ELEVATED72

Predictability risk rising in 1st-down personnel groupings

Pending:51 CRITICAL

Conflicts:3

Confidence:84%

Methodology

The FlyIQ Professional
Grading Standard

A 6-axis, EPA-calibrated, confidence-scored, coach-governed grading system designed for league adoption. Every grade is explainable. Every grade has uncertainty. Every correction makes it smarter.

Why a new standard?

Existing grading systems grade players on a single axis, hide their rubric, don't report uncertainty, and don't learn from coaches. The result: coordinators use the grades as rough inputs but can't defend them in staff meetings, draft rooms, or contract negotiations. FlyIQ's standard is built so that every grade can be traced back to the snap, explained in English, and challenged by a coach — and when the coach is right, the system learns.

The 6 Evaluation Axes

Every player on every snap receives an independent score on each of these 6 dimensions. They are scored separately because they measure fundamentally different things — and conflating them produces a grade that's easy to compute but impossible to act on.

Status: Decision, Technique, Outcome, and Effort are live on every player grade today. Impact and Difficulty are in active development and will roll out in the next engine version.

Decision

0–100 per snap

“Did they make the right read given what was available?”

Examples by position:

QB: Read progression accuracy, throw-or-hold decision, audible recognition
WR: Route adjustment based on coverage leverage, option-route read
OL: Correct protection ID, combo-block handoff timing
DB: Run/pass key read, coverage responsibility assignment
LB: Blitz/coverage decision, fill-gap recognition

Why separate? A great decision that leads to a bad outcome (tipped ball, dropped pass) should still grade well. PFF penalizes the outcome. FlyIQ separates the two.

Technique

0–100 per snap

“How well was the physical skill executed?”

Examples by position:

QB: Footwork, release point, ball placement, pocket presence
WR: Route crispness, break speed, hand-catch vs. body-catch
OL: Hand placement, leverage, anchor strength, pass-set depth
DB: Backpedal fluidity, hip flip speed, trail technique, ball skills
DL: Hand usage, rush moves, counter-move timing, disengagement

Why separate? Technique degrades under fatigue. Tracking it separately reveals which players are physically falling off in the 4th quarter — and which are coachable.

Outcome

0–100 per snap

“What actually happened? (EPA-anchored)”

Examples by position:

Completion vs. incompletion, yards after catch, turnover
Sack allowed, pressure created, run stuff, tackle for loss
EPA delta from pre-snap expected value to post-play value

Why separate? The result matters. But it doesn't tell the whole story. A QB who makes a perfect throw that gets dropped shouldn't be penalized on Outcome the same way as on Decision. Separating axes lets coaches distinguish process from result.

Effort

0–100 per snap

“Was maximum effort given throughout the play?”

Examples by position:

Pursuit to the ball carrier (defenders)
Blocking through the whistle (OL/TE)
Finishing routes even when not the primary read (WR)
Getting off blocks and flowing to the play (LB/DL)

Why separate? Effort is the one axis that's 100% controllable. A player with elite technique who loafs on 20% of snaps is a liability. Effort is detectable from tracking data and All-22 film, and it's the first thing every coaching staff wants to see.

Impact

0–100 per snap

“How much did this player affect the play relative to their responsibility?”

Examples by position:

A CB who gets beaten on a 60-yard TD = high impact (negative)
A backside OG whose defender doesn't reach the play = low impact (neutral)
A blitzing LB who causes a fumble = high impact (positive)

Why separate? Not all snaps matter equally for a given player. Impact weighting ensures a shutdown corner who erases a WR (low target count, low yards allowed) grades well even though nothing 'happened' on their side. PFF's system penalizes players who don't show up in the box score — Impact rewards them.

Difficulty

0–100 per snap

“How hard was this situation?”

Examples by position:

Coverage type (press man vs. soft zone)
Opponent quality (SP+ ranking of the opposing unit)
Game state (score differential, time remaining, field position)
Leverage index (how much this play affected win probability)

Why separate? A great play in a meaningless 4th-quarter blowout against an FCS team is not the same as a great play in a 4th-quarter tie against a top-10 defense. Difficulty is the context multiplier that makes cross-game and cross-opponent comparisons honest.

The Aggregation Pipeline

Per-snap scores flow through a 6-layer pipeline before becoming the published FlyIQ Grade. Each layer adds transparency, not opacity.

Per-Snap Evaluation

Every player on every snap receives all 6 axis scores from the CV + attribution model. This is the atomic unit — everything else aggregates from here.

Situational Weighting

Snaps are weighted by leverage (win probability impact), quarter, opponent strength, and game pace. A 3rd-and-8 in the 4th quarter of a close game contributes more than a 1st-and-10 in a blowout.

Game Aggregation

Per-snap scores → per-game profile. Bayesian shrinkage pulls small-sample games toward the position mean. Each game gets a confidence score based on detection quality, player ID accuracy, and snap count.

Season Aggregation

Per-game → per-season. Snap-weighted with the reliability dampener: 1-game players get 30% z-score credit, 5+ game players get full credit. Trend computation tracks whether the player is improving or declining.

8-Layer Validation

Every grade passes through Identity, Position, Coverage, Grade, Temporal, Constraints, Cross-Dimensional, and Scheme Novelty validators before publishing. Any layer can flag a grade for human review.

Governed Learning

Coaches can AGREE (confirm), CORRECT (change + train), or FLAG (queue for review) any grade. Corrections feed back into the model. The system gets smarter with every reviewed rep.

The FlyIQ Grade Profile

The final output is not a single number. It's a profile:

FlyIQ Grade

Reliability-dampened, z-score-stretched display grade (30–99)

6-Axis Radar

Profile

Spider chart of Decision / Technique / Outcome / Effort / Impact / Difficulty

Confidence

92%

How much we trust this number (0–100%), based on sample size, ID quality, volatility, and games

Interval

± 3.2

Statistical uncertainty band — wider for volatile or small-sample players

Position Pctile

94th

Rank within the position group after Bayesian shrinkage

Trend

↑ +4.2/wk

Game-by-game trajectory — improving, stable, or declining

Clutch

1.12×

Performance in high-leverage situations vs. baseline — clutch multiplier

Narrative

Text

Auto-generated English explanation of what drove the grade and why

Audit Trail

Per-snap

Every axis score, every snap, every validation result — traceable and challengeable

How This Compares

	PFF	ESPN QBR	NextGen Stats	FlyIQ Pro
Positions graded	All	QB only	All (physical)	All
Axes	1	1	Physical only	6 independent
Context adjustment	None	Win prob	None	Leverage × opponent × difficulty
Transparency	Opaque	Opaque	Partial	Full rubric per snap
Uncertainty	None	None	None	Confidence + interval
Speed	24–48 hrs	Next day	Real-time	Hours (CV pipeline)
Learning	None	None	None	Coach-governed corrections
Auditable	No	No	No	Yes — per-snap trace
Comparable	Grader drift	—	—	Z-score normalized, position-relative

Ready to see it in action?

1,900+ games and 195,000+ plays are already graded in the live engine.

View Player Grades →See Pricing

flyiQ

The FlyIQ Professional
Grading Standard

The 6 Evaluation Axes

The Aggregation Pipeline

The FlyIQ Grade Profile

How This Compares

Ready to see it in action?

Ask AI

Play Review

Model Claim

Evidence (4 rules)

Why Confidence is Low

Your Decision

Review History (1)

Coach Notes

The FlyIQ Professional
Grading Standard

The 6 Evaluation Axes

The Aggregation Pipeline

The FlyIQ Grade Profile

How This Compares

Ready to see it in action?

The FlyIQ ProfessionalGrading Standard

The 6 Evaluation Axes

The Aggregation Pipeline

The FlyIQ Grade Profile

How This Compares

Ready to see it in action?

Ask AI

Play Review

Model Claim

Evidence (4 rules)

Why Confidence is Low

Your Decision

Review History (1)

Coach Notes

The FlyIQ ProfessionalGrading Standard

The 6 Evaluation Axes

The Aggregation Pipeline

The FlyIQ Grade Profile

How This Compares

Ready to see it in action?

The FlyIQ Professional
Grading Standard

The FlyIQ Professional
Grading Standard