Seven out of ten isn’t a 70% prediction

“Seven of the last ten” is a description of ten matches that already happened. It is not, by itself, a calibrated chance of the next result.

What the number actually says

In Newcastle’s saved sample before the Leeds game on 14 September, both teams scored in seven of ten eligible league matches. The displayed 70% comes directly from seven divided by ten.

That calculation does not account for the exact strength of the next opponent, the confirmed lineup, tactical changes or the price of a bet. A historical rate can be useful evidence without doing all of those jobs.

Ten games can move quickly

In a ten-match sample, one result is ten percentage points. When a new match enters and the oldest one leaves, a single replacement can change seven of ten to six of ten. That movement need not mean the team suddenly became a different side.

Dates matter too. Ten league games might cross a summer break, a transfer window or a managerial change. A larger-looking percentage does not remove those questions.

Two teams do not always mean twenty matches

If both teams played each other within the selected histories, that match may appear in both records. A combined sample should count the underlying match once, not twice. FootyTotals pools distinct eligible matches for that reason.

The Leeds–Newcastle all-games example happens to contain 20 distinct matches, with BTTS in 13. Other fixtures may have fewer distinct results even when both individual cards show ten.

Read the count with the percentage

Use the percentage to compare patterns quickly, then check the numerator, denominator and actual results. Keep disagreements between all-games and venue-specific form visible. The aim is to understand the evidence more clearly, rather than make the number carry more certainty than it contains.