Showing journal schema alpaca. Portfolio figures are unavailable: no Alpaca credentials are configured here, and this page never needs them — it only reads the journal.
From candidate to trade
Candidates priced
262,394
every viable strike pair, every expiry, every scan
Reached the AI
120
0.05% of everything priced — the rest failed a deterministic gate first
Approved to trade
89
0.03% of everything priced
AI calls avoided
262,274
pre-screening before spending a token
What the risk engine turned away
The deterministic gates, ranked by how many candidates each one stopped. A candidate can fail several at once, so the shares add up to more than 100% — the tallest bar is the constraint actually binding. This is the whole design in one table: the gates do the deciding, and only 120 of 262,394 candidates were ever worth asking the AI about.
| Why it was turned away | Gate | Candidates | Share | |
|---|---|---|---|---|
| Already exposed to this underlying | duplicate_exposure | 239,328 | 91.2% | |
| Open interest too thin | open_interest | 181,493 | 69.2% | |
| Volume too thin | volume | 164,510 | 62.7% | |
| Portfolio risk budget spent | portfolio_risk | 131,443 | 50.1% | |
| Bid/ask spread too wide | liquidity_spread | 81,175 | 30.9% | |
| Daily loss circuit breaker | daily_loss | 8,687 | 3.3% | |
| Credit received too small | credit | 2,078 | 0.8% | |
| ai_decision_rejected | ai_decision_rejected | 31 | 0.0% |