🧰 Layer 2 Models

The quant models behind the desk.

Forward-looking predictive models, the self-learning loop, multi-path expected-value arbitrage, real trade receipts, set-level heat, and the sealed-product market. All updated every 4 hours by the cron pipeline. Math published at /methodology.

🔥 Set Heat — which sets are running

Per-set composite: avg score + BUY density + recent momentum

⏳ Loading set heat…

📦 Sealed Market — boxes / ETBs / bundles

Separate market from singles · top buys + undervalued picks

⏳ Loading sealed market…

📒 Trading Journal — actual track record

Real trades logged from the desk · realized P/L + open positions
Realized P/L
Portfolio ROI
Win Rate
Open Positions

Recent Closes

Sold Card Tax Cost Net Proceeds P/L ROI Days
⏳ Loading…

Open Positions

Item Card Qty Cost / Unit Total Cost Days Held
⏳ Loading…

🧠 Self-Learning Engine — the AI is evolving

Weights re-tune from realized 7-day returns · Pearson-correlation optimizer
Weights Source
Retrains
FW Emissions
Signals Tracked

⏳ Loading model weights…

📊 Past Picks
Loading scoreboard…

💰 Arbitrage Opportunities — expected-value plays

Multi-path EV model: raw→PSA10, raw→PSA9 floor, PSA9→PSA10 cross-grade, retail flip. Fees + grading + ship modeled.

⏳ Loading arbitrage scanner…

🔮 Future Winners — the AI hunt list

Predictive model: momentum + scarcity + accessibility + chase-variant signal + tier premium. Capped at $5,000 entry for accessibility.

⏳ Loading Future Winners…