New Orleans was 89% to win at halftime, then never scored again — and lost to Jacksonville on even yardage

We read all 23 drives of Jacksonville's 24-20 preseason win at New Orleans on 15 August, and the final box score is a near-perfect draw that hides a game decided by timing. The Saints and Jaguars finished within two yards of each other — 277 to 275 — split 18 first downs to 19, and each cashed two red-zone trips (New Orleans on four tries, Jacksonville on five). New Orleans even averaged slightly more per snap, 4.5 yards a play to 4.2. Read straight across, the sheet looks like a coin flip. That evenness is an average of two opposite halves. New Orleans scored on four of its first eight possessions and led 20-6 at the break. ESPN's win-probability model put the Saints at 89.4% late in the second quarter and still 87.2% to open the third ^mmodeled. Then New Orleans ran four more drives — a missed field goal, two punts, and a possession that expired with the game — and did not score again. Jacksonville, which had thrown both of its interceptions before halftime, turned the ball over zero times after it and scored on three of its four second-half possessions: 18 unanswered points — and ESPN's win-probability line ran from near-certain New Orleans to zero ^mmodeled, without a single second-half takeaway to mark the turn. The scoring order tells it plainly. New Orleans opened with an Audric Estime 9-yard touchdown run, added a 1-yard Zach Wilson touchdown pass to Oscar Delp, and kicked two Tanner Brown field goals to reach 20-6 at the half. After the break Jacksonville answered with a Hunter Long touchdown catch and two-point conversion, a 36-yard Cam Little field goal — capping a 17-play drive — and a J'Mari Taylor 2-yard touchdown run with 4:55 left for the final margin. Every one of the game's three turnovers happened in the first half; the second half, which New Orleans lost 18-0, was turnover-free for both sides. Jacksonville held the ball 35:12 to 24:48 and won without out-gaining anyone. The yards came out even, and the second-half turnover margin was zero; what set the result apart was which half each team did its scoring in [analysis].

Strip the game to its two totals — 277 yards, 275 yards — and you would call it a toss-up that could have gone either way [analysis]. The complete drive-by-drive read says something more specific: this was two teams playing the same game in opposite orders. New Orleans front-loaded its offense — efficient early, a 2-play touchdown drive among four scores before the break — then produced nothing across four second-half possessions. Jacksonville back-loaded its — two field goals and two interceptions in a 6-point first half, then three scores on four drives after it. The totals match because the halves cancel, not because the teams were even at any single point in the game [analysis]. The win-probability arc is the clearest single view of what the box score conceals [analysis]. A model that sat near 89% for New Orleans at halftime and ended at zero, with no second-half turnover and no blowout to explain the swing, is describing a collapse that the yardage and turnover columns never register ^mmodeled. What moved was execution and sequence: New Orleans stopped finishing drives — the missed field goal, the two punts — while Jacksonville started finishing them, including a 17-play march for three points and the go-ahead touchdown drive late. These are co-occurring facts in the record — even totals, opposite-half scoring, a win-probability line that fell without a takeaway — not a demonstrated mechanism; the data show what happened and when, not why the Saints went quiet [analysis].

The honest ceiling here is that this is a preseason game, and one game ^sspeculative. A second-half shutout in August plausibly reflects roster churn and vanilla playcalling as much as any in-game turn — factors we did not isolate ^sspeculative. We could not confirm how ESPN's model is trained or whether it adjusts for preseason conditions, so we do not lean on "89.4%" as a settled probability — we report it as the model's stated output at that moment, not a claim about what would have happened ^mmodeled. And a single game supports no broader pattern — nothing here should be carried to the regular season or to either team's season outlook ^sspeculative. What is not in doubt is the closed record we re-derived from the box score and drive log: 277 yards to 275, a 20-6 halftime lead, an 18-0 second half, all three turnovers before the break and none after, and New Orleans scoring on four of its first eight possessions and none of its last four.