Degenerate IndexQuantitative research terminal
Loading research…
Preparing the latest market snapshotTimestamp unavailableResearch and education only

Chronological checks

Model Performance vs. Baseline

These checks use later observations that were not available when the original model reading was made. Positive edge means the model beat a simple comparison after time moved forward.

Awaiting historical checks

Whole-strategy results

Returns, risk, and the simple comparison

Market-aware evidence

How each model behaved in different climates

Outcomes are matched only to the climate known on the original decision date.

Sharpe inference

Could the risk-adjusted result be luck?

This check does not create a signal. It challenges a strategy's reported risk-adjusted return for limited history, unusual returns, repeated trials, costs, and statistical dependence.

Research allocation

Did the modeled mix earn its complexity?

The learned allocator is compared with a simple chronological alternative. If it does not clear the comparison, the dashboard labels it diagnostic instead of presenting it as validated.

Glossary & model map

How to read the research

Each section answers a different market question. Expand a category for technical definitions and the practical interpretation.

Daily Top Pick
Daily Top Pick
The highest-ranked single ticker in the current model output. It combines classification, expected direction, and supporting technical context; it is not a trade instruction.
Daily Pick Board
Buy
Market climate and next-session direction are aligned positively in the combined classification. This is the strongest positive research label, not a trade instruction.
Long
The next-session direction is positive, but the broader market-climate evidence is not fully aligned.
Sell
The next-session direction is negative in the combined classification. It is a model reading, not a guarantee that the asset will decline.
Hold
The component readings do not support a sufficiently clear positive or negative classification.
Expected move
The model’s estimated next-session percentage change. It is displayed separately from the classification because direction, market climate, and magnitude answer different questions.
Model readings and their jobs
Market climate
Groups return and trading-range behavior into broad conditions. Other tools can use that condition as context without calculating it again.
Next-day direction
Estimates the sign and size of the next daily move from shared market-history inputs.
Range expansion risk
Estimates whether the next session may move more than its recent normal range. It describes movement risk, not direction.
Chart-structure setups
Converging-price and rounded-base continuation checks can identify a setup, then use historical outcomes to show how comparable structures behaved.
Long-term move conviction
Checks whether a large multi-month move is supported by participation or looks vulnerable to stalling or reversal.
Specialist research roles
Can give a signal
The tool can produce a dated directional or event reading when its required inputs are available. The reading still needs its own historical validation.
Supports other signals
The tool adds quality or risk context to another setup. It is not presented as an entry-timing call by itself.
Awaiting data
The required dated input is missing. The card remains visible with the reason, and no neutral or zero signal is invented.
Historical evidence and Sharpe inference
Chronological validation
The model makes a reading using only information available at that time, then the later outcome is used to score it.
Baseline
A deliberately simple comparison used to test whether added model complexity contributed useful information.
Sharpe inference
A statistical challenge to an observed Sharpe ratio after accounting for sample size, unusual returns, dependence, costs, and repeated strategy trials. It validates evidence; it does not create a market signal.
Same-day options positioning
0DTE
Same-day-to-expiration options research. It describes an intraday positioning diagnostic, not a guaranteed trade or a recommendation to trade options.
Dealer gamma exposure (GEX)
An estimate of how option-market positioning may change dealers’ hedging demand as the underlying price moves. The estimate depends on an explicit dealer-position convention and available open-interest data.
Candidate pin
The strike with the largest modeled absolute gamma concentration. It is a potential area of intraday price influence, not a predicted closing price.
Call wall / put wall
Strikes with the largest modeled call-side or put-side gamma concentration. They describe positioning concentrations rather than guaranteed support or resistance.
GEX concentration
The share of modeled absolute gamma exposure located at the candidate pin. Higher concentration means the estimate is less evenly distributed across strikes.
Daily Pair Trade
Daily Pair Trade
The highest-ranked eligible relationship that also meets the configured spread-displacement threshold. It is a research setup, not an execution instruction.
Daily Long
The relatively discounted leg of the selected mean-reversion relationship. It does not imply that the security is forecast to rise independently.
Daily Sell
The relatively rich leg of the same relationship. It does not imply that the security is forecast to decline independently.
No active setup
No eligible relationship crossed the configured spread-displacement threshold. The research remains blank instead of promoting a lower-quality classification.
Qualification and rank
Entry setup
An eligible relationship whose standardized spread crossed the configured research threshold.
Eligible pair
A relationship that passed the configured correlation, cointegration, multiple-testing, and mean-reversion filters.
Research only
A completed comparison that did not satisfy every eligibility requirement. It remains visible for transparency but is not promoted as a qualifying relationship.
Pair Finder metrics
Correlation
The degree to which two securities’ returns moved together in the measured sample. Correlation alone does not establish a stable long-run relationship.
Cointegration
A statistical test for whether a linear combination of two non-stationary price series behaves as a mean-reverting process.
FDR-adjusted p-value
A significance measure adjusted to control the expected false-discovery rate across many simultaneous pair tests. Lower values indicate stronger evidence under the configured procedure.
Z-score
The current spread’s distance from its estimated mean, measured in standard deviations. Its sign indicates direction; its magnitude indicates statistical displacement.
Hedge ratio
The estimated quantity of the second security used to construct a historically balanced spread against one unit of the first.
Half-life
The estimated number of trading sessions required for half of a spread deviation to decay under the fitted mean-reversion model.
Interpreting the measures
Breakout structure
Converging price boundaries and their measured breakout levels. Historical breakout probability describes comparable past setups, not certainty.
Rounded-base continuation
A rounded price recovery with a consolidation area. Qualification and historical target outcomes provide context for the setup.
Sample size
The number of observations used to estimate a result. Small or dependent samples increase uncertainty.
Confidence
A model-specific probability or strength measure. Confidence is not a guarantee and different models may define it differently.
Spread displacement
The spread’s distance from its historical mean in standard deviations, also called its Z-score.
Adjusted p-value
A significance measure adjusted for the number of relationships tested to control false discoveries.
Conviction
The degree of participation supporting a move; it is distinct from the projected return.
Data quality
Current, limited, stale, failed, or unavailable evidence. Missing inputs are never treated as Hold.
Scores, outcomes and supporting measures
Sharpe ratio
Average excess return divided by return volatility, annualized when labeled. A larger historical value does not establish repeatability.
Total return
Cumulative change over the stated evaluation period, including only the cost and income assumptions explicitly supplied.
Annualized return
A historical return expressed as an annual growth rate. It is not a forecast of the next year.
Worst decline
The largest peak-to-trough loss in the tested return series, also called maximum drawdown.
Hit rate
The share of evaluated outcomes meeting the stated model target. Interpret it with its target, baseline and sample size.
Edge over baseline
The reported difference between a model result and its simple comparison, measured on the same target and period. The metric definition must be supplied to interpret the difference.
Classification quality
Area under the receiver operating characteristic curve: how well a binary score separates the two outcome classes across thresholds. This is distinct from accuracy.
Training error
Prediction error measured on the training data. It does not measure performance on later unseen observations.
Climate persistence
Estimated probability that a market-climate state remains in that state for the next modeled interval.
Climate switch probability
Estimated probability of a change in the modeled market-climate state.
Breakout probability
The reported historical rate or modeled probability of a qualified breakout under the supplied setup definition.
Target rate
The historical share of comparable qualified structures that reached their measured target.
Reward for each unit of risk
Estimated distance to a target divided by distance to the setup’s risk boundary. Costs and execution can change the realized ratio.
Failure probability
The model’s estimated chance that a price move fails under its stated definition. The target and horizon must be known to interpret it.
Post-earnings response
Price behavior following dated earnings events, compared with similar historical observations.
Earnings-to-cash quality
A supporting measure of how consistently reported earnings convert into cash flow. It does not establish entry timing.
Cash-conversion risk
A supporting assessment of weak or unstable conversion of reported earnings into cash.
Macro Context
Dated economic-release and interest-rate observations that help explain the environment surrounding a model reading.
Intraday order-flow pressure
Trade or order-book imbalance over short intervals and its measured relationship with later price changes.
Variation explained
The share of observed variation captured by the stated statistical relationship. It is not a causal claim.
Forward horizon
The number of future intervals over which an outcome is measured. Interpretation depends on the interval length.
Quality score
A model-defined supporting score of company cash-flow quality. Its inputs, scale and limits should be supplied with the reading.
Risk score
A model-defined supporting assessment of cash-conversion risk. It is not a probability of investment loss.
Red flags
The count of checks that met the published model’s risk-flag criteria.
Repeated trials
The number of variants considered when selecting a result. Testing many variants increases the chance of selecting a lucky result.
Minimum history
The estimated observation history needed for the stated inference threshold. Dependent observations may provide less information than their raw count.
Modeled exposure
The aggregate invested share in the published public research comparison. It describes the simulation’s exposure.
Cash reserve
The aggregate share assigned to cash in the public research comparison.
Spot
The reference price of the underlying security at the option snapshot time.
Contribution
A supplied component’s influence on the combined ticker reading. Without the publisher’s scale, contributions should not be compared across models.

Every output is a research classification. None is a guaranteed price, return, or trade recommendation.