Why NBA Playoff Series Usually Favor the Better Team
A best-of-seven series gives each team repeated chances to adjust. That makes it different from a single-elimination tournament, but it does not make an upset impossible or turn seeding into a complete forecast.
The useful question is not whether a seed is “safe.” It is how per-game win probability, home-court scheduling, injuries, and matchup adjustments combine across a series.
From One Game to a Series
If a team had the same independent win probability in every game, a series probability could be calculated from the possible paths to four wins. Real series are less tidy: venues alternate, players get hurt, coaches change rotations, and the result of one game can affect later assumptions.
That is why a seven-game model should expose its per-game probabilities instead of relying on an unsourced historical seed table. Seed is a useful summary of regular-season performance, not a substitute for a matchup forecast.
What Home Court Actually Changes
Under the NBA’s 2-2-1-1-1 format, the team assigned home-court advantage hosts Games 1, 2, 5, and 7 when needed. A series ends as soon as either team reaches four wins. The size of any home-court adjustment should come from a documented model or current market comparison; it should not be treated as a permanent universal number.
Build separate home and away probabilities, then stress-test the series result. If a small home adjustment radically changes the conclusion, the forecast is fragile and should be presented that way.
Matchup Inputs That Can Override Seed
- Availability and plausible minutes for high-usage players.
- Whether either team can create or prevent efficient shots.
- Turnover, rebounding, and foul-rate matchups.
- Rotation depth when games are close together.
- The range of adjustments each coach can make after a loss.
These inputs also explain why a lower seed can be stronger than its label suggests. Regular-season record may reflect injuries, schedule, or a roster that changed before the playoffs.
Keep the Forecast Auditable
Record the series price, your per-game assumptions, and the date of the analysis. After the series, separate a bad result from a bad process: an injury learned later is different from an availability risk ignored at decision time.
For reproducible constant-probability calculations and sensitivity examples, see the NBA Playoff Probability Guide. For single-elimination paths, see March Madness Odds.
Q: Why does a longer series generally help the stronger team?
A: Under an independent, constant-probability model, a team with a per-game win probability above 50% has a higher chance of winning a best-of-seven series than a single game. Real series can have changing probabilities; the model does not prove that a particular seed is stronger.
Q: Does the No. 1 seed always win?
A: No. Seeding omits matchup details, current availability, and uncertainty.
Q: How should home court enter a series model?
A: Use an explicit, sourced or clearly labeled assumption for home and away games, then show how sensitive the series probability is to that assumption.