NFL Draft Analytics · Tight Ends

Draft Intelligence
for Tight Ends

Machine-learning models trained on real NFL combine data, college production stats, and NFL career outcomes for 456 tight ends from 2000–2025.

456TEs in Database
0.96RF Model AUC
20Real Features
272026 Prospects
🔍
Player Lookup
Search any TE from 2000–2025. Real draft data, college stats, and NFL career outcomes.
Explore database →
🎯
Scout Any Prospect
Input combine metrics and college stats. Project draft position, success probability, and contract value.
Run projection →
📋
2026 TE Class
Full model rankings for all 27 TEs who participated in the 2026 NFL Combine.
View rankings →
⚡
Compare Players
Side-by-side comparison of any two TEs across combine, college, and career metrics.
Compare now →
Recent Hits & Misses
Model Validation Highlights
A sample of players the model correctly identified as hits or misses based on pre-draft signals.
Player Database

Player Lookup

Search any TE from 2000–2025 to retrieve their draft score, combine profile, college stats, and career outcomes.

Browse
All Drafted TEs
Click any column header to sort. Click a player row to view their full profile.
Prediction Tool

Scout Any Prospect

Enter combine measurements and college production stats. The model projects draft pick, success probability, and contract value.

① Combine Metrics
② College Stats
③ Review & Predict

Measurements

Speed & Explosiveness

Agility

Career Totals

Best Season

Last Season & Efficiency

Combine Summary

College Summary

Missing values are automatically imputed using the median from 456 real TEs in our dataset. The model works best when at least 40-yard dash, career yards, and career TDs are provided.
2026 NFL Draft

2026 TE Draft Class

Model projections for all 27 tight ends who participated in the 2026 NFL Combine, ranked by predicted success probability.

Click any card for detailed breakdown
Head-to-Head

Compare Players

Side-by-side comparison of any two TEs across combine athleticism, college production, and NFL career outcomes.

Quick Matchups
Classic Comparisons
Methodology

Model Information

How the ensemble model works, what data it uses, and how to interpret its outputs.

Data Sources
  • ✓ NFL Combine measurements (2000–2025)
  • ✓ College football receiving stats (2004–2025)
  • ✓ NFL career receiving stats (2000–2024)
  • ✓ PPA (Predicted Points Added) per play
  • ✓ Usage rate by down and distance
Model Architecture
  • ✓ Random Forest (200 trees, max_depth=5)
  • ✓ Gradient Boosting (100 estimators, lr=0.1)
  • ✓ Logistic Regression (C=0.5, regularized)
  • ✓ Ensemble: 50% RF + 35% GB + 15% LR
  • ✓ AUC: 0.96 (RF), 0.995 (GB), 0.70 (LR)

Feature Importance (Random Forest)

Derived from real feature_importances_ across 456 TEs. Athleticism dominates over college production in this model — vertical jump and 40-time are the top two features.

Success Definition

Success = top 40% of drafted TEs by cumulative career receiving EPA (Expected Points Added). The EPA threshold computed from our dataset is +26.3 EPA. Players above this threshold are classified as hits. This metric captures both volume and efficiency — a player can rack up yards and still have a low EPA if their targets weren't helping their team score.

Historical Draft Quadrants (2000–2024)

Every drafted TE classified by model prediction vs. actual NFL outcome.

Interpretation Guide

Output What It Means How to Use It
Success Probability Estimated chance of exceeding +26.3 career EPA Use >55% as draft target threshold
Predicted Pick Where the model thinks this player should be drafted Compare to consensus boards for surplus value
Surplus Value Model pick minus actual pick (positive = undervalued) Target players with +20 or more surplus
Rookie AAV Expected annual salary based on draft slot (2024 scale) Estimate cap cost of drafting this player
TE Draft Intelligence
TAMIDS Student Data Challenge · Texas A&M · 2026
Trained on 456 TEs · NFL Combine + CFB (PPA) + NFL EPA data · Educational use
Team Name: BALLIQ · Website Developed By: Prayag Peram