SYS:ONLINE
VER 2.4.1
LAT 33.9425° N
LON 118.4081° W
DATASET: 0 RECORDS
CLASSES: 0
UPDATED: 2026.04.11
STATUS: ACTIVE

Can your prospect
be the franchise?

0 quarterbacks — 0 draft classes — composite scoring
Draft Board
×
Head to Head
×
Historical Comps
×
Play Style — Cosine similarity on z-score normalized metrics. Matches QBs with the same profile shape regardless of how good they are. A mid-tier gunslinger will match an elite gunslinger because they play the same way.

Performance — Inverted euclidean distance on min-max normalized metrics. Matches QBs who performed at the same level. An elite prospect matches other elite prospects, not just QBs who play similarly.

8 metrics (unweighted): BTT%, TWP%, ADOT, TTT, P2S%, Deep Grade, Pressure xANY/A, Rush YPG.
Score guide: 85%+ strong · 70–85% good · 55–70% moderate · <55% weak
Style Drift
×

Did their college game translate to the pros? Percentile-ranked across 5 style dimensions. Click a row to expand year-by-year breakdown.

Lab
How the Composite Score Works

The composite score is a single number (0–100) that measures how good a QB looked statistically in college. It does not measure arm strength, athleticism, or leadership — only what shows up in the data. That's why some elite NFL QBs who were "raw" in college (Mahomes, Allen, Lamar) rank lower, while polished college producers (Cam Newton, Russell Wilson, Baker Mayfield) rank higher.

We grade every QB across 14 metrics grouped into six categories. For each metric, we rank the QB against every other QB in our database (140+ prospects since 2011). A QB in the 90th percentile on a metric scored better than 90% of all QBs — regardless of era or conference.

Not all metrics are weighted equally. The ones that best predict NFL success carry more weight:

Core Production
TPS
12
Total Production Score. Combines passing efficiency (xANY/A), rushing, and volume. G5 and FCS production is discounted.
BTT%
9
Big-time throw rate. Accurate passes into tight coverage.
Adj Cmp%
4
Completion % adjusted for drops, spikes, and throwaways.
Aggressiveness
MOF Grade
10
Middle-of-field throws into safety coverage. The hardest read in football.
Deep Grade
8
Accuracy on 20+ air-yard throws.
Int Grade
6
Intermediate (10–19 yd) passing grade.
Short Grade
5
Short (<10 yd) passing grade.
Over 10+ Yd%
3
Share of throws traveling 10+ air yards.
Pocket Presence
P2S%
5
How often pressure turns into a sack. Lower = better.
Fumble Rate
3
Fumbles per dropback. Lower = better.
Scrambling
Rush YPG
6
Designed + scramble rushing yards per game.
Yds/Scramble
5
Average yards gained per scramble.
Clean Pocket
Clean Pkt xANY/A
6
Production from a clean pocket.
Pressure
Pressure xANY/A
10
Production when the pocket collapses. Separates stars from game managers.

After calculating the weighted average, two adjustments are applied:

  • Conference adjustment — G5 QBs have their score multiplied by 0.93 and FCS by 0.82, since weaker competition inflates stats.
  • Red flags — If a QB trips a statistical warning (e.g., high sack rate, low depth of target, high turnover rate), a graduated penalty is applied based on severity. Barely over the line = small penalty; way past it = bigger penalty.

QBs are then sorted by composite score and assigned tiers: T1 = top 15%, T2 = 15–40%, T3 = 40–70%, T4 = bottom 30%.

Technical Details

Normalization: Percentile Rank

Each metric is converted to a 0–1 percentile by counting how many QBs in the database have a strictly lower value. A QB at the 85th percentile scored higher than 85% of all QBs on that metric. This is immune to outliers — one extreme value doesn't compress everyone else toward the middle, unlike min-max normalization. For inverted metrics (P2S%, Fumble Rate), the percentile is flipped: 1 - percentile.

Weights (92 total points)

Category Metric Weight % of Total r(wAV)
Core ProductionTPS1213.0%0.357
BTT%99.8%0.376
Adj Cmp%44.3%0.183
AggressivenessMOF Grade1010.9%0.406
Deep Grade88.7%0.324
Int Grade66.5%0.272
Short Grade55.4%0.249
Over 10+ Yd%33.3%0.117
Pocket PresenceP2S%55.4%0.215
Fumble Rate33.3%
ScramblingRush YPG66.5%0.290
Yds/Scramble55.4%0.250
Clean PocketClean Pkt xANY/A66.5%0.283
PressurePressure xANY/A1010.9%0.403

r(wAV) = Pearson correlation between the college metric and NFL career Weighted Approximate Value across 114 drafted QBs (2011–2024). Weights are informed by these correlations but also incorporate domain knowledge — for example, fumble rate has a low linear correlation but represents a real ball-security concern that scouts value.

Conference Multiplier

After computing the raw weighted average, the score is multiplied by a conference factor: P5 × 1.00, G5 × 0.93, FCS × 0.82. This bakes competition strength directly into the composite rather than applying a post-hoc tier demotion.

Red Flags (Graduated Penalties)

Five thresholds trigger graduated penalties. The further past the threshold, the larger the deduction (capped at 3 points per flag):

penalty = min(3, 2 × |actual − threshold| / threshold)

Flag Threshold What it catches
P2S% > 20%Pressure → sack rateCollapses under pressure
ADOT < 9Avg depth of targetOverly conservative / checkdown-heavy
Deep Att% < 14.5%Deep attempt shareNever pushes the ball downfield
Dropbacks < 800Career volume (2015+ only)Small sample — stats less reliable
TWP% > 4.5%Turnover-worthy playsReckless decision-making. Note: Mahomes (4.0%), Allen (4.4%), Lamar (4.0%) all clear this threshold.

What This Model Can't Measure

This is a statistical production model. It cannot evaluate arm talent, release mechanics, football IQ, leadership, scheme fit, or development trajectory. QBs like Patrick Mahomes (#33 composite) and Josh Allen had mediocre college stats but elite physical tools. Conversely, Mac Jones and Bryce Young had excellent college stats but face questions about NFL ceiling. Use this as one input alongside film study and scouting reports, not as a complete evaluation.

Data Sources
  • PFF — Career scouting grades, xANY/A, TPS, and situational splits (2011–2026)
  • Pro Football Reference — NFL draft picks, career stats, wAV (2011–2025)
  • PFF Passing Summaries — Season-level data for 980+ college QBs (2014–2025)
Feature Requests
Column Groups
Feature Request
What would make this tool more useful?