of your casino is the figure that applies. For more on that number, see the RTP guide.
The Core Model
Probability of reaching a multiplier
In the standard model, the probability that a round reaches at least multiplier m is:
P(reach m) = RTP / m = 0.965 / m
Two checks show the formula is sensible:
- At m = 1, P = 0.965. That means 3.5% of rounds would end before anyone can cash out, which is exactly the house edge showing up at the lowest point of the curve.
- Each time you double the target, the probability halves. Reaching 4x is half as likely as reaching 2x.
Why every target has the same expected value
Suppose you stake 1 unit and plan to cash out at multiplier m. You receive m units back with probability 0.965 / m, and nothing otherwise:
EV of the payout = (0.965 / m) x m = 0.965
The m cancels out. On every unit staked you get back 0.965 on average, so you lose 0.035 per unit, or 3.5 per 100 staked, whatever target you pick. That is simply the house edge (100% - 96.5%). Because each part of a partial cash-out is a bet with its own target, splitting a stake across two targets does not change this either.
Cash-Out Targets Compared
The table uses a 100-unit stake. "Expected loss" is the long-run average loss per round. "Standard deviation" measures how far a single round typically lands from that average, and it is computed from the model as 100 x sqrt(0.965 x m - 0.965^2).
| Target | Chance of reaching it | Roughly 1 in | Expected loss per 100 | Standard deviation per 100 |
|---|---|---|---|---|
| 1.1x | 87.7% | 1.1 | 3.5 | 36 |
| 1.5x | 64.3% | 1.6 | 3.5 | 72 |
| 2x | 48.25% | 2.1 | 3.5 | 100 |
| 3x | 32.2% | 3.1 | 3.5 | 140 |
| 5x | 19.3% | 5.2 | 3.5 | 197 |
| 10x | 9.65% | 10.4 | 3.5 | 295 |
| 20x | 4.83% | 20.7 | 3.5 | 429 |
| 50x | 1.93% | 51.8 | 3.5 | 688 |
| 100x | 0.965% | 103.6 | 3.5 | 978 |
| 1,000x | 0.0965% | 1,036 | 3.5 | 3,105 |
| 5,000x | 0.0193% | 5,181 | 3.5 | 6,946 |
Worked example at 2x: you win 100 with probability 0.4825 and lose 100 with probability 0.5175, so EV = 48.25 - 51.75 = -3.5. At 10x: you win 900 with probability 0.0965 and lose 100 with probability 0.9035, so EV = 86.85 - 90.35 = -3.5. Same loss, very different experience.
Variance: Same Average, Very Different Ride
What the standard deviation tells you
The expected loss is identical for every target, but the spread is not. A low target such as 1.1x wins most rounds in small amounts and occasionally loses a full stake. A high target loses almost every round and now and then pays a large amount.
Over n rounds, the expected loss grows in proportion to n, while the standard deviation grows only with the square root of n. That has a useful consequence, using a normal approximation and 10,000 rounds of 100 units:
- At 2x: expected loss 35,000, standard deviation about 9,990. Finishing ahead needs a result 3.5 standard deviations above average, roughly a 0.02% chance.
- At 10x: expected loss 35,000, standard deviation about 29,500. Finishing ahead needs about 1.2 standard deviations, roughly a 12% chance.
This does not mean high targets are better. The expected loss is the same, and the chance of ending ahead comes with the risk of a much larger deficit. It means that luck dominates short sessions, and the 3.5% edge only becomes visible over a very large number of rounds.
Losing streaks
At a 2x target the chance of missing is 51.75% per round. Rounds are independent, so streaks follow simple powers:
| Consecutive misses at 2x | Probability | Roughly 1 in |
|---|---|---|
| 3 | 13.9% | 7 |
| 5 | 3.7% | 27 |
| 10 | 0.14% | 726 |
Streaks of this length are normal, not evidence that something is broken. For how this feeds into staking plans, see the bankroll guide.
The 5,000x Cap
Spaceman's maximum multiplier is 5,000x. In the standard model the chance of a round reaching 5,000x is 0.965 / 5,000 = 0.0193%, or about 1 in 5,181 rounds. That is why the cap barely affects the maths: only a tiny slice of the probability sits above it.
If the developer truncates the curve at the cap, the expected value of any target at or below 5,000x is unchanged in this model, and targets above it do not exist. We cannot confirm how the real game treats the top end, so do not read the table's last row as a promise of how often 5,000x occurs. Our maximum win article looks at what the cap means in practice.
Independence and the Gambler's Fallacy
Each round is generated independently by a certified random number generator. The probabilities above apply fresh to every round:
- After ten rounds below 2x, the next round still has the same chance of reaching 2x. It is not "due".
- After five wins in a row, the chance of the next win has not changed.
- Patterns in the history chart are what random data looks like to a human eye. They carry no information about the next round.
This is also why tools that claim to forecast rounds cannot work.
What the maths does not allow
- No cash-out target has a better expected value than another.
- No staking pattern turns a negative expected value positive. Progression systems change the shape of results, not the average.
- Past rounds give no information about future rounds.
Practical takeaways
- Treat any stake as the price of entertainment, with an expected cost of 3.5 per 100 wagered.
- Choose a target by how much variance you can tolerate, not because it seems smarter.
- Decide a budget before playing and stop when it is gone.
Use Our Spaceman Analytics Tools
Analyze Spaceman data with our live statistics, distribution analysis, trend charts, and probability calculator. All tools are free and require no registration.
Related Guides
Game Guides:
Strategy & Analysis:
Scam Warnings:
Platform Guides:
Responsible gambling: This article is educational and is not financial or betting advice. Gambling is only for adults where it is legal, and only with money you can afford to lose. If gambling is causing you harm, contact BeGambleAware or Gamblers Anonymous.