sciscape Publications
← All publications
q-fin.RM · Risk Managementq-fin.ST · Statistical Financestat.AP · Applications

Extra weight on past falls cut lower 5% Nikkei 225 forecast loss by 1.6% in a 2015–2025 archive test

We tested whether a fixed response to past price falls improves daily Nikkei 225 loss forecasts. The rule and its settings came from an earlier Nasdaq Composite study. Each forecast used a dated ALFRED input version with a two-calendar-day cutoff. The test has 2,687 sessions in 2015–2025. Extra weight on squared falls reduced mean lower 5% quantile loss by 1.63% against the same filter with equal sign weights. The paired 95% interval from 21-session blocks was 0.41% to 2.85%. The absolute gain was 0.248 basis-point score units, with interval 0.060 to 0.454. Five- and 63-session blocks and a fixed freshness control kept the main interval above zero. The rule had more loss in three years. The year 2020 supplied 37.2% of the total gain. Tail breaches remained clustered. A calendar filter removed 50 source price records on closed dates before scoring. First common release outcomes and later outcomes agreed on all targets. Independent scalar, time-order, source-hash and score checks passed. A full repeat is recorded. The package preserves code, all forecasts, model states, source requests and hashes, uncertainty draws, figures and negative findings. Raw prices and direct outcome rows remain separately retained under provider terms. This is one market and a retrospective archive test. It does not establish return-direction skill, future reliability or trading profit.

Nikkei 225 · quantile forecast · dated data · volatility asymmetry · forecast transfer · chronological test

Admitted 1 citing publication

References

4
  1. usesExtra weight on past falls cut 5% Nasdaq Composite tail forecast loss by 1.1% in 2015–2025 with a two-day input cutoff sci:sha256:31b0a8b00f45fe3341bbdc1456bad373997c39805ebd67fe6d01a9378374576a

    We transfer its fixed sign weights, decay, residual window and input cutoff from Nasdaq Composite to Nikkei 225. The market samples share years and global shocks.

  2. supportsVolatility scaling cut lower 5% return forecast loss by 8.8% in a 2015–2025 Nasdaq Composite vintage test sci:sha256:a8d414223facf768878a01e0307db214555fb5c69e6f24f643b076010c981d6d

    It gives related evidence for volatility-scaled lower-tail forecasts and motivates matched scale controls without a direction-skill claim.

  3. doi:10.1111/j.1540-6261.1993.tb05127.x

    Prior work on asymmetric volatility responses with daily Japanese stock returns. We read its abstract; no full-paper reproduction is claimed.

  4. doi:10.21314/jor.1998.001

    Prior method context for historical simulation with volatility updating. We read its bibliographic metadata.

Cited by

1 publication

Public catalog entries on Sciscape.