Two peer-reviewed papers say the last half hour of the US session continues the day's direction. Gao, Han, Li and Zhou ("Market intraday momentum", JFE 2018) found it on SPY 1993-2013; Baltussen, Da, Lammers and Martens (JFE 2021) widened it to sixty futures markets over 1974-2020 and attributed it to gamma-hedging demand. Both samples end before 0DTE options became half the SPX tape.
We retested on 1,085 SPX sessions, 2022-04-14 to 2026-08-20, one observation per day: the rest-of-day return (open to 30 minutes before the close) against the last-30-minute return.
The unconditional effect is gone
| sample | days | slope | t |
|---|---|---|---|
| 2022 | 175 | +0.009 | +0.4 |
| 2023 | 250 | +0.002 | +0.1 |
| 2024 | 252 | -0.016 | -0.7 |
| 2025 | 249 | +0.016 | +1.0 |
| 2026 | 159 | -0.012 | -0.6 |
| all | 1,085 | +0.006 +/- 0.009 | +0.6 |
Flat overall, flat in every single year, signs alternating. Anyone still trading "the last half hour continues the day" on the index is trading a sample that ended before the market they are in existed.
The one place it still shows
The hedging-demand mechanism makes a sharper prediction than either paper could test: continuation should live where dealers are short gamma. Conditioning on a tape-signed 0DTE dealer book read at 15:30, the slope adds +0.055 (t +3.1) on short-gamma closes, and the public open-interest convention book sees none of it. The interaction is positive in all five years - but short closes are only 15% of sessions, the magnitude swings by an order of magnitude, and the final-40-session holdout rests on four short days. Five-for-five on sign is worth reporting; it is not worth a position.
Full tables, the named caveats, and a recipe to rebuild everything from the public per-session files (spot minutes, the aggressor-signed volume grid, per-strike IVs) are in the full post: https://firmtape.com/research/is-intraday-momentum-still-alive
All 1,000+ finished sessions are free to replay at https://firmtape.com/sessions
