AstraEye Labs
Sign InEnter the Lab
AstraEye Labs · Forecast record · ATLAS
ATLASForward test running

Volatility Outlook forecast record

AstraEye records the Volatility Outlook's forecasts before their outcomes are known, and judges them once, on 17 May 2027, against standards declared on 12 November 2026. The result will be published on this page on 17 May 2027, whatever it shows.

  1. 01
    What is recorded
  2. 02
    How it is judged
  3. !
    The result, 17 May 2027
Impersonal research. AstraEye Labs is not a registered investment adviser, does not give financial advice, and does not know your goals or risk tolerance. This research does not account for your individual goals, risk tolerance, financial situation, or other circumstances. Past performance does not predict future results.
01The record

What is recorded

Every week, from the close of the last session of the week and before the next session opens, AstraEye records the Volatility Outlook's forecast of realized volatility over the next 5, 21 and 63 sessions for every stock and fund it holds ten years of prices for. Every slot is settled as a forecast or as an abstention with its reason, so the record cannot leave out a forecast that turned out badly. A forecast is scored once, when the sessions it covers have closed, against the realized volatility measured from the same daily prices.

Standards declared
12 November 2026, before the first forecast
Judged and published
17 May 2027
02The standards

How it is judged

  1. Each of the 5- and 21-session horizons is judged on its own; the 63-session horizon is reported, not judged.
  2. At least 100 effectively independent resolved forecasts, counting forecasts made on the same date as one where they move together.
  3. A lower average loss (QLIKE, the usual loss for variance forecasts) than both baselines declared with it, the trailing realized volatility over the same number of sessions and an exponentially weighted average, each with its confidence interval above zero after allowing for testing two baselines at two horizons.
  4. No SEC industry sector, and not funds as a group, more than 10% worse than the trailing baseline.
  5. Windows holding an earnings report better than the trailing baseline, with a 95% interval above zero.
  6. A horizon passes only if it meets all four. An inconclusive result is not a pass.

No category is left out: every stock and fund the test covers is in the record, funds that hold futures contracts and leveraged or inverse funds included.

03The result

Published on 17 May 2027

No result of this test is stated before 17 May 2027. On that date this page will publish whether each horizon met the standards above, and list every forecast the test scored, the ones that missed beside the ones that did not, whatever the result.

A record of past forecasts over the period shown. A past record does not guarantee or predict future results.