APEX factor · Options

Options factor — explained

Options flow reveals what informed traders are betting on the next 30-90 days. Heavy call buying with rising implied volatility precedes positive returns; heavy put buying plus put-skew steepening precedes drawdowns. Unusual single-day volume — multiples of open interest — is the strongest variant of the signal.

Where this comes from

Academic anchor

Pan-Poteshman 2006 — The Information in Option Volume for Future Stock Prices
Demonstrates that the put-to-call volume ratio at the option-class level predicts the underlying stock's next-week return: stocks in the lowest decile of P/C ratio outperform the highest decile by ~40 bps per week. The effect is strongest in low-volume options (where informed-trader share is highest) and around earnings announcements. Cremers-Weinbaum 2010 extended the result to volatility skew — when out-of-the-money put implied volatility is unusually high relative to OTM call IV, the underlying tends to fall. Both signals are informed-trader proxies that arrive earlier than equity-market price discovery.
Plain English

What it actually measures

Options trade with leverage. A directional bet that costs $10k of equity capital costs $1k of premium for the equivalent move via options, so informed traders systematically prefer options for short-horizon expressions of view. The footprint they leave is observable: a sudden 5x spike in call volume on a dormant strike with no equity-side news is somebody who knows something. The factor watches for three patterns — put/call ratio extremes, IV skew shifts, and unusual-volume single-day spikes — and weights them by the depth of the option chain (deep chains are dominated by hedgers, thin chains by directional speculation).

No calibration constants

Math sketch

pc_ratio    = put_volume / call_volume                    // 30-day rolling
pc_z        = z_score(pc_ratio against ticker's history)
iv_skew     = IV(25Δ_put) - IV(25Δ_call)                  // OTM put-call IV gap
skew_z      = z_score(iv_skew against ticker's history)
unusual_flag = (max(daily_vol / open_int) > τ) ? 1 : 0
options_raw = -1·pc_z + (-1)·skew_z + unusual_flag · sign(net_call_flow) · κ_unusual
options = z_score(options_raw)

Sign flips on pc_z and skew_z so the factor reads bullish-when-high. The unusual-flag's contribution is signed — high call-volume unusual activity adds a positive boost; high put-volume unusual activity subtracts. Threshold τ for 'unusual' and the boost coefficient κ_unusual are not disclosed; the structure follows Pan-Poteshman 2006 plus Cremers-Weinbaum 2010 plus the standard 2× open-interest convention used by industry options-flow tools.

Pipeline

How DeepVane implements it

Options chain data comes from FMP's options API at a 1-hour cadence during market hours, falling back to end-of-day if the upstream is throttled. Implied volatilities are mid-quote, no synthetic surface fitting. The factor refreshes daily at 06:00 UTC plus an inline pull during the universe sweep when an unusual-flag fires (we don't want to wait until tomorrow to register a 50× volume spike). Strike-level granularity is preserved — we keep the full call+put volume per strike per expiry, then aggregate to the ratio + skew + unusual signals.

One coherent posterior

How it composes with APEX

Options pairs strongly with Short Interest in the SHORT SQUEEZE SETUP confluence pattern — when extreme short interest co-occurs with sudden bullish call-volume unusual activity, the squeeze trigger is statistically validated by informed flow. Options also reinforces PEAD: heavy directional call flow into an earnings print plus a SUE beat plus continued call flow post-print is a high-confidence drift setup. Bearish options flow into a pharma Phase 3 catalyst is the OPTIONS-PHARMA RISK confluence pattern — informed traders are pricing trial failure ahead of read-out.

Honest limitations

When it fails

Three honest limitations. (1) Hedge contamination. A 10% drop in an underlying triggers index-fund put buying that has nothing to do with company-specific information; we partially mitigate via individual-name vs index-relative skew, but the noise is real. (2) Data freshness. The 1-hour FMP cadence misses intraday flash spikes — a single-bar volume burst won't surface until the next aggregation. (3) Coverage. Small-cap and biotech tickers have thin options markets; the signal is barely meaningful for any name with avg daily option volume under ~5k contracts. We currently only score options for tickers above that threshold; the rest receive a neutral 50.

Read next

Related factors

Short InterestPEADMomentum

See Options score on a real ticker

Every ticker page shows the per-factor decomposition. The Options score is one of twelve composing the 0–100 APEX composite.

Try NVDA →Full methodology35 invariants live