Playoff Series Probability, Explained
After reading this you can turn a single-game win probability into the chance of winning a best-of-3, 5 or 7 series, read the full distribution of series lengths, and explain why longer series favor the stronger team.
What the calculator does
A playoff series is a race. In a best-of-7, the first team to win 4 games takes the series. The calculator starts from one number, your chance of winning a single game, and computes the chance of winning the whole series.
Here is the hook. Suppose you win 55% of individual games against an opponent. That sounds like a modest edge. In a best-of-7, your series win probability climbs to about 60.8%. The longer format stretches your small per-game edge into a larger series edge, because luck has more games to average out.
The same math answers the reverse question. A 45% underdog wins a single game almost half the time, but wins a best-of-7 only about 39.2% of the time. Short series protect underdogs; long series expose them.
When to use it, and when not to
Use this when you already have a defensible single-game win probability and the games are close to independent. A rating system such as the Elo Rating Calculator gives you exactly that number: an expected score between 0 and 1 that you can read as a win probability.
The model assumes each game is a fresh coin flip with the same probability. That assumption breaks in real sport in three common ways.
- Home advantage. Most series alternate venues, so your per-game probability is not constant. A single average number smooths this over.
- Momentum and adjustments. Coaches change tactics between games, and injuries accumulate. The model treats game 7 exactly like game 1.
- Draws. The formula assumes every game produces a winner. For sports with ties, resolve them first or use a different tool.
Do not feed this a probability you have not justified. A 5-point swing in the single-game number (say 55% versus 60%) moves the best-of-7 result from 60.8% to 71.0%. Garbage in, confident garbage out.
The formula and why it works
Let p be your single-game win probability and let the series be first to n wins (n = 4 for best-of-7). You win the series exactly when you reach n wins before the opponent does. The clean way to count this is the negative binomial: the probability that your n-th win happens in game number g.
Read it left to right. To clinch in game g, you must win that last game (one factor of p), and among the previous g-1 games you must have exactly n-1 wins and g-n losses. The binomial coefficient \binom{g-1}{n-1} counts the orderings of those earlier games, p^{n} collects your n wins, and (1-p)^{g-n} collects the losses.
The series can end anywhere from game n (a sweep) to game 2n-1 (all games needed). Sum over those to get the total series win probability.
Each term is the chance of clinching in game g. Add them and you have the chance of winning at all. The same sum, with terms kept separate, is the distribution of series lengths.
Worked example: 55% favorite in a best-of-7
These are the tool's defaults: win_pct = 55, format = bo7, so p = 0.55 and n = 4. The series can last 4, 5, 6 or 7 games.
- Win in 4 (sweep): \binom{3}{3} \, 0.55^{4} = 1 \times 0.09151 = 0.09151.
- Win in 5: \binom{4}{3} \, 0.55^{4} \, 0.45 = 4 \times 0.09151 \times 0.45 = 0.1647.
- Win in 6: \binom{5}{3} \, 0.55^{4} \, 0.45^{2} = 10 \times 0.09151 \times 0.2025 = 0.1853.
- Win in 7: \binom{6}{3} \, 0.55^{4} \, 0.45^{3} = 20 \times 0.09151 \times 0.09113 = 0.1668.
- Add them: 0.09151 + 0.1647 + 0.1853 + 0.1668 = 0.6083.
So a 55% single-game favorite wins the best-of-7 about 60.8% of the time. Your per-game edge of 5 percentage points became a series edge of almost 11 points.
The distribution of series lengths
The single win probability hides useful detail. The full distribution tells you how many games the series is likely to run, counting both the favorite's and underdog's clinch games. For the 55% case, add the mirror terms where the opponent clinches (swap p and 1-p).
| Games played | Favorite clinches | Underdog clinches | Total |
|---|---|---|---|
| 4 | 0.09151 | 0.04101 | 0.1325 |
| 5 | 0.1647 | 0.09021 | 0.2549 |
| 6 | 0.1853 | 0.1218 | 0.3071 |
| 7 | 0.1668 | 0.1218 | 0.2886 |
The four totals sum to 1.000, as they must. A 6-game series is the single most likely outcome at 30.7%. The expected number of games is the length-weighted average: 4(0.1325) + 5(0.2549) + 6(0.3071) + 7(0.2886) = 5.77 games. Even a real edge produces series that usually run long.
