Everything! Analytics · Advanced

A Historical Study with the Archive

Conditioning forward outcomes on a Nations index's decile: an illustrative walk-through of how to build a historical study from the downloadable archive — and what to be careful about when you do.

The core question a historical study tries to answer is: when a Nations index was in a particular part of its historical distribution — say, the top decile, or below the 25th percentile — what tended to happen to [some outcome variable] over the next [some horizon]? The downloadable archive makes this question answerable. Here is the methodology, illustrated step by step, with an honest account of what this kind of study can and cannot tell you.

Step 1 — Export and sort the archive

Download the CSV for the index and underlying you want to study (e.g., VolDex® on a major equity ETF). The CSV gives you a date-indexed daily series of closing-level readings going back to the earliest available date. Open it in any spreadsheet or analysis environment. Sort by date ascending — you now have the raw material for a conditional study.

Step 2 — Assign decile or percentile buckets

For each daily observation, compute its percentile rank within the data that was available at that date — not using the full archive, which would introduce look-ahead bias. This is the critical step: use only past data, not the future. For each observation on date t, rank it against all observations before date t. Assign it to a decile (1 through 10, where 10 is the top decile). Alternatively, use the lifetime and 52-week percentile ranks already computed by the Everything! platform — noting that those are full-sample ranks, so they carry a mild look-ahead quality that is worth acknowledging in your study design.

Step 3 — Measure forward outcomes

For each entry date (each observation in the decile of interest), measure the chosen outcome variable at a fixed forward horizon. For a vol-mean-reversion study, the outcome might be: the VolDex® reading 30 days later (did it revert?). For a premium-collection study, it might be: the realized vol over the next 30 days versus the VolDex® reading at entry (did implied exceed realized?). For a directional study, it might be: the ETF's total return over the next 20 trading days. Match the outcome to the question you are asking.

Step 4 — Tabulate by decile

Compute the average (and ideally the median and distribution) of your outcome variable for each decile of entry readings. Present these side by side. If there is a monotonic pattern — outcomes systematically better or worse as you move from low-decile to high-decile entry — that is consistent with a historical conditional relationship. If the pattern is noisy or non-monotonic, the decile of entry may not have been informative for that particular outcome.

Pairing with the Nations Strategy Backtester

The CSV archive from Everything! is designed to pair directly with the Nations Strategy Backtester. The Backtester accepts the archive as an input signal series and runs it against configurable strategy rules — entry when the reading crosses a percentile threshold, exit on reversion to the median, position sizing based on the z-score. The archive provides the signal history; the Backtester handles the options strategy mechanics. Together they let you test not just "what happened to the index" but "what would a strategy based on this signal have produced" — including transaction costs and the real-world path of the options, not just the expiry P&L.

Educational content from Nations Indexes. This case study is illustrative — all study designs, decile distributions, and workflow diagrams are schematic examples, not representations of real backtested results. VolDex® is a mark of Nations Indexes. Past patterns do not guarantee future performance. Nothing here is investment advice.