DIA 'Turnaround Tuesday' over the last ~3 years — red-Monday conditioning
The celebrated "Turnaround Tuesday" doesn’t show up in DIA over the last ~3 years. I measured close-to-close Tuesday returns conditional on the prior day being a down Monday (N=58) and compared them to Tuesdays