ELEVATED72
Pending:51 CRITICAL
Conflicts:3
Confidence:84%
Loading...
No game selected
Welcome to FlyIQ Ask AI. I can help you analyze game film, find specific plays, generate reports, and identify coaching opportunities. Try asking me to: - Find clips matching specific criteria - Generate teach tape for a player or position group - Build a scouting or post-game report - Identify the top fixable errors from any game All answers are source-grounded in your actual game data and evaluation pipeline.
04:01 AM
Quick actions
AI responses are grounded in your game evaluation data
Play 14 - PA Boot RT
Model evaluation score and confidence
Offensive line failed to execute zone blocking scheme, resulting in immediate pressure from the defensive end. The quarterback was forced to throw off-platform, leading to an inaccurate pass.
OL_ZONE_BLOCK_001
Left tackle failed to reach the defensive end on the zone step
OL_COMBO_BLOCK_003
Guard-tackle combo did not transfer cleanly to second level
QB_POCKET_002
Quarterback climbed pocket when edge was still available
WR_ROUTE_DEPTH_001
Routes reached correct depth for play-action concept
Supporting Data
Pressure Time
1.8s
Avg Pressure Time
2.6s
Qb Throw Accuracy
62%
Expected Points
-1.2 EPA
No game selected
Cmd+Enter to save
No notes yet
Add your first note above
69 plays · Max divergence: 34pts · The gap between lines = where the scoreboard lies
What's driving the gap between traditional and grade-adjusted WP?
OL pass block grades dropped from 74.2 (season avg) to 55.8 in Q3. Interior pressure rate 41% vs 18% season average.
DL run stop grades cratered from 71.0 to 48.9. Opponent averaging 6.2 YPC in second half vs 2.8 in first half.
Opponent QB decision grade surged to 91.3 in Q3. Adjusted to tempo, reading blitz packages, exploiting Cover-1 tendency.
CB1 coverage grade dropped to 41.2. Safety help arriving late — fatigue-correlated decline starting play 38.
QB accuracy under pressure fell from 68% to 44%. Pocket collapsing 0.4s faster than Q1-Q2.
Punter averaging 48.3 net yards, pinning opponent inside 20 on 2 of 3 punts in Q3.
YBC dropped from 2.4 to 0.6 in second half. Opponent front 7 dominating at point of attack.
How real-time momentum feeds into the grade-adjusted model
Plays where grade-adjusted WP diverged most from traditional — the moments the model saw what the scoreboard couldn't
First play where OL grades dropped below 50. Interior pressure already building — this was the canary in the coal mine.
DL gap integrity grade hit season low. Grade model detected run-fit breakdown that traditional WP missed entirely.
Largest single divergence. Traditional WP still at 58% because 21-14 looks fine. Grade model at 30% — it saw the unit collapse before the scoreboard did.
Safety coverage grade bottomed out. Grade model already priced in the collapse — traditional WP just now catching up as score ties 21-21.
Grade model was at 10% when traditional was still 36%. The grades predicted this outcome 12 plays earlier.
Convergence point — traditional WP finally reflects what grades showed 20 plays ago. Scoreboard caught up to reality.
Games where traditional WP was high but grade-adjusted WP saw the loss coming
Positive = traditional WP overestimates (scoreboard lying). Negative = hidden edge the scoreboard misses.
Why this model sees what traditional models miss
You are winning 21-14 but your offensive line grades dropped 18 points in the last 10 plays, your secondary is grading below 45, and the opponent QB decision grade just hit 91. Traditional models see a 7-point lead. FlyIQ sees a team about to lose. The grade-adjusted win probability is the truth the scoreboard has not caught up to yet.