Research tools for long/short equity

Macro prediction down to the industry level, built on a methodology of self-defined industry groups. Machine-learning-scored insider buys. Accounting-anomaly flags from the footnotes. One purpose: cut the front end of the research process.

Log in

01 — The problem

Economists are the wrong instrument

The standard indicators lag, get revised, and rarely translate into positions. Even with the CPI print a month in advance, most managers would struggle to make money on it.

This model tracks something narrower: changes in the quantity of money against changes in real output at different stages of production. And instead of an economist’s forecast, it reads the people with the most direct view of the economy, officers and directors of public companies. They see their order books. The model sees the conditions around them, and scores their buying accordingly.

The same discipline applies to disclosure. When management adds a risk factor, it is because they see a legitimate new risk. Those changes are read year over year, and they inform the macro work as well.

02 — The signal

Not every insider buy means something

Most don’t. The model rates each buy on the buyer and on the macro backdrop, generalized to a 3×3 grid, Strong, Neutral, Weak on each axis. When a name you follow prints an insider buy, you log in and see in seconds whether it carries information or should be ignored.

A Weak/Weak rating is not a short signal. It is an instruction that the buy tells you nothing, which is, on its own, worth knowing before an analyst spends a week on the name.

+26% median 1-yr return, Strong/Strong buys
−11% median 1-yr return, Weak/Weak buys
75% hit rate, short-rated industry groups, since 2021
55% hit rate, long-rated industry groups, since 2021

03 — The map

Every industry group, positioned in the regime

Scored buys aggregate into self-defined industry groups built for macro modeling, shipping sits with midstream energy, not with a catch-all “industrials.” Each group’s raw and adjusted score is mapped against the current macro regime and its forecast path.

The output is one picture: which industries the median company is likely to rise in, and which it is likely to fall in. That is the starting point of a screening process, delivered before the process starts.

Macroeconomic regime map and industry groups by macro regime
The regime path with 9-month forecast, and all industry groups scored against it.

04 — In practice

Semiconductors, called both ways

Short in January 2022, ahead of the drawdown. Long in January 2024, and again in mid-2025.

The lower panel shows why: the composition of insider buying, rated positive, neutral, negative, turned before the industry did. The adjusted score is the same series shown for every industry group on the platform.

Semiconductors raw vs. adjusted Z-score with insider buys by rating bucket
Semiconductors: raw vs. adjusted score, with insider buys by rating bucket.

05 — The footnotes

The footnotes, already read

Filings across coverage are scraped for accounts receivable sold. DSO is adjusted for it. Free cash flow is adjusted for it, and for buybacks against stock-based compensation. Risk factor sections are presented as a diff against the prior year, so what management newly considers a legitimate risk is visible at a glance.

If sold receivables are one of your red flags, the list of who is doing it, and what their numbers look like after adjustment, is already built.

06 — Access

What’s on the platform

Subscription includes the macro map, scored insider buys in large-capitalization companies, and year-over-year risk-factor changes across coverage. Data on the site is generalized by design.

Arms-length bespoke research is available to a limited number of institutional clients.

Log in / Register