Methodology

How we score market reaction

SMRS — the SueWallSt Market Reaction Score — is a 0–10 composite measuring how severely a stock reacted to a corrective disclosure relative to its own volatility, market-cap impact, volume anomaly, and sector isolation. Every other module on the site — damages, recovery, shock map, litigation study — extends or consumes it.

What it measures

The four sub-scores

Each component captures a distinct dimension of severity. Composites are universe-percentile ranked, winsorized at the 1st/99th percentile, then scaled 0–10.

AR
Vol-adjusted shock

Abnormal 2-day return divided by 60-day residual volatility (z-score units).

MCL
Market-cap impact

Shares outstanding × |price drop|, log-scaled.

AV
Volume shock

(Volume D0 + Volume D1) ÷ (2 × 30-day average volume).

SR
Sector isolation

Move relative to sector after controlling for the broad market.

Composite formula

How the score is calculated

Weighted sum across the four sub-scores. Vol-adjusted shock leads, market-cap dollars second.

SMRS = 0.35 · AR + 0.25 · MCL + 0.20 · AV + 0.20 · SR
Tier bands

What the tiers mean

SMRS < 4
Mild

Routine volatility.

SMRS 4 – 6
Moderate

Notable but not extreme.

SMRS 6 – 8
Severe

Strong fraud-style reaction.

SMRS ≥ 8
Extreme

Catastrophic, lawsuit-quality reaction.

Damages

How we estimate shareholder damages

For each case, we estimate aggregate shareholder damages — the losses of investors who bought stock at inflated prices during the fraud period and held through the corrective disclosure.

  1. 01
    Disclosure impact

    When a corrective disclosure causes the stock to drop, we isolate the company-specific decline by subtracting the S&P 500 return on the same day. Multi-disclosure cases measure each one and layer the inflation.

  2. 02
    Damaged shares — one-trader model

    Every share in the float has equal probability of trading on any given day. We track shares purchased during the class period and estimate how many survived to each disclosure. Trading volume is halved to exclude intraday round-trips.

  3. 03
    PSLRA lookback cap

    Federal law caps per-share damages at the trading price minus the stock’s 90-day mean post-disclosure price. For recently-filed cases we use a provisional lookback (refreshed every 6 hours) until 90 calendar days have passed.

The one-trader model is the most aggressive standard damages model — figures are preliminary estimates for comparison purposes, not litigation-ready calculations.
Recovery forecast

Predicting the 30-day path

Beta-matched cohort modeling. We answer: of historical disclosures with similar post-event sensitivity to the market, how did they move over the next 30 trading days?

  1. 01
    Estimate beta

    For every scored event we fit a 60-day post-disclosure beta against the S&P 500. Once 60 trading days have passed the beta becomes final; before then the cohort match is held back.

  2. 02
    Match the cohort

    A target case is matched against historical events whose 60-day beta falls within an adaptive band (±0.10 to ±0.50, widened until the cohort is large enough to be informative).

  3. 03
    Aggregate the trajectory

    We compute the median 30-day post-disclosure return path of the matched cohort and surface the IQR fan. Confidence is reported as a function of cohort size, not a p-value.

The recovery sweep refreshes every 12 hours. See the live tool at Calculate / Recovery Forecast.

Shock map

Classifying events by behavioral quadrant

A two-axis grid plots every scored event by SMRS against its pre-event volatility percentile. The four quadrants give defense and plaintiff counsel a shared vocabulary for arguing severity.

A
Fundamental Repricing

High SMRS, low pre-event vol — market revalued the company.

B
Panic / Overshoot

High SMRS, high pre-event vol — may be overreaction.

C
Routine Volatility

Low SMRS, low pre-event vol — odd, but not severe.

D
Noise

Low SMRS, high pre-event vol — within expectations.

Open the interactive grid at Explore / Shock Map.

Litigation study

Does SMRS predict lawsuit incidence?

A logistic regression of lawsuit indicator on SMRS, controlling for raw drop magnitude, sector, and market-cap bucket. Run nightly across the universe of scored drops.

We fit lawsuit ~ SMRS + drop% + sector + size and report the marginal SMRS coefficient, its 95% CI, the AUC of the full model versus a drop-only baseline, and the decile-bucketed lawsuit rate. Inference requires at least 100 lawsuit-class events; below that threshold we render the descriptive sections only and label the page Preliminary.

Live results at Explore / Litigation Study.

Recovery patterns

Severity vs. recovery

The cross-sectional view: across every closed case, how does SMRS at disclosure correlate with the realized 30-day return path?

We bin the case library by SMRS tier and plot median trajectories with IQR fans. This is where the structural relationship — that more severe shocks tend to mean-revert less — is visible. See the interactive view at Explore / Recovery Patterns.

Caveats

What it does not tell you

SMRS is not a causation finding, not legal advice, and does not predict outcomes deterministically. It is a normalized severity index over a shared universe of U.S. equity disclosures. See Data Sources for upstream provenance and refresh cadence.