Why a Small Sample Misleads in Sports
A few games are a noisy sample. A hot week can appear without any change in skill, and it does not set a price.
Sports produce a number every night, and the night is a small sample of a longer skill. People promote the night anyway, because they watched it. A batter who goes 5-for-5, a team that wins four in a row, a goalkeeper who stops every shot in one match: each of these can be real as a description of that sample and useless as a new estimate of the person. The size of the sample is the part the highlight leaves off.
What small means in trials
Five at-bats are five trials. Four games are four trials of a team result that itself hides hundreds of smaller plays. A percentage built on five trials will jump. Three hits in five at-bats is .600. One hit in the next five is .200. The two weeks can belong to the same batter. What a batting average measures is the fraction. The sample-size point is that the fraction needs a denominator you would not be embarrassed to say out loud.
A useful habit is to refuse a rate when you cannot also state the count. .600 on 5 at-bats and .600 on 200 at-bats are not the same sentence. The first is a short story. The second is a long record. Variance in a short session is the same idea away from sports: the cloud is wide when the count is small.
Why the eye picks the extreme
Leagues contain many players and many weeks. Extreme weeks must occur somewhere. The broadcast finds them, because an ordinary 1-for-4 is not a segment. If you only consume the segments, you will believe extreme weeks are the normal state of good players. They are the visible tail. Coincidences and the birthday problem explain why a striking match is likely once the number of quiet comparisons is large. A league is a factory for those comparisons.
The repair is boring on purpose. Look up the count. If the count is a weekend, file the number under weekend. Do not file it under identity.
Streaks are samples too
A winning streak is a count of games, not a substance that accumulates in the ball. Four wins can come from a strong team, from a soft schedule, or from a few close bounces. The streak itself does not say which. Treating the next game as owed, either as a continuation or as a snap-back, imports the gambler's fallacy into a sport. The next game is its own event, played by whoever is actually available, against a particular opponent.
You can still notice that a team has played well. You notice it with a large count and with the roster written beside it. A streak is an invitation to look at those, not a substitute for them.
No price is hiding in the sample
A small sample does not contain a betting recommendation. Markets, where they exist legally for adults, add a margin and a list of conditions this article does not price. The literacy task is prior to any of that: describe the sample honestly, or decline to make a claim. Most hot weeks, described honestly, are claims you should decline.
Anyone can practice this with a box score and no money. Minors especially should keep it there. Understanding a sample is not a step into a wager.
Frequently asked questions
How many games make a sample large enough?
Large enough for the question. A weekend cannot support a new career verdict. A season can support a cautious one. State the count either way.
Can a hot streak be skill?
It can include skill. The streak alone does not separate skill from a short run of ordinary variance. Use a longer record.
Should a small sample change a wager?
This article does not recommend wagers. A small sample should not be promoted into a confident claim of any kind, priced or not.