October 31 on SPY: 15 sessions from Halloween
Has SPY done better in the 15 sessions from October 31 than in any 15 sessions?
From 2000 to 2025, this rule did better than holding SPY for the same number of sessions in 16 of 26 years. Its average trade returned +0.68% gross, against +0.56% for holding SPY over the same number of sessions: an average excess of +0.12%.
Years it did no better than holding: 2000, 2003, 2007, 2008, 2010, 2011, 2012, 2015, 2018, 2025.
The Halloween effect is usually stated as owning stocks from late October to spring. This page tests a narrower, fixed rule: enter at the first open on or after October 31 and hold 15 sessions, compared with holding SPY for any 15 sessions.
Named after “the Halloween effect”, usually a claim about the run from late October to spring. This checks a fixed 15 sessions from October 31, whatever weekday that lands on: a stand-in for the informal claim, not the claim itself. No registered strategy tests that claim; its study did not hold up.
Research through 2025-12-31; prices through 2026-10-16. Gross, before any trading cost or tax. Computed on history: see the hypothetical performance notice and how these pages are computed.
SPY
Enter October 31, hold 15 sessions
Every figure is gross, and compares each trade with holding SPY for the same number of sessions.
Summary
+0.68%
+0.56%
+0.12%
+1.38%
+0.12%
62%
16 of 26 years
69%
18 of 26 trades
69%
18 of 26 years
+1.88%
65%
15 of 23 windows
-2.39%
-0.21%
26Trades 26 1.0First half -1.46%Second half +1.71%Best year, per trade 2023 +8.34% 2008 -19.09%
: 2026, out of sample
Every trade that closed after 2025-12-31, through 2026-10-16. None of them reached a figure above: this is the rule on prices it was not measured on. Gross, at the open.
Trades closed
0
Return, summed
-
Excess, summed
-
over holding SPY the same length
Position now
None
Next trades
- Enter at the open of Nov 2, 2026, exit at the open of Nov 23, 2026
- Enter at the open of Nov 1, 2027, exit at the open of Nov 22, 2027
- Enter at the open of Oct 31, 2028, exit at the open of Nov 21, 2028
Yearly excess with rolling 4-year average
Each bar is a year's mean excess per trade over holding SPY for the same length. The line is the mean of four consecutive years, drawn at the last of them; a missing year breaks it.
Every year (26)
| Year | Trades | Return | Mean excess | Total excess | |
|---|---|---|---|---|---|
| 2000 | 1 | -4.35% | +0.56% | -4.91% | -4.91% |
| 2001 | 1 | +7.11% | +0.56% | +6.55% | +6.55% |
| 2002 | 1 | +3.28% | +0.56% | +2.72% | +2.72% |
| 2003 | 1 | -1.10% | +0.56% | -1.66% | -1.66% |
| 2004 | 1 | +3.48% | +0.56% | +2.93% | +2.93% |
| 2005 | 1 | +4.04% | +0.56% | +3.49% | +3.49% |
| 2006 | 1 | +1.75% | +0.56% | +1.20% | +1.20% |
| 2007 | 1 | -7.08% | +0.56% | -7.63% | -7.63% |
| 2008 | 1 | -18.53% | +0.56% | -19.09% | -19.09% |
| 2009 | 1 | +6.33% | +0.56% | +5.77% | +5.77% |
| 2010 | 1 | +0.52% | +0.56% | -0.03% | -0.03% |
| 2011 | 1 | -5.47% | +0.56% | -6.03% | -6.03% |
| 2012 | 1 | -1.79% | +0.56% | -2.35% | -2.35% |
| 2013 | 1 | +1.60% | +0.56% | +1.05% | +1.05% |
| 2014 | 1 | +2.90% | +0.56% | +2.35% | +2.35% |
| 2015 | 1 | +0.51% | +0.56% | -0.05% | -0.05% |
| 2016 | 1 | +2.93% | +0.56% | +2.38% | +2.38% |
| 2017 | 1 | +0.78% | +0.56% | +0.22% | +0.22% |
| 2018 | 1 | -1.77% | +0.56% | -2.32% | -2.32% |
| 2019 | 1 | +2.22% | +0.56% | +1.67% | +1.67% |
| 2020 | 1 | +8.20% | +0.56% | +7.65% | +7.65% |
| 2021 | 1 | +2.30% | +0.56% | +1.75% | +1.75% |
| 2022 | 1 | +2.12% | +0.56% | +1.57% | +1.57% |
| 2023 | 1 | +8.89% | +0.56% | +8.34% | +8.34% |
| 2024 | 1 | +3.10% | +0.56% | +2.54% | +2.54% |
| 2025 | 1 | -4.38% | +0.56% | -4.93% | -4.93% |