FEI and Drive Efficiency: Rating College Football by Possessions, Not Plays

Most advanced football ratings grade the game one play at a time. That’s the world of success rate, EPA, and the play-level engine inside SP+ — and it’s a powerful way to see the sport. But there’s an older, complementary idea that asks a different question: forget the individual snaps; did the drive do its job? That is the premise of the Fremeau Efficiency Index (FEI) and the broader family of drive-based efficiency metrics, which evaluate each possession as a single unit.

The one-sentence version: FEI rates teams by what their possessions accomplish — points produced relative to expectation given where the drive started — opponent-adjusted, possession-based, and stripped of garbage time.

The drive as the unit of analysis

Play-level metrics treat the play as the atom of football; drive-level metrics treat the possession as the atom. The logic is that football is ultimately a game of drives that end in one of a few outcomes — a touchdown, a field goal, a punt, a turnover, a turnover on downs, or the end of a half. What you really want to know is how often a team turns possessions into points, and how often its defense denies the same.

FEI, developed by Brian Fremeau, builds its rating from exactly that. Its core ingredients are:

• Possession-based. The model evaluates each drive as a whole, not each play within it. A team that scores in four plays and a team that scores in twelve both produced one scoring drive.
• Value relative to starting field position. A drive that begins at the opponent’s 30 is expected to produce more points than one that begins at your own 10. FEI grades each drive against that expectation, so a touchdown from your own 5 is worth more credit than one handed to you by a short field.
• Opponent-adjusted. Like every serious rating, FEI recalibrates for schedule — scoring efficiently against elite defenses counts more than the same output against weak ones.
• Garbage time down-weighted. Drives that occur once the outcome is effectively decided are discounted, so a meaningless late touchdown doesn’t inflate a rating.

That last point is a theme across good modern metrics; we walk through how analysts even define and strip garbage time in our companion pieces. The result is a rating that answers a clean question: per possession, and against this schedule, how good is this team at turning drives into points and stopping the other team from doing so?

An illustrative drive: why field position is the whole game

Consider two touchdown drives, both ending in seven points. On a play-level scoreboard they look identical. On a drive-value scoreboard they are not. Here is a made-up example to show the mechanic:

The expected-point figures above are invented to illustrate the idea, not real values. But the point is real: the offense that drove 92 yards exceeded expectation by far more than the offense handed a 25-yard field, so a possession-based metric credits it more. Flip it to defense, and the unit that forced a punt after giving up a short field gets credit for limiting the damage. Counting only “points scored” would miss all of that; FEI is built to capture it.

How drive efficiency complements play-level metrics

Drive efficiency is not a competitor to play-level analysis — it’s the other half of the picture. Think of three lenses that fit together:

• Efficiency (play level) — how often you stay on schedule, measured by success rate and EPA. See our success rate and EPA explainer. This is the floor.
• Explosiveness (play level) — how much your good plays are worth, the ceiling. See explosiveness and yards per play.
• Drive efficiency (possession level) — whether all of that actually converts into points per possession, accounting for field position and finishing.

The connection runs both ways. A team can be efficient and explosive on a per-play basis yet leave points on the field — red-zone stalls, turnovers, missed kicks — and drive efficiency will expose that gap. Conversely, a team with mediocre per-play numbers but excellent field position and finishing can grade out better by possession than its play stats suggest. Reading FEI alongside SP+ — which leans more heavily on the play-level engine — gives you two honest, partly independent views of the same team.

Reading offense, defense, and special teams

Like most complete ratings, drive-efficiency systems split into units, and each tells you something specific:

• Offensive drive efficiency — points produced per possession relative to expectation. High values mean an offense that consistently converts drives into points, not just yards.
• Defensive drive efficiency — the same from the other side: points allowed per possession versus expectation. A strong defense forces empty possessions and makes opponents earn long fields.
• Special-teams drive efficiency — the often-ignored third unit. Special teams shape starting field position on nearly every drive (kickoffs, punts, returns) and convert short fields into points (field goals). Because drive value is so sensitive to field position, special teams matter more in a possession-based rating than fans usually assume.

Splitting the units this way lets you diagnose a team precisely. A squad with a great offensive rating but a weak special-teams number is winning the field-position battle the hard way; one with strong defense but poor offense is leaning on its stops to survive.

Caveats: small samples and field-position effects

Drive efficiency is a sharp tool, but it has honest limits — some of them the flip side of its strengths:

• Few drives per game means early noise. A team runs dozens of plays but only a handful of possessions per game. With a small number of drives early in the season, a couple of fluky outcomes (a tipped interception, a missed chip-shot field goal) move the rating more than they should. Be even more patient with drive metrics in September than with play-level ones.
• Field-position effects cut both ways. The model’s great virtue — accounting for starting field position — also makes it sensitive to how field position is earned. A team that constantly starts on short fields thanks to its defense and special teams may post gaudy offensive numbers; the adjustment helps, but interpreting the splits together still matters.
• It’s a model, not a result. As with SP+, a possession-based rating can favor a team that lost a close game, because it’s measuring quality of play, not counting wins. That’s a feature — just don’t confuse it with the standings.

The bottom line

FEI and drive efficiency reframe football from a sequence of plays into a sequence of possessions, asking the question the scoreboard ultimately cares about: did the drive produce points it should have, against this opponent, from this spot on the field? Used next to play-level efficiency and explosiveness, it rounds out the picture — catching the finishing, field-position, and special-teams effects that per-play numbers can miss. No single rating is the truth. But a team that grades out well by drive and by play is one you can believe in.