How format length amplifies the edge
Hold the single-game probability fixed and change the format. The favorite's series win chance grows as the series lengthens, because more games give the stronger team more chances to prove itself and fewer chances for a fluke to decide everything.
| Single game | Best of 3 | Best of 5 | Best of 7 |
|---|---|---|---|
| 0.50 | 0.5000 | 0.5000 | 0.5000 |
| 0.55 | 0.5748 | 0.5931 | 0.6083 |
| 0.60 | 0.6480 | 0.6826 | 0.7102 |
| 0.65 | 0.7183 | 0.7648 | 0.8002 |
Read the 0.60 row. A 60% favorite wins about 64.8% of best-of-3 series but 71.0% of best-of-7 series. Note also that a true coin flip stays at exactly 50% in every format. Length does not create an edge; it magnifies an edge that already exists.
Reading and interpreting the result
Three numbers matter, and they answer different questions.
- Series win probability
- The bottom line: your chance of winning the whole series. Use it for a fair betting price or a bracket forecast.
- Length distribution
- How long the series is likely to run. Useful for "will it go the distance" markets and for scheduling expectations.
- Expected number of games
- The average series length. At p = 0.55 it is 5.77 games; at a true 0.50 it peaks near 5.81, and at a lopsided 0.70 it drops to about 5.34 because sweeps get common.
One reassuring check: the favorite's and underdog's series probabilities always add to 1, because one of them must win. If your two numbers do not sum to 1, you made an arithmetic slip.
Common mistakes
Convert home and away probabilities into a single number honestly. If you win 62% at home and 48% away in a 2-2-1-1-1 format, weight by how often each venue occurs rather than taking a flat average of 55%. The venue mix changes with series length.
The most frequent error is treating the series probability as if it equals the single-game probability. It never does, except at exactly 0.50. A 60% game favorite is not a 60% series favorite; in a best-of-7 they are a 71.0% favorite.
The second error is assuming independence where it does not hold. If a team is far stronger at home, real series are more predictable than the constant-probability model suggests, because the schedule loads the favorite's games onto their strong venue. Treat the output as a baseline, not a guarantee.
The third error is over-trusting the input precision. Rating models carry error bars. If your Elo-derived probability could plausibly be 0.53 or 0.57, report a range for the series result (about 0.56 to 0.62 in a best-of-7), not a single decimal.
Related tools
Get the single-game input from a rating with the Elo Rating Calculator, or build it from scoring rates. For soccer, the Soccer Score Simulator (Poisson) turns two expected-goals numbers into win, draw and loss probabilities. For tennis, the Tennis Match Probability Calculator chains serve-point win rates up to a full match. To simulate a whole table rather than one matchup, use the League Season Simulator. For a single decisive shootout, the Penalty Shootout Probability Calculator handles conversion rates and sudden death.
Frequently asked questions
Why does a 60% favorite lose a best-of-7 so often?
Because 71.0% to win means 29.0% to lose. Best-of-7 rewards the favorite, but four games out of seven is still a modest sample. Roughly three times in ten, the underdog wins enough of those games to take the series.
Does home advantage change the answer a lot?
It can. This tool uses one constant probability. If your edge is concentrated at home and the schedule gives you more home games, your true series chance is higher than the flat-average input suggests. Model the venues separately if the split is large.
What single-game probability should I enter?
Use one you can defend from ratings or scoring data. An Elo expected score works directly. If you only have a hunch, remember that a 5-point change in the input moves the best-of-7 result by about 10 points, so guess carefully.
Can I model a first-to-N race that is not a standard playoff?
Yes. The custom option sets any target number of wins. First to 2, first to 10, the same negative binomial sum applies, from game n to game 2n-1.
What is the longest a best-of-7 can run, and how likely is game 7?
Seven games. At p = 0.55 the series reaches game 7 about 28.9% of the time. Closer matchups reach game 7 more often; at a true 0.50 it is about 31.3%.
Related tools
Elo Rating Calculator
Expected score and rating change for any two Elo ratings and K-factor.
Soccer Score Simulator (Poisson)
Win/draw/loss and most likely scorelines from two expected-goals numbers.
Badminton Match Probability Calculator
From serve-rally win % to game and match probability with the 30-point cap.
League Season Simulator
Monte Carlo the rest of the season: title, top-N and relegation odds from the current table.
Squash Match Probability Calculator
From serve-rally win % to game and match probability with PAR-11 scoring.
Table Tennis Match Probability Calculator
From serve-point win % to game, deuce and match probability for best-of-3/5/7.