Fractiz
Pattern Detail
Pre-Holiday Effect
Whether the session right before a US market holiday produces different RTH returns than a normal day.
Reliable effect
None
No window is distinguishable from the baseline.
Baseline Avg Return
+0.065%
Across 4,724 NQ sessions (1d)
Windows tested
1
0 reliable, 1 within noise
Summary
No window beats the all-session baseline beyond chance. Every difference below is within the margin of error, the seasonal effect is not measurable here.
Per-Window Stats
| Window | Sessions | Avg Return | % Up | Std Dev | Δ vs Baseline | Verdict |
|---|---|---|---|---|---|---|
| Pre Holiday | 56 | +0.072% | 60.7% | 1.11% | +0.007% | Noise |
| Baseline (all sessions) | 4,724 | +0.065% | 55.6% | 1.41% | — |
"Reliable" = the window's average return differs from the all-session baseline beyond a 95% chance threshold, on at least 30 sessions. Everything else is within the margin of error.
Reliable effect
None
No window is distinguishable from the baseline.
Baseline Avg Return
+0.042%
Across 4,724 ES sessions (1d)
Windows tested
1
0 reliable, 1 within noise
Summary
No window beats the all-session baseline beyond chance. Every difference below is within the margin of error, the seasonal effect is not measurable here.
Per-Window Stats
| Window | Sessions | Avg Return | % Up | Std Dev | Δ vs Baseline | Verdict |
|---|---|---|---|---|---|---|
| Pre Holiday | 56 | +0.157% | 62.5% | 1.07% | +0.116% | Noise |
| Baseline (all sessions) | 4,724 | +0.042% | 54.5% | 1.24% | — |
"Reliable" = the window's average return differs from the all-session baseline beyond a 95% chance threshold, on at least 30 sessions. Everything else is within the margin of error.
What this pattern measures
A pre-holiday session is the last trading session immediately before a US market holiday. Detected empirically from gaps in the data: any session where the next trading session is more than the expected one calendar day away (or more than three days for a Friday) is treated as a pre-holiday session.
Definitions used on this page:
- Sessions are aggregated from RTH bars only (08:30 to 15:00 Central Time for CME equity index futures).
- Return is
(close − prior close) / prior close. - Holidays are inferred from the data itself. There is no hardcoded NYSE holiday list, so the detection survives changes to the holiday calendar over time.
- Baseline is every session in the sample. The pre-holiday bucket is the small subset preceding a non-weekend market closure.
Why it matters
The pre-holiday effect is one of the oldest “anomalies” in equity research: returns on the session before a holiday have historically averaged higher than typical sessions. Explanations include short-covering ahead of an illiquid period, behavioral anchoring, and a quirk of measurement. The numbers below show whether the effect persists in this instrument over the sample period.
Sample size is small: only about 9 to 11 US market holidays per year, so even 18 years of data produces under 200 pre-holiday sessions. Treat any apparent edge with proportional skepticism.
How to read the numbers
- The Pre-holiday bucket counts sessions immediately before a holiday.
- The baseline holds every session in the sample. Compare directly.
- Delta vs baseline is the difference of means. With sample sizes this small, a delta in either direction can be noise rather than signal.
- % Up is informative independently: even a tiny mean edge with a clearly elevated share of up sessions suggests a directional bias.
What’s not here
- Day-after-holiday returns (a separate but related question).
- Holiday-by-holiday breakdown (e.g., Thanksgiving vs July 4).
- Volume conditioning (some pre-holiday sessions trade much thinner than normal).
FAQ
Is the pre-holiday effect real?
It is one of the oldest anomalies in equity research: returns on the session before a holiday have historically averaged higher than typical sessions. This page tests whether that persists in this instrument by comparing the pre-holiday bucket to a baseline of every session, with holidays inferred empirically from gaps in the data rather than a hardcoded calendar. The delta vs baseline on the page is the difference of means, and with these sample sizes a delta in either direction can be noise.
What is the pre-holiday effect?
A pre-holiday session is the last trading session immediately before a US market holiday, detected here whenever the next trading session is further away than the expected one calendar day, or more than three days for a Friday. The claim is that these sessions earn higher RTH returns than normal, often attributed to short-covering ahead of an illiquid period or behavioral anchoring. Sample size is small, only about 9 to 11 holidays per year, so even 18 years leaves under 200 such sessions. Treat any apparent edge on the page with proportional skepticism.