PairDesk runs the same protocol every week on about 1,900 US names. A desk build runs it on yours: your universe, your data, your delivery shape, on your cadence. What you receive is the computation and the caveats, never a forecast, and the method document is complete enough that your own people can reimplement and check every number.
Engle-Granger in both orientations against finite-sample critical values, hedge ratios by total least squares, half-lives with the small-sample correction, a held-back period for out-of-sample checks, and a false-discovery rate against a null built from real series rotated out of alignment. Capacity comes from actual dollar volume, with the leg that binds it identified per pair. Same-issuer, same-underlying and same-exposure pairs are flagged rather than hidden. The data source's survivorship gaps are measured and printed, not assumed away.
| universe | state |
|---|---|
| US stocks and ETFs | built, running weekly since August 2026 on about 1,900 liquid names; the free build is the full 21 August desk |
| NSE F&O universe | built first, on a survivorship-correct panel back to 2010 |
| Your list | any market with a daily price history, quoted after a look at the data |
from $6,000
half on signing, half on the first accepted build; ongoing builds priced per period in the same letter
Scoped and quoted in writing before anything is built, against a stated universe, data source, delivery format and cadence. No revenue share, no success fee, no payment that depends on the desk making money. The first build arrives within three to four weeks of the data being available.
The engagement runs under the published desk-build letter: you own every build and the method document; the engine stays mine; your data and ideas never inform my own book, and nothing from my book enters your desk.
Write with the universe, the data you hold or want used, and the shape you want the desk delivered in. You get a written scope and a price back, and nothing is built until you have both.
WRITE FOR A SCOPE