Anomaly Signal Co.

Ventures

Two Platforms. One Signal.

BioAlpha Terminal

Timestamped. Quantitative. Admissible.

BioAlpha Terminal operates on a simple principle. In biotech markets, the data moves before the news. The system monitors 975 companies daily through twenty-two independent quantitative signals, documenting exactly when trading behavior, insider activity, and market microstructure shifted relative to public disclosure. Indication-specific probability of success models provide clinical context: oncology Phase 3 base rate of 3.4%, vaccines 33.4%, CNS 8%. Insiders trading against these odds is the story the data tells.

Every reading is timestamped, sourced, and structured for evidentiary use. The output is a forensic record of what moved, when it moved, and how far ahead of the announcement it moved.

Primary

Litigation Intelligence

The litigation product is a timestamped forensic record. It shows when volume deviated from baseline, when insiders transacted, when the options market re-priced, and when the public was finally informed. The time gaps between these events are the core of a securities fraud case. BioAlpha packages them with source attribution, confidence scoring, and statistical context for counsel review.

Insider Detection: SEC Form 4 cluster analysis identifying when multiple C-suite executives trade within days of each other before material announcements

Insider Quality Scoring: Role-based transaction weighting. CMO and CSO trades weighted 3x, CEO 2x, CFO 1.5x. Maps each trade to the insider's specific access level and proximity to material information

Informed Trading: PIN and VPIN probability scoring quantifying the fraction of trades driven by private information in the weeks preceding catalysts

Options Market Divergence: Risk-neutral density extraction from options chains showing what sophisticated traders were pricing in before announcements. Flags divergence between options positioning and equity market, including elevated put/call ratios preceding negative catalysts

Dilution Forensics: Six-signal composite tracking cash depletion, S-3 filing chains, share count velocity, insider pre-dilution selling, and options positioning

S-3 Filing Detection: Real-time monitoring of shelf registrations filed 3-6 weeks before offering announcements. Connects S-3 filing dates to insider selling windows and offering announcements into a single forensic sequence

Volume Anomalies: Bayesian changepoint detection identifying exact dates of trading activity regime shifts inconsistent with public information

Regime Detection: Hidden Markov Model identifying sustained ACCUMULATION, QUIET, and DISTRIBUTION states. A DISTRIBUTION regime beginning weeks before a negative announcement, combined with insider selling, proves coordinated exit

Forecasting: Magnitude prediction models comparing actual stock movement against expected distributions, identifying statistically improbable outcomes

Case Strength Index: Predicts motion-to-dismiss survival probability, settlement range, and evidence priority for biotech securities class actions. Calibrated on 2,000+ resolved cases and validated against SEC enforcement outcomes. A hierarchical Bayesian classifier and log-linear settlement estimator quantify the signals that actually move cases: insider trading magnitude, options foreknowledge, regime state, and class period duration. The system then matches your case to its five closest historical analogs via proprietary algorithm and provides the outcome.

Signal Generation

The timing data also generates investment signals. When signals converge ahead of a catalyst, the system scores the setup. 80% accuracy. 4.86 Sharpe ratio. Walk-forward validated on 3,972 historical events.

Research Reports

For biotech and biopharma employees with equity in their company. The reports show where your company sits on the catalyst timeline, what insider and microstructure signals indicate, and what historical base rates predict for your therapeutic area.

975 Companies monitored daily

22 Independent signals

80% Forecast accuracy

SignalForce

The vulnerability no one is measuring

SignalForce is issuer-side defense for the algorithmic read of earnings. A new layer of AI and systematic readers parses your press release, filings, slide deck, prepared remarks, and Q&A in milliseconds, weighing the gaps between what you wrote, what you said, and what the numbers were. A strong quarter can read as a weak one, and the funds reading it act on the read.

We read your presentation the way those systems do, then help you shape the next one so the same facts land the way you intend, across what you create, how you prepare for Q&A, and how you deliver it. Eight patents pending.

Read the full SignalForce page →

4,400 Companies exposed to it

8 Patents filed to stop it

1 Product