Third down is where drives live or die — but raw conversion rate hides how hard the down-and-distance really was.
Published June 6, 2026 · NFL Analytics
Broadcasters call it the money down for a reason. Third down is the hinge on which a possession turns: convert, and the drive lives to fight another set of downs; fail, and the offense almost always punts the ball away. There is no down with more leverage - the swing in expected outcome between success and failure is enormous, and it happens on a single snap.
That leverage makes third-down conversion rate one of the first numbers fans and broadcasters reach for. A team converting "45% on third down" sounds clearly better than one converting 35%. But raw third-down rate is one of the most misleading headline stats in football, and understanding why is the difference between watching the scoreboard and actually reading the game.
The reason raw third-down rate misleads is that it averages together situations that have almost nothing in common. A 3rd-and-1 is a near-coin-flip in the offense's favor - a quarterback sneak or a power run converts it a large majority of the time. A 3rd-and-12 is a long shot the defense expects to win, with the offense forced into an obvious passing situation against a coverage built to defend the sticks.
Two teams at 42% third-down rate can be wildly different if one faced 3rd-and-short all year and the other 3rd-and-long
Lumping those together produces a single percentage that hides the only thing that really matters: what distances was this team facing? An offense that posts a mediocre overall third-down rate while constantly facing 3rd-and-10 may actually be converting at an impressive clip for those distances - it is just stuck in bad spots. Another offense might post a gaudy rate simply because it keeps arriving at 3rd-and-2.
This is why serious analysis breaks third down into distance buckets rather than reporting one number. The standard split looks something like short, medium, and long:
| Bucket | Typical distance | Illustrative conversion range | Reality |
|---|---|---|---|
| Short | 3rd-and-1 to 3rd-and-2 | ~60% to 70%+ | Conversion is likely - the offense is favored |
| Medium | 3rd-and-3 to 3rd-and-6 | ~40% to 50% | Genuine coin-flip territory |
| Long | 3rd-and-7 to 3rd-and-10 | ~25% to 35% | Conversion is unlikely - the defense is favored |
| Very long | 3rd-and-11+ | well under ~25% | A long shot; often effectively a punt down |
Once you accept that every third-down distance carries its own baseline conversion probability, the smarter metric writes itself. Instead of asking "what share did they convert?", ask "how many did they convert relative to what an average offense would have converted from those same distances?" That is conversions over expected.
Conversions Over Expected = Actual conversions − Sum of league-average conversion probabilities for each situation faced
The mechanic mirrors every other "over expected" metric in the analytics toolkit - the same logic behind completion percentage over expected and rushing yards over expected. You assign each third down its expected conversion probability based on distance (and ideally field position and time), add those up to get an expected number of conversions, and compare it to what actually happened.
Hypothetical figures for illustration only.
Suppose an offense faced 200 third downs. Based on the distances of each, an average offense would have been expected to convert about 78 of them. This offense actually converted 90.
90 − 78 = +12 conversions over expected
That +12 is the real signal. It says this team beat the situations it was handed - which is a very different and more meaningful statement than "they converted 45%," because it controls for whether they were living in 3rd-and-short or fighting out of 3rd-and-long all season.
Here is the insight that reframes the entire down: the best way to be great on third down is to rarely face a hard one. Third-down distance is not handed out at random - it is the direct product of how the offense performed on first and second down. Gain six yards on first down and you set up a manageable 2nd-and-4, which sets up an inviting 3rd-and-short. Lose two on first down and take an incompletion on second, and you have backed yourself into a 3rd-and-12 the defense is licking its chops over.
This is precisely what early-down efficiency metrics like EPA and success rate are built to capture. A play on first or second down that gains positive expected points - that keeps the offense ahead of the chains - is doing the quiet work of turning future third downs from long shots into coin flips into near-locks. When you read a team's success rate on early downs, you are largely reading a leading indicator of how comfortable its third downs are going to be.
The flip side belongs to the defense. A unit that wins early downs - stuffing first-down runs, generating second-down pressure - is manufacturing the 3rd-and-long situations where defenses hold a big edge. Forcing the offense into very long third downs is one of the most reliable ways to get off the field, which is part of why early-down defense is so valuable even when it does not show up in third-down stats directly.
One more reason to distrust the raw number: third-down conversion rate is noisy and volatile from year to year. The sample is modest - a team faces only so many third downs in a season - and the down is high-variance by nature. Many third downs are coin-flip medium-distance situations where the bounce of a contested catch, a borderline spot, or a single broken tackle decides the outcome.
The result is that third-down rate regresses heavily toward the mean. Teams that finish a season near the top in third-down conversion are not reliably the ones who finish near the top the next year, and the same goes for the cellar. A chunk of any extreme third-down rate is simply the small-sample luck of how those coin flips happened to land.
Third down also sets up the down that follows it, and the two are best understood together. When an offense fails to convert, the result is not automatically a punt - it is a fourth-down decision. A failed 3rd-and-1 frequently rolls into a 4th-and-1, where the math often favors going for it; a failed 3rd-and-12 rolls into a 4th-and-long where punting is usually correct. The leverage of third down, in other words, partly flows from how the resulting fourth down is handled.
Smart offenses even plan across the two downs - calling a third-down play with the knowledge that a short gain still leaves a very makeable fourth down, which changes the risk calculus on third down itself. The two downs form a single sequential decision, and treating third down in isolation misses that connection.
Third down is the leverage down - convert and the drive survives, fail and you punt - but raw third-down conversion rate is misleading because distance is everything: 3rd-and-1 is a near-lock and 3rd-and-12 is a long shot, and a single percentage hides which a team actually faced. Good analysts judge third down by distance buckets and by conversions over expected, which controls for the situations a team was handed. The deeper truth is that you win third down on first and second down: early-down efficiency sets up manageable third downs, so a great offense may face fewer and shorter ones. And because the raw rate is volatile and regresses hard year to year, lean on the underlying early-down efficiency for the real signal - and remember the failed third down simply hands off to a fourth-down decision.
Want the code behind these metrics? Work through the 45-chapter NFL analytics tutorial.
Browse tutorials Free tools