Ticker intelligence.

Every completed ticker regime calculation, with the production method, validation evidence, probability, score, and earnings controls retained for audit.

139
calculated tickers
104
bullish > +9
1
earnings window
Separate tickers with spaces, commas, AND, or OR.

45-day downside & efficiency

Loss probability measures a finish more than 15% below the starting price after 45 calendar days—not a drop at any point during the period. Efficiency is the pipeline’s opportunity score: 100 × (positive-return probability − loss probability). Higher is more favorable, but it is not expected return.

Risk snapshot: 2026-10-01 · 119 tickers calculated. Missing or differently dated estimates remain unavailable. Limited tail-event evidence means low probabilities deserve extra caution.

How to read these metricsOpen guide
Regime score · −100 to +100
Bearish
−100 to −51
Moderately bearish
−50 to −10
Neutral
−9 to +9
Moderately bullish
+10 to +50
Bullish
+51 to +100

45D positive

Estimated probability of finishing above the starting price after 45 calendar days, using the selected positive-return model. Higher means a greater estimated chance of a gain, not a larger expected gain.

45D loss >15%

Estimated probability of finishing more than 15% below the starting price after 45 calendar days. Lower is more favorable. This is an end-of-period loss, not the chance of touching that level or suffering a peak-to-trough drawdown during the period.

Efficiency score

The pipeline’s opportunity score: 100 × [P(positive return) − P(loss greater than 15%)]. Higher is more favorable on this measure. For example, 70% positive probability and 10% severe-loss probability give a score of 60—not a forecast of a 60% return. It does not measure return per dollar invested.

Opportunity ratio

Shown in View evidence: P(positive return) ÷ P(loss greater than 15%). The same 70% and 10% example gives 7×. Higher is more favorable, but tiny loss estimates can inflate the ratio. It is not a gain-to-loss payoff ratio; neither metric accounts for the size of gains or losses.

Tail evidence

Counts severe-loss events in out-of-sample validation. Limited tail events means fewer than five events; Adequate event count means at least five. This is a sample-size warning, not a confidence percentage or proof that a low loss probability is reliable.

Not calculated / Date mismatch

Not calculated means no usable downside estimate was imported for that ticker. Date mismatch means its downside and regime snapshots differ. Missing values appear as a dash, are sorted last, and must not be interpreted as zero risk.

Selected model

The main table’s model and OOS Brier refer to positive returns. Downside candidates are selected separately by their lowest out-of-sample Brier score and shown in View evidence. QQQ’s positive-return score retains the standing market model.

OOS Brier

Average squared error between predicted probabilities and outcomes, tested on observations outside each model’s training sample. Lower is better; it is not a percentage accuracy. Compare candidates on the same outcome and validation sample. Rare losses can produce very low Brier errors even when crash prediction remains uncertain.

Downside model candidates

Historical baseline uses the ticker’s training-period severe-loss frequency. Market / QQQ applies the QQQ downside probability unchanged to the ticker. Ticker four-factor fits the four inputs to that ticker’s severe-loss outcome. For QQQ, the last two are the same model, not independent evidence.

Direct score

The ticker’s score from the four-factor model, retained for comparison even when another candidate wins production selection.

Market score

The current QQQ market-regime score used by the market-identity candidate.

Factor contribution

Each standardized factor’s addition to or subtraction from the model logit. Positive values raise probability; negative values lower it.

Validation rows / events

Rows are dated out-of-sample predictions used to evaluate the models. Events are the observations that actually ended with a loss greater than 15%. Overlapping 45-day windows are not independent trials and may reflect the same market downturn.

Risk snapshot

The analysis date of the imported downside estimates, not a live quote timestamp. Efficiency and the ratio require a downside estimate dated the same as the ticker’s positive-return estimate. Training observations fit the models; validation observations evaluate their forecasts.

Earnings window

Identifies when a ticker is within the model’s configured proximity period around a reported earnings date. It is event context, not an instruction to act or refrain from acting.

Data quality

Whether the inputs passed the production checks. The detail panel explains any unavailable or cautionary condition.

Regime score = 200 × P(positive return) − 100. A regime score of 0 means a 50% positive-return probability. Efficiency score uses a different formula: 100 × [P(positive return) − P(loss greater than 15%)]. An efficiency score of 0 means those two probabilities are equal. Neither score is a predicted return or a stand-alone trading instruction.

139 of 139 tickersClick a row for model evidence