How to compare different sample windows

You see Receiver A averaging 25 yards and Receiver B averaging 20. A looks ahead. Then you notice A's number uses the latest four appearances, while B's uses six. You haven't compared the same slice of their records.

A sample window tells you which observations were selected. “Last four games,” “last thirty days” and “whole season” can select different records. Even the same phrase can count team games or player appearances differently.

Different windows aren't automatically wrong. One might be useful for studying a recent role change, another for describing a longer history. The problem is presenting their difference as if the window choice doesn't matter.

In this invented example, both players’ latest four appearances cover the same dates. A’s receiving totals are 22, 26, 18 and 34 yards. That is 100 yards in four appearances, or 25 per appearance.

B's latest four are 28, 28, 28 and 28: 112 yards, averaging 28. Their two earlier appearances had four yards each. Include those earlier eight yards and the six-appearance total becomes 120, averaging 20.

Now the original comparison makes sense. B's 20 came from a longer window that included two low-output appearances. Across the same recent four, B averages 28 and A averages 25. The ranking reverses because the question changed.

B's longer average isn't an error. The earlier games belong in the six-appearance question. The recent four-game value simply describes a narrower part of the same history.

Don't silently replace the longer value and imply it was incorrect. Name both measures if both are useful: 20 across six appearances, 28 across the latest four. Then explain why the recent window matters to your question.

If the source says “last four team games,” check whether the player appeared in all four. An inactive game and an unavailable game record are different from a recorded zero-yard performance. The window label alone may not settle that detail.

After matching the window, read the units. Receiving yards per appearance isn't the same as receiving yards per target. In baseball, a pitch-specific contact value isn't automatically comparable with a full offensive metric that includes walks and strikeouts.

Check the source and filters too. Two providers may use different definitions. A postseason-only window can exclude games that a recent-game window includes. Keep those choices attached to the numbers.

A clear comparison names the measurement, window and count for both sides. When those differ, say why. The goal is to explain a useful difference, not create one by mixing incompatible records.