Hit Rate Matrix — % of prospects who became Star or Starter in first 5 NFL seasons by DMX decile (D1=best). Covers 2001–2023 draft classes. RB D1 at 93.1% is the highest validated hit rate of any tier across any position in the dataset.
93.1%
RB D1 Hit Rate
n=29
75.9%
TE D1 Hit Rate
n=29
70.0%
QB D1 Hit Rate
n=30
69.7%
WR D1 Hit Rate
n=66
Hit Rate by DecileStar+Starter %
Loading…
Offense Curveshit rate by decile
IDP Curveshit rate by decile
Residual Analysis — actual 5-year VBD minus DMX-predicted VBD. Reveals what the model systematically misses. Model R²: RB 14.7% | QB 12.3% | WR 9.4% | TE 6.9%. Low R² is informative, not a failure — it quantifies exactly how much career variance DMX explains vs factors invisible at draft time (injuries, system fit, coaching).
Position
Filter
Residual Distributionactual − predicted VBD
DMX vs Actual VBDscatter · color = career tier
WR Residual Leaderboardsorted by magnitude—
Draft Capital Efficiency — VBD ROI by draft round. 1st-round RBs (picks 1–10) produce 220.8 average 5-year VBD at an 85.7% hit rate — the single best dynasty investment by both metrics. Round 2-3 QBs have a 2.9% star rate from 34 prospects — the worst.
85.7%
RB Rd1 Hit Rate
picks 1-10 · n=63
220.8
RB Rd1 Avg VBD
5-yr cumulative PPR
73.8%
WR Rd1 Hit Rate
picks 1-10 · n=126
2.9%
QB Rd2-3 Star Rate
1 star from 34 prospects
Metric
Draft Capital Efficiency by Position & Round Tier
Full BreakdownVBD · Hit% · Star%
Positional Scarcity Waves — % above replacement-level fantasy value year-by-year post-draft. Dynasty sell windows are visible in the slope of each curve. RBs cliff. WRs plateau. TEs slow-burn. QBs are bimodal — all-or-nothing.
Position
Deciles
% Above ReplacementRB
Avg VBD by Year Post-Draft
Position Insights
Select position and click Update.
K-Means Prospect Archetypes — clusters all historical prospects into 8 archetypes based on ATH/DPOS/AWP component profiles using Lloyd's algorithm. Each archetype has a known hit rate distribution, VBD expectation, and dynasty risk profile. This is the platform's highest-IP-novelty analysis.
Position
K (clusters)
Axes
ATH vs AWP Scattercolor = cluster archetype
Archetype Profilescentroid ATH · DPOS · AWP + hit rate
Run clustering to generate
Archetype Hit Rate ComparisonStar+Starter % by cluster
Cross-Era Analysis — how has DMX evolved across draft classes? Track the D1 threshold, class composition, and component drift from 2000 to 2026. Also compare any two draft classes head-to-head on ATH/DPOS/AWP profiles.
Position
Component
WR Avg DMX by Draft Year
Class Comparison Toolside-by-side ATH · DPOS · AWP radar
Class A
vs Class B
Combine-to-Career Regression — which combine tests actually predict NFL success, controlling for draft capital? Key findings: DPOS dominates all individual tests at every position. 40 time is the most overrated metric. Broad jump is the best single athleticism predictor for QBs and TEs. AWP (college production) outperforms every combine test except DPOS.
~1.5%
ATH Alone R²
all positions — athleticism barely predicts
-0.25
TE 40-Time Corr
best single combine test correlation
0.202
QB Broad Jump Corr
best combine test for quarterbacks
13.4%
RB DPOS R²
draft capital vs VBD — strongest signal
View
Combine Test Correlation with 5-Yr VBDby position
AWP Predictive Premiumstandalone R² vs ATH vs DPOS vs DMX
Key Insight
DPOS (draft capital z-score) is the single most predictive variable at every position — explaining 5.6–13.4% of 5-year VBD variance on its own. AWP (college production) adds 3.2–8.2% of additional standalone signal, making it the most valuable athleticism-independent predictor. ATH alone explains only 1.4–3.0% — raw athleticism from the combine has minimal direct predictive power when separated from draft position and college context. The 40-yard dash is particularly overrated: it explains near-zero variance for WRs and RBs. Broad jump and vertical are better single-test predictors for most positions. The DMX composite is superior to any individual test by design — this data confirms the architecture choice.
Distribution Analysis — DMX score distributions within and across deciles. Shows how cleanly the model separates quality tiers. Also includes individual prospect ATH/DPOS/AWP radar profiles for scouting context — select any draft year to see the full class breakdown.
Position
View
Year (Radar)
WR DMX Distribution by Decile
Draft Board Heat MapATH · DPOS · AWP strength by draft class
Interactive Scatter — plot any DMX component pair against career outcomes. Hover for player names and stats. Color = career tier. Reveals where predictive signal concentrates and where it breaks down entirely.
Position
X Axis
Y Axis
DMX vs 5-Yr VBD — WRhover for player · color = career tier
Star
Starter
Contributor
Bust
AI Analytics Assistant — natural language queries grounded in the actual DFF dataset. Full context injected automatically: DMX/DPX methodology, hit rates, R² values, combine regressions, draft capital efficiency, and residual analysis findings. Saves queries to your library for future reference.
Natural Language AnalyticsPowered by Claude
Cmd+Enter to submit
Saved Query Library
No saved queries yet.
Context: DMX formula, all position hit rates, R² values, combine correlations, draft capital data, residual analysis, 25-yr historical scope.