ANY/A Explained: Adjusted Net Yards Per Attempt, the Best Simple QB Stat

The Best Simple Quarterback Stat

If you want one publicly available number to rank quarterback play without resorting to a black-box model, the smart pick is Adjusted Net Yards per Attempt - ANY/A. It is built entirely from a passing box score, you can compute it on the back of an envelope, and it correlates with team scoring and winning better than completion percentage, raw yards per attempt, or even the traditional passer rating. That mix of simplicity and predictive power makes it a favorite of analysts who still want something they can fully explain.

One-line idea: ANY/A takes plain yards per attempt and fixes its three biggest blind spots - it rewards touchdowns, punishes interceptions, and finally counts sacks as the failed pass plays they are.

The Formula, Term by Term

Start with the ancestor of every passing efficiency stat, yards per attempt: passing yards divided by pass attempts. It already blends volume and effectiveness into a per-play rate. But it treats a 12-yard completion and a 12-yard touchdown the same, ignores the disaster of an interception, and pretends sacks never happened. ANY/A patches all three with a single expression:

ANY/A = (Pass Yards + 20 × Pass TD − 45 × INT − Sack Yards) / (Pass Attempts + Sacks)

Or, in math form:

$$ \text{ANY/A} = \frac{\text{Yds} + 20\,\text{TD} - 45\,\text{INT} - \text{SkYds}}{\text{Att} + \text{Sk}} $$

Every piece is doing a specific job:

Term	What it does
Pass Yards	The base value - total passing yardage on the day or season.
+20 × Pass TD	A touchdown bonus. A passing score is worth roughly 20 yards of additional credit because reaching the end zone is far more valuable than the raw yardage suggests.
−45 × INT	The interception penalty. Each pick costs 45 yards. Turnovers flip possession and field position, so the hit is large and deliberately asymmetric to the touchdown bonus.
− Sack Yards	Yardage lost to sacks is subtracted from the numerator, so taking sacks directly drags the rate down.
+ Sacks (denominator)	Sacks are added to attempts, so each one counts as a dropback that failed. This is the "Net" in Adjusted Net Yards.

Why +20 and −45? These weights are conventions from Pro-Football-Reference, chosen so the bonuses and penalties line up with the historical point value of touchdowns and turnovers. They are not arbitrary like the caps in passer rating - they are rough estimates of real point impact, and the heavier interception penalty reflects that turning the ball over hurts more than a score helps.

A Worked Example

Example: computing ANY/A for a hypothetical game

Round numbers invented to show the arithmetic - not real game data.

QB A's line: 300 passing yards, 3 passing touchdowns, 1 interception, 35 pass attempts, and 3 sacks for 20 yards lost.

First the numerator:

300 + (20 × 3) − (45 × 1) − 20 = 300 + 60 − 45 − 20 = 295

Then the denominator:

35 attempts + 3 sacks = 38 dropbacks

So:

ANY/A = 295 / 38 ≈ 7.8

Plain yards per attempt here would be 300 / 35 ≈ 8.6, which looks great. ANY/A trims it to about 7.8 because the interception and the three sacks were real, drive-stalling events that simple YPA ignored. That gap between 8.6 and 7.8 is exactly the information ANY/A adds.

As a rough orientation for the scale, league-average ANY/A in the modern pass-friendly NFL typically sits somewhere in the neighborhood of 6 to 7 yards, with strong full seasons pushing into the 7s and 8s and rough seasons falling into the 5s. Treat those as ballpark reference ranges rather than precise cutoffs - the league average drifts year to year.

Why It Beats Completion % and Passer Rating

ANY/A predicts winning so well because its ingredients are the things that move the scoreboard: explosive yardage, touchdowns, turnovers avoided, and dropbacks that don't end in a sack. Compare that with the popular alternatives:

vs. Completion %

Completion percentage rewards throwing short and safe. A checkdown artist can post a gorgeous completion rate while moving the offense nowhere. ANY/A is yardage-weighted, so it does not reward empty completions.

vs. Raw YPA

Yards per attempt is the right shape but blind to scores, picks, and sacks. ANY/A is YPA with those three corrections bolted on - same intuition, far more complete.

vs. Passer Rating

Passer rating never counts sacks, uses arbitrary 0-2.375 caps, and is built on 1970s constants. ANY/A counts sacks, has no artificial ceiling, and uses weights tied to real point values.

The payoff

One transparent yards-based number, no model required, that historically tracks team scoring and wins about as well as anything you can compute by hand.

ANY/A+: Indexing to the Era

Raw ANY/A still has one problem it shares with passer rating: passing has gotten easier over time, so a 7.0 in a low-scoring era is more impressive than a 7.0 today. The fix is to index it the way baseball indexes OPS+ - rescale so that 100 always equals the league average for that season, with the spread normalized so each point above or below 100 means the same thing across years.

100

Exactly league average for that era

115+

Clearly above average passing

85

Clearly below average passing

ANY/A+ is what lets you say a 1970s passer and a modern passer were "equally good relative to their peers." Like OPS+, it is era-adjusted by construction, so cross-generation comparisons become fair.

How It Relates to QBR and EPA

ANY/A sits on a ladder of quarterback metrics that trade simplicity for completeness. Think of it as the most sophisticated thing you can build from a box score, one rung below the full play-by-play models.

Metric	Built from	Trade-off
ANY/A	Box score: yards, TD, INT, sacks, attempts	Fully transparent and hand-computable, but counting-stat based.
EPA per dropback	Play-by-play down, distance, field position	Captures situational value, but needs play-level data.
Total QBR	Play-by-play plus credit-division and leverage weights	Richest context, but proprietary and debatable.

The good news is that ANY/A usually agrees with the fancier metrics on the big questions. When ANY/A and EPA disagree, it is often because of exactly the things ANY/A cannot see - which brings us to the caveats. For the full comparison of the model-based side, see our breakdown of passer rating vs. QBR and EPA vs. DVOA.

What ANY/A Still Misses

It's still a counting stat

ANY/A is built from aggregate box-score totals, not individual plays. It cannot tell a third-and-long conversion from a meaningless completion in garbage time. Down, distance, and game state are invisible.

It ignores rushing

Like passer rating, ANY/A counts only pass plays. A quarterback whose legs add real value - scrambles, designed runs, third-down conversions on the ground - has part of his game uncounted.

No context for pressure or scheme

A clean pocket and a collapsing one look identical in the totals. ANY/A cannot separate the quarterback from his protection, his scheme, or the quality of receivers running open.

Sack credit is shared in reality

ANY/A charges every sack to the quarterback, but sacks are partly the offensive line's fault. Pairing ANY/A with pressure-rate and pass-block metrics gives a fairer picture.

Use it as a screen, not a verdict. ANY/A is excellent for a fast, defensible ranking of passing efficiency. When it flags a surprising name, dig deeper with accuracy and pressure context - start with CPOE to separate a quarterback's accuracy from his supporting cast.

The bottom line

ANY/A is yards per attempt grown up: it adds a roughly 20-yard bonus per passing touchdown, subtracts a roughly 45-yard penalty per interception, and counts sacks as failed dropbacks by removing sack yardage from the top and adding sacks to the bottom. Those corrections tie it to real point value, which is why it predicts scoring and winning better than completion percentage or the capped, era-inflated passer rating while staying simple enough to compute by hand. Index it to 100 as ANY/A+ for fair cross-era comparisons. Just remember its honest limits - it is still a counting stat that ignores rushing, situational leverage, and the scheme and protection around the quarterback - so treat it as the best quick screen, then confirm with play-by-play metrics like EPA, QBR, and CPOE.