Reading Results

Read the results

Guide home | Dashboard tour

Read the source, timestamp, and quality flags before the calculated values. The same rule applies to terminal output, agent replies, and dashboard charts.

Data and quality

Field or label	Meaning	What to do
provider_used and source	Origin recorded by the calculation	Distinguish synthetic data, user snapshots, and provider data.
Capture time	Time associated with the input	Distinguish it from the artifact-generation time.
degraded	Reduced quality or an unavailable required input	Read the reason before using the values.
notes	Other observations and assumptions	Read them even when degradation is false.
Missing IV	No usable implied volatility for that contract	Read the skip counts. Do not substitute a guessed value.
unlimited	Unbounded modeled gain or loss	Do not treat it as missing data or a finite amount.

An artifact can be internally valid without being current or useful for a particular research question. Synthetic screenshots demonstrate the interface, not market performance.

Greeks and positioning

The ladder contains contract sensitivities. Read its units block before combining or comparing values. In particular, a unit change in volatility is different from a one-percentage-point change. Vega is per 1.00 of volatility, so divide by 100 for the per-point figure. Theta, charm, veta, and color are per calendar day. The skipped block counts the contracts left out of the ladder and why. The sample ladder keeps 22 of 62 contracts inside the default strike band and counts the other 40 under out_of_band.

Dealer exposure adds open interest and an assumed position sign to contract gamma. The assumed dealer holdings are not observed holdings. A wall or flip is conditional on that assumption and the available chain.

Structures and ranking

Read the legs and quantity signs before the summary. Check whether stock ownership, collateral, or a second expiry is part of the structure.

Probability of profit is conditional on the stated model. The comparison score orders structures under that model. It is neither a recommendation nor evidence of an achievable return.

For a two-expiry structure, inspect how the surviving leg is valued. The plan's probability.model field states the method: a lognormal underlying at the near expiry, at the near chain's at-the-money implied volatility, with the far leg marked at the volatility it carries today and the marking curve integrated numerically over the scan grid. A maximum found over a scan window does not establish a global maximum. An entry delta ratio does not establish bounded tail risk.

Simulation

The GARCH-t simulation models the underlying from price history. It answers a different question from a probability calculated with chain IV.

If converged is false, do not quote the posterior quantiles as reliable estimates. Inspect the parameter diagnostics and repeat the analysis with appropriate settings. More draws do not guarantee convergence.

Backtests and paper positions

A backtest uses modeled premiums and omits trading costs and execution effects named in its honesty statement. Overlapping entries require the reported block treatment for uncertainty estimates. Read the benchmark over the same windows.

A paper position records a plan before later marks. It has no broker execution, fill verification, or account reconciliation. A missing later quote can make a position unmarkable. Entry marks are mid quotes, not fills. A close settles the structure at intrinsic value against the supplied or newest spot. The result note says the profit still assumes the entry mid was achieved.

Next: Research examples.