How the Scanner Works

The symbol universe

ACondor scans three sources of symbols, merged into one universe each cycle:

1. Manual allowlist: symbols you explicitly permit in the dashboard. Starts with a core set of liquid names: SPY, QQQ, IWM, AAPL, MSFT, AMZN, NVDA, GOOGL, META, TSLA, AMD, XLF, and others.
2. News-promoted symbols: when the news classifier produces a high-confidence company-specific headline (≥ 0.80 confidence), that symbol is added to the scan universe for 48 hours. Capped at 100 symbols.
3. Screener-promoted symbols: ACondor pulls tastytrade's public watchlists daily (tasty IVR, Liquid Symbols, Liquid ETFs, NASDAQ 100, High Options Volume) and adds names that meet minimum IV Rank and liquidity rating. Capped at 250 symbols. Refreshes every 24 hours.

Entry gates (every symbol must pass all of these)

1. Liquidity check: the underlying must have sufficient average daily volume and option open interest. Known-liquid symbols (deep-market ETFs and mega-caps) bypass the per-cycle probe and always pass.
2. IV Rank floor: each engine has its own floor. Mechanical: ≥ 30. Earnings: ≥ 50. Contraction: ≥ 80 with a confirmed recent jump. A symbol below its engine's floor is skipped for that engine.
3. Earnings window: the mechanical and contraction engines skip any symbol with earnings inside their DTE window. The earnings engine requires earnings inside the window.
4. Bid/ask spread: the actual proposed legs are checked at the time of entry (not just the pre-screened symbol). If any leg's bid/ask spread exceeds 10% of mid, the trade is skipped.
5. Buying power check: before placing any order, ACondor submits a margin dry-run to tastytrade to get the exact buying power effect. If deploying that trade would exceed the allocation cap or leave less than the 20% reserve, the trade is skipped.
6. Post-exit cooldown: if a symbol was recently closed, it sits out for a cooldown period. Profitable exits clear in 2 hours. Stop-loss and news exits cool for 72 hours.
7. Risk layer gates: VIX kill switch, macro halt calendar (FOMC, CPI), daily/weekly loss circuit breakers, sector concentration cap, position count cap.

Candidate scoring

Candidates that pass all gates are scored:

• credit / max_loss is the credit-to-risk ratio (tastytrade's “1/3 of width” rule targets ≥ 0.25)
• POP = 1 − Σ|short deltas| (probability all shorts expire worthless)
• bias_multiplier = 1.0 baseline, 1.20× for directional pair structures when a strong bias is present

The top score does not get the only shot. ACondor ranks every structure that clears the score floor, then walks the ranking best-first: the first structure that also passes the entry-time checks (the real bid/ask width of the exact legs about to be traded, the per-trade risk cap, and a collision check against options you hold yourself) is the one entered. If the highest-scoring structure quotes too wide to fill fairly, or risks more than the per-trade cap allows, the next-best structure is evaluated instead of the trade being abandoned. Every structure passed over is logged with its own reason.

If no structure clears, the cycle logs a skip and waits for the next interval.

What you see on the dashboard

The Scanner tab in the ACondor dashboard shows the most recent cycle's candidates: every symbol evaluated, its IV Rank, the engine that evaluated it, the proposed structure, the score, and the reason it was skipped or entered. This is the primary tool for understanding what the bot is seeing and why it is or is not trading.