Every number here traces back to a column the league export actually contains, and each one shows its own inputs so you can check the work. Two things are deliberately missing, and it's worth knowing why:

There's no EPA. Real Expected Points Added needs down, distance and field position on every single play. The export is a box score — there are no plays in it. An "EPA" built from this data would be a made-up number wearing a famous name, and you'd be betting on it. ANY/A below is the real efficiency metric this data can support.

There's no target share. Madden reports catch percentage as 100% for everyone — it's counting catches over catches, not over targets. So targets are genuinely unknowable here. What you get instead is catch share and yard share, which are exact.

Every carry splits in two. The yards before contact are what the blocking handed him. The yards after contact are what he took himself. RYOE is that second half measured against a league-average back — the yards he created that an average back wouldn't have.

Why it matters for props: yards before contact belong to the offensive line and swing hard week to week. Yards after contact are the back, and they hold up. A rusher living on RYOE is a better bet to clear his line than one living on the blocking.

A quarterback's real yield per dropback, once touchdowns, interceptions and sacks are all accounted for.
The efficiency number that tracks winning most closely. A quarterback who piles up yards but throws picks and takes sacks will rank below a quieter one who does not.
It is a rate, not a volume: a backup with six good throws can post a high one.

Whether a quarterback completed more passes than he should have, given how disruptive the defences he faced have been.

Positive means he beat the coverage he was actually up against. It is how you tell a quarterback with a soft schedule from one who has earned his completion rate.

This is not true CPOE — that needs throw-by-throw tracking that Madden's export does not carry. Ours is an honest approximation, which is exactly why it is labelled rather than passed off as the real thing.

Splitting a receiver's yards the same way. Yards in the air are the throw and the route; yards after the catch are the player. Two receivers with identical yardage can be completely different bets — the YAC guy needs only a short throw to pay off.
Catch share is not target share. Targets aren't in the export. This is his share of the catches his team actually made.

The export tracks catches allowed per defender, which is real coverage data. Stop rate is deflections plus interceptions over everything thrown his way. A low stop rate on high volume is a defender opposing offenses have found.