A sudden surge in trading volume is often read as climax: the last buyers or sellers rush in, the move exhausts itself and reverses. The literature on volume and price changes (Blume and Easley; Karpoff) gives the idea some standing, but there is no named party forced to trade on the wrong side.
What singled it out was the lab's factor delta scan. Volume was the only factor whose predictive relationship had strengthened in 2026, and the scan's own explanation was retail inflows and options activity. "Most improved" was taken as a reason to test it.
How it was tested
Pre-registered on 2026-09-02, the second of three candidates from that scan, together with the ATR% spike reversion and the one-day return reversal. Signal: the 20-day average volume above its 60-day rolling mean plus 2 rolling standard deviations. The pre-registration described a contrarian rule in both directions; the script ran it short, held for 5 sessions. The decision rule asked for a Sharpe of at least 0.6 before the formal gates. One run.
Where it died
The run lost money on raw returns: Sharpe −0.36, t −1.60 over 4,875 trades. The post-mortem drew the obvious inference that high volume precedes continuation, not reversal. That opposite reading was later pre-registered as a long strategy of its own and failed Gate 7 on 2026-10-14 with t −8.58.
On 2026-10-30 the method audit opened the original backtest and found the rolling statistics computed across all symbols together, inside the signal's query string, instead of per symbol. The threshold that defined a "spike" was therefore not the one registered, and the verdict was changed to void. The 2026-09-02 numbers stay on the page as what the broken script produced, not as a measurement of the idea.
The record
Thesis
A surge in a stock's 20-day average volume far above its recent norm marks an exhausted move that reverses. It was the only factor whose predictive relationship had strengthened in the lab's 2026 factor scan, attributed to retail inflows and options activity.
Blume, L. & Easley, D. (1983). Trading volume and returns.
Karpoff, J. (1987). The Relation Between Price Changes and Trading Volume. Journal of Financial and Quantitative Analysis.
Specification
signal
20-day average volume above its 60-day rolling mean + 2 rolling standard deviations
direction
short, as run (the pre-registration describes a contrarian rule in both directions)
exit
5 sessions
caveat
The rolling statistics were written inside the signal's query string over the whole factor panel, not grouped by symbol — the same construction the sibling one-day-reversal test later found had not run as specified. The candidate itself came from the factor delta scanner that the lab later found to ignore cross-sectional co-movement. Treat the result as weak evidence.
Gates
Pre-registrationpass
Was the hypothesis, its parameters and its pass/fail rule written down before any data was looked at?
Raw significance (legacy)not evaluated
Are the strategy's raw returns statistically different from zero?
pre-registered decision rule: Sharpe ≥ 0.6, then Gates 1-8
t raw-1.60sharpe-0.36trades4,875
Raw per-trade returns, no benchmark adjustment — and not a valid measurement: the script computed the rolling statistics over the whole panel at once, not per symbol, so the rule tested was not the rule registered.
Why it died
Void, reclassified after the October 2026 method audit: the backtest script computed its rolling mean and deviation across all symbols together instead of per symbol, so it never ran the rule as registered. The numbers it produced are kept for the record only: Negative raw returns (Sharpe −0.36, t −1.60 over 4,875 trades): high volume in this universe preceded rallies more often than reversals, so betting on reversal lost money.
What it taught
"Most improved factor" is a statement about variation, not about an edge. The post-mortem's own reading — high volume precedes continuation — was later tested as its own pre-registered long hypothesis and also failed, with an even stronger negative sign: the extreme tail of volume underperforms.
Verdict history
2026-09-02Pre-registered from the factor delta scan; single run rejected No alpha
2026-10-14The opposite (long continuation) reading is tested separately and fails Gate 7 with t −8.58
2026-10-30Reclassified as void by the method audit: rolling statistics computed over the whole panel, not per symbol Void