×

⚡ WAR ROOM — Your AI Draft Edge

This is not another rankings list. WAR ROOM uses 4 machine learning models trained on 5+ years of NFL data to predict every player's fantasy output — and it beats ADP by 58%.

🎯 How To Use This Board

1. The Player Pool — Who To Draft

Players are ranked by VBD (Value Based Drafting), not raw points. This accounts for positional scarcity — a RB scoring 180 pts is more valuable than a QB scoring 250 pts, because replacement QBs score more than replacement RBs.

Column	What It Means
VBD	How many more points this player gives you vs a waiver-wire replacement at the same position. Higher = draft earlier.
PROJ	Projected season fantasy points (half-PPR).
RANGE	80% confidence interval [low–high]. Wide range = risky. Narrow = reliable.
RISK	low / medium / high. Based on prediction uncertainty. High risk = boom or bust.
BYE	This player's bye week. Red = conflict with your other starters.

2. My Roster — Your Team

Click DRAFT to add a player. They auto-fill into the correct roster slot (QB → QB slot, RB → RB1/RB2/Flex/Bench, etc.). Your projected season total is shown at the bottom.

3. Target Intel — The Edge

Panel	What It Does For You
🎯 Pick Advice	Top 5 players to draft next, based on VBD + your roster needs. "MUST DRAFT" = you still need that position.
📊 Positional Scarcity	Shows which positions have the steepest talent cliffs. If RB drops off hard, grab RBs early.
⚠ Bye Conflicts	Warns if drafting a player creates a bye week hole. Never start a week with empty slots.
💤 Sleepers	Players our model ranks WAY above ADP — the market is sleeping on them.
📉 Busts	Players ADP overvalues — the market is paying too much for name recognition.
🔒 Handcuffs	Backup RBs for your starters. If your 1st-round RB goes down, their backup takes over. Draft insurance.

4. Opponent Picks

Click OPP to mark a player as drafted by an opponent. This removes them from your pool and recalculates VBD/scarcity in real-time.

🧠 Key Concepts

VBD — Why It Beats Raw Points

Imagine a QB projected for 250 pts and a RB projected for 180 pts. Who's more valuable? The RB. Because the 12th-best QB scores ~170 pts (only 80 pts less), but the 24th-best RB scores ~120 pts (60 pts less with way more starters needed). VBD measures the gap between a player and the replacement you'd get for free.

Why We Beat ADP

ADP is a popularity contest — it reflects name recognition and recency bias. Our model uses statistical regression, age curves, and 4 ML models to find players the market misvalues. The biggest edge is at QB — ADP consistently undervalues dual-threat QBs because their rushing points aren't obvious from traditional stats.

The 4 Models

Each player's projection is the average of 4 independent models:

• Ridge — Stable baseline, prevents wild predictions
• Random Forest — Captures age cliffs and breakouts
• XGBoost — Best at complex feature interactions
• CatBoost — Best generalization, wins on most positions

Averaging 4 models reduces overfitting and produces more reliable projections than any single model.

Data: nflverse (via nfl_data_py), FantasyPros ADP, NFL schedule · Walk-forward validated · No data leakage · Train/test R² gaps < 0.02