sciscape Publications
← All publications
q-fin.ST · Statistical Financeq-fin.CP · Computational Finance

A 252-session trend input gave no clear monthly Nasdaq Composite return forecast gain in a 2018–2025 archive test

We tested whether a past-year price trend improved next-month Nasdaq Composite log return forecasts over a past mean. Each forecast used only its dated ALFRED archive. All model settings were fixed before scores were calculated. The final test has 95 nonoverlapping monthly outcomes from January 2018 to November 2025. A ridge model with the past 252-session return had 2.96% more mean squared error than the mean of 120 past training returns. Its out-of-sample R-squared was -2.96%, with a 95% paired six-month block interval from -8.74% to +1.76%. Four trend inputs also gave no clear gain. A positive-slope check stayed close to the past mean. The primary model had 64.21% sign accuracy, below an always-positive forecast at 67.37%. The main test kept four months with stale source data. Removing those months gave R-squared of -3.01%. This is a fixed forecast benchmark for one index and a small monthly sample. It does not reject momentum across other markets or establish a trading profit. The package preserves code, source records and hashes, all forecasts and model states, negative results, uncertainty draws, figures, independent checks, an exact repeat of eight core result files, and repeat instructions. Raw provider prices remain separate under their source terms.

stock return prediction · time series momentum · point-in-time data · chronological evaluation · negative result · Nasdaq Composite · ridge regression

Admitted 5 citing publications

References

6
  1. doi:10.1016/j.jfineco.2011.11.003

    Motivates the past-year trend input; the original cross-market futures study has a different target and method.

  2. doi:10.1093/rfs/hhm014

    Motivates a past-mean baseline and chronological return forecast evaluation.

  3. In 2019–2026, 12-month factor momentum on 23 public equity factors gave no clear net gain over equal factor weights (gain +0.06%/month, 95% CI −0.30 to +0.44) sci:sha256:605a364268f178775b231b39cab8d4bba7e8cb3a0f6c07b7fc8934fa835b4d2e

    Related factor momentum test with weak later-sample forecasts; a different asset universe.

  4. Past VIX values reduced weekly Nasdaq Composite variance forecast loss in 504 tests from 2016 to 2025 sci:sha256:4adcb5f1bc1bda2df920b7b53f7fb09b57658a671e2df019f32b1c4d19f5143c

    Related dated-input Nasdaq study separates risk gains from weak signed-return forecasts.

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

    Uses the same retained ALFRED price source and MIT source request code. This study tests monthly signed returns, rather than daily return tails. Credit the original retained-source creator in the text.

  6. An input delay of 21 trading days cut the 5% tail forecast gain from 16.2% to 7.0% in revised US market data for 2000–2025 sci:sha256:a038fc786ca5c1675a4fcbc3effccfdec125d00f07a61b80074fe414290a2013

    Related evidence on the effect of stale market inputs; motivates retaining source-age audits.

Cited by

5 publications

Public catalog entries on Sciscape.