"They've won six straight." "They can't buy a win." Streaks dominate sports conversation. But do they predict what happens next? The answer is: sometimes a little, often less than people think. This guide explains how to separate genuine form from normal randomness.
Randomness creates streaks
Even pure randomness produces streaks that look meaningful. If you flip a coin 100 times, you'll almost certainly see a run of five or more heads somewhere in the sequence. Nobody thinks the coin is "hot."
Sports work similarly. A team that wins 55% of its games will have plenty of winning and losing streaks over a season, just by chance.
Worked example: a team with a 55% chance in every game. The chance of winning five in a row starting from any given game is 0.55^5 ≈ 5%. But over an 82-game NBA season, there are many starting points, so a five-game winning streak at some point is very likely. Its fans will think the team got hot.
When form is real
Some streaks reflect real changes:
- A key player returns from injury.
- A new coach or system starts working.
- A young team improves as players develop.
- A roster change adds or removes talent.
- Schedule: a streak against weak opponents may say little about true strength.
The key question: is there an explanation beyond results? If a team's underlying performance, like scoring margin or chance creation, improved too, the form may be real. If results improved but the underlying numbers didn't, luck may be playing a big role.
Results vs process
Analysts separate results (wins, goals) from process (the underlying performance that drives results).
Examples of process stats:
- Soccer: shots and chance quality.
- Hockey: shot attempts and scoring chances.
- Basketball: shooting efficiency and turnovers.
- Football: yards per play and success rates.
- Baseball: quality of contact and strikeout/walk rates.
A team winning with poor process stats is often due to cool off. A team losing with strong process stats is often due to improve. This is sometimes called regression to the mean.
Close games and luck
Close games are close to coin flips. A team that wins many one-score or one-goal games in a row is often riding luck. Records in close games tend to even out over time.
Worked example: two teams each finish 10-6 in a season.
- Team A won its games by an average of 10 points and lost close games.
- Team B won six one-score games and was outscored overall.
Over the next season, Team A is more likely to remain strong. Team B's record overstated its quality.
How markets react to streaks
The market adjusts for streaks, sometimes a bit too much. Popular teams on long winning runs can become slightly overpriced because casual fans want to back them. Teams on long losing runs can become slightly underpriced.
This doesn't mean you should always fade streaks. It means you should ask whether the price has moved more than the underlying performance justifies.
The "hot hand" debate
For years, researchers argued over whether the "hot hand" in basketball shooting was real. Early studies said no; later work suggested small real effects once statistical biases were corrected. The takeaway for predictions: even if hot hands exist, the effect is probably smaller than fans believe, and it's hard to use in advance.
How to evaluate form
- Look at underlying performance, not just wins and losses.
- Consider the opponents. Strong results against weak teams mean less.
- Check for real changes: injuries, roster moves, coaching.
- Weight longer samples more than short ones. Recent games matter, but not as much as people think.
- Compare with the market. Has the price already adjusted?
Your own streaks
Streaks apply to your picks too. A winning streak doesn't mean you've cracked the code. A losing streak doesn't mean you're bad at this. Both are normal. Avoid raising stakes after wins or chasing after losses. See bankroll management and the gambler's fallacy.
Practice spotting real form
Next time a team goes on a streak, check whether its underlying stats improved. Make a free play-money pick on MockSport either way, and note your reasoning. Over a season, you'll learn how often streaks continue and how often they fade. Browse the odds page.
FAQ
Are winning streaks predictive?
Sometimes a little, especially when underlying performance improved. But many streaks are consistent with normal randomness.
What is regression to the mean?
It's the tendency for extreme results to be followed by more average ones. Teams with unusually good or bad luck usually drift back toward their true level.
Should I bet against teams on long winning streaks?
Not automatically. Check whether the price has moved more than the team's underlying performance justifies.