How to read a sports research table

You open a matchup report. One number is huge, another cell is green, and a rank looks promising. Where do you start? With the small words around the number.

Take the 280 in the card above. It could mean total yards, yards per game, pass attempts or something else entirely. The labels turn it into a clear statement: Harbor's offense averaged 280 passing yards across its last four completed games.

That gives you a team, a measurement and a window. Now you know what you're looking at. It describes those four games; the next game hasn't happened yet.

Use the same approach in any sport. Read “shots on goal per game” as a whole phrase. “Shots,” “shots on goal” and “goals” count different things. A color can help you find a cell, but the label tells you what it means.

An average folds several games into one number. That's handy for a quick comparison, but it can hide how different those games were.

Here are Harbor's four passing totals. Add 140, 220, 260 and 500: that's 1,120 yards. Divide by four games and you get the 280 on the report.

Now add Valley's defense. It allowed 250 passing yards per game over its own last four games. Harbor's row measures yards produced. Valley's row measures yards allowed.

Both use yards per game, so you can read them side by side. But they faced different opponents. You're comparing two sets of past games, not watching the teams play each other.

The ranks make the matchup look interesting: Harbor is fifth for passing yards produced; Valley is 27th for passing yards allowed. Here's the detail that makes those ranks useful: first means most yards for offense, but fewest allowed for defense.

A team sample and a player sample aren't always the same. Imagine a Harbor receiver who played in only two of those four games and finished with 60 receiving yards in total.

If you want their average when they played, divide by two appearances: 30 yards. If you spread their contribution across all four team games, divide by four: 15 yards. Same 60 yards, different question.

You now have enough to say something helpful: “Harbor produced 280 passing yards per game and ranked fifth over four games. Valley allowed 250 and ranked 27th over its own four games. That gives me a reason to look closer at Harbor's passing matchup.”

The next thing to inspect is how those yards happened. Harbor's 500-yard game lifts the average. Was it a different opponent, more pass attempts, or several long completions? The game belongs in the sample; looking at it helps you understand the sample.

Then check who will actually play and whether the report is current. Past production is easier to use when the current team still resembles the team in those games.

Avoid turning the two averages into a made-up forecast. Splitting 280 and 250 to get 265 is easy arithmetic, but the table gives you no reason to expect 265 next game. And fifth plus 280 isn't two pieces of independent support: that rank is based on the same yardage measurement.