Time Horizons

What a time horizon means on the performance dashboard, and how each bucket is computed.

Written By [Scanr.ai] Eoin

Last updated 2 months ago

Time Horizons

Time horizons are performance cohorts โ€” they group a scan's performance into forward-looking time buckets. Each bucket holds the average of whichever metric you've selected, measured over that horizon.

Concretely, a horizon answers the question: "if you had bought a stock at the moment it was scanned, then sold it this far in the future, how would the scan have performed?"

An example

Take "1 Day Sharpe Ratio". This means:

What is this scan's Sharpe ratio if you look at the performance of stocks up to one day after they were scanned โ€” i.e. if you bought each stock at the time of the scan, sold it one day later, and computed the Sharpe ratio across all of that scan's results.

Change the horizon to "1 Month" and the same metric now reflects buying at scan time and selling one month later. Change the metric to Win Rate and you get the share of those trades that were profitable over that same holding period.

So every performance number is really a pair: a metric (what is measured) and a horizon (over how long it is measured).

The available horizons

HorizonHolding period
1D1 day after the scan
3D3 days after the scan
5D5 days after the scan
10D10 days after the scan
1M1 month after the scan
3M3 months after the scan
6M6 months after the scan
1Y1 year after the scan
YTDFrom the scan date to the end of the year

Why it matters

A scan that looks strong on a 1-day horizon may be a short-term momentum setup, while one that shines at 3 months is better suited to swing or position trading. Comparing the same scan across horizons tells you how long its edge tends to last, and switching the horizon lets you match a scan to your own trading style.

Every performance card lets you pick the horizon independently, so you can view the same ranking through a short-term or long-term lens without changing the underlying scan.

Related: see Performance Metrics for what each selectable metric measures.