What Is Pass Rate Over Expectation? PROE and Fantasy Opportunity Explained

Pass rate over expectation, usually shortened to PROE, compares how often an offense drops back with how often a model expected it to drop back in those situations. It helps distinguish a tendency from a circumstance. A team throwing while chasing a large deficit is not necessarily making the same strategic choice as a team throwing when the situation ordinarily invites a run.

Information cutoff: October 19, 2026 at 4:32 AM Central. This is a methodology guide, not a current NFL team ranking, injury report or weekly projection. Every numerical example is a hypothetical Football Nation teaching exercise. The model probabilities are invented for explanation, not claimed outputs for an actual game.

The expected part is the essential part

nflfastR documents xpass as a probability of a dropback and pass_oe as the difference from that expectation. In its implementation, the play result is represented by a pass indicator and the difference is multiplied by 100. For a selected collection of qualifying plays, average those differences to obtain a rate measured in percentage points.

Ben Baldwin’s model description identifies situational inputs including down, distance, time, score and field position. A high expected passing rate therefore is not a prediction that the offense will pass well. It is an estimate that this type of situation tends to produce a dropback.

Keep that distinction separate from efficiency. Our EPA explainer examines how a play changes modeled scoring value. PROE describes the mix of plays relative to a baseline, not the value those plays generated.

Example one: calculate the difference

Consider four hypothetical qualifying plays. Assign expected dropback probabilities of 20%, 80%, 60% and 40%. Suppose the offense drops back on the first, second and fourth plays, but runs on the third. Each dropback counts as one and the designed run as zero.

Teaching calculation	Result
Expected dropbacks	0.20 + 0.80 + 0.60 + 0.40 = 2.0
Actual dropbacks	3
Expected dropback rate	2 ÷ 4 = 50%
Actual dropback rate	3 ÷ 4 = 75%
PROE	75% minus 50% = plus 25 percentage points

The result is not a 25% completion advantage, 25 extra attempts or a 25 point fantasy bonus. It says the observed dropback rate exceeded this model’s average expectation by 25 percentage points in this tiny sample.

The same arithmetic can be done one play at a time. The four differences are plus 80, plus 20, minus 60 and plus 60 percentage points. Their average is plus 25. The model did not expect half of an actual play to occur; probabilities add to an expected count across possibilities.

Example two: higher PROE can coexist with fewer passes

Now compare two fictional offenses over the same reporting window. Offense A has 50 qualifying plays and 30 dropbacks. Its actual rate is 60%, against a 50% expected rate, producing plus 10 percentage points of PROE. Offense B has 80 qualifying plays and 44 dropbacks. Its actual rate is 55%, against a 60% expected rate, producing minus five.

A is further above its situational baseline, yet B has 14 more dropbacks. This is why you should never sort by PROE and assume you have sorted by receiving opportunity. The total number of plays still matters. So does what happens after the quarterback drops back.

A dropback is not automatically a target

nflfastR’s pass indicator includes sacks and scrambles. Those are passing process outcomes, but they do not create an ordinary receiver target. A chart based on that indicator cannot be read as a box score count of official pass attempts.

For a fantasy receiver, follow the opportunity chain: offensive plays, dropbacks, throws that can generate targets, the player’s routes and his involvement. Our route participation guide explains why being present for a passing play and releasing into a route are different facts.

Then consider distribution. In a separate simplified target example, one receiver gets seven targets from 28 team passes while another gets eight from 40. The first has the larger share, 25% versus 20%, but fewer actual targets. Use our target share guide to connect the team environment to the individual role.

Check the sample before calling it a coaching identity

Record the provider, model version, cutoff and play filters. Ask how penalties, kneels, spikes and missing predictions are handled. Do not substitute a neutral situation passing rate for PROE: a filtered descriptive rate and a model comparison are different calculations.

When combining games, use the underlying plays. A 20 play game should not receive the same weight as an 80 play game if the question is the combined rate across all 100 plays. Conversely, an equal game average is a different summary that must be labeled. There is no benefit in hiding the weighting choice.

How Football Nation would use it

For waivers, a sustained increase in passing tendency is a reason to investigate receivers whose routes are also growing. It is not enough to elevate an option who remains outside the primary rotation. Confirm the actual opening, available teammates and evidence of targets before spending.

For a trade, separate the thesis from the price. If the only argument is that the offense will pass more, ask whether the player will participate in those plays and whether the asking price already assumes the improvement. Yards per route run can add a production lens, but it cannot supply missing routes.

The strongest counterargument is that raw volume is more directly useful to fantasy managers. That is fair for describing current opportunity. PROE adds value when you want to understand how much the observed volume depended on situation. It does not replace the volume or prove that a coach’s next choice is fixed.

Common questions

Does positive PROE mean a good passing offense? No. It means more dropbacks than the selected model expected. Execution is a separate question.

Is negative PROE always bad for a receiver? No. A large offense, concentrated targets or efficient production can still create useful opportunities. Judge the whole role rather than one team rate.

Can I compare two providers directly? Only after checking their definitions, models and filters. Similar labels do not guarantee identical baselines.

Football Nation conclusion

Read PROE as passing tendency relative to a stated situation model. Then count the real opportunity and identify who receives it. Your next checkpoint is the next comparable game: does the tendency persist, do routes expand and do targets reach the player whose fantasy price you are evaluating?

Sources checked October 19, 2026 at 4:32 AM Central: nflfastR expected pass documentation, its calculation source, its play classification documentation and Ben Baldwin’s model methodology, published February 5, 2021. No current team statistics are taken from that older methodology article.