Morning Brief: WTI at $104.12 Ahead of ISM and Friday Payrolls

MacKinlay's event-study procedure asks researchers to state which firms enter a sample and to inspect bias created by that choice. A seasonal Gann date can be examined with the same discipline, although MacKinlay did not test Gann dates. First list every eligible asset-date observation using a rule set before viewing returns. Then report which entries have usable prices in the fixed window and why others are absent. The fraction with data is coverage; the reaction rate among observed entries is a different number. Suppose an editorial worksheet lists 30 opportunities. Twenty have complete price records, and 12 meet a predeclared reaction rule. The observed hit rate is 12/20, or 60%, while data coverage is 20/30, or 66.7%. The ten missing outcomes might all fail or all qualify. Under those two extremes the full-list hit rate lies between 12/30, or 40%, and 22/30, or 73.3%. These are arithmetic bounds, not confidence intervals. Record missing reasons separately: unavailable historical quotes, market closures, and eligibility changes call for different repairs. Do not silently replace a missing symbol with one that performed well or drop an inconvenient year. Name the asset universe, date rule, price source, and reaction criterion before running the worksheet. If the missing range changes the conclusion, obtain better records or describe the finding as unresolved.

Try it

An editorial Gann-date study prelists 30 asset-date opportunities. Complete data exist for 20, and 12 of those meet a fixed yes/no reaction rule. Compute data coverage, the observed hit rate, and the lowest and highest possible full-list hit rates if all ten missing outcomes are unknown.

Limits

These bounds assume ten unknown binary outcomes. They do not correct nonrandom missingness, estimate a probability, or prove calendar causality. MacKinlay's article concerns economic events and equities. The Gann-date worksheet and its arithmetic are editorial examples, not his empirical findings.

Source

A. Craig MacKinlay, Event Studies in Economics and Finance, Journal of Economic Literature 35(1), 1997, printed p. 15, section 2, 'Procedure for an Event Study,' paragraph beginning 'After identifying the event'; university scan PDF page 3. Original peer-reviewed finance-methods article, pp. 13-39, hosted as a university scan. The missing-outcome bounds are an explicitly editorial extension of its sample-selection warning.

A longitude tells you where a planet is, not whether someone could see it. JPL Horizons observer quantity 48 estimates moonlight scattered by Earth's atmosphere at the target's position. Its sky_SNR compares the target's estimated visual surface brightness with that moonlit background. Think in linear brightness units for this ratio, not differences between logarithmic magnitude numbers. Below one, the target is dimmer than the modeled moonlit sky. Above one, it clears that particular background test, but a detector still needs enough sensitivity to record it. The field has a narrow domain. Horizons reports it for an observer on Earth's surface when the target and Moon are above the local horizon, the Sun is at least in astronomical twilight, and the target is neither Sun nor Moon. A target radius is also needed for sky_SNR. Outside the model's conditions, n.a. means unavailable, not a zero-brightness sky. In a synthetic worksheet, a target brightness of 7 against background brightness of 10 gives a ratio of 0.7. If the target brightness is 16, the ratio is 1.6. Record the site, timestamp, target, and validity conditions before using a visibility label to frame a market inspection window. Neither ratio supplies a price forecast.

Try it

In a hypothetical Earth-surface observer table with a known target radius, suppose both objects are above the horizon and the Sun meets the astronomical-twilight condition. Background linear brightness is 10 units. Target A has 7 units and target B has 16. Compute each sky_SNR and classify its moonlight-background test. What should quantity 48 report instead if the Moon is below the horizon?

Limits

Horizons estimates idealized moonlight scattering, not an observed sighting. Its manual excludes local light pollution, weather, and galactic brightness; the estimate has stated uncertainty. The ratio depends on site and target conditions. It neither validates historical visibility claims on its own nor demonstrates a relationship with market returns.

Source

NASA/JPL, Horizons System Manual, Definition of Observer Table Quantities, quantity 48, Sky brightness and target visual signal-to-noise ratio (SNR). Primary NASA/JPL Solar System Dynamics technical manual, version 4.98e (August 25, 2026). The brightness values in the worksheet are synthetic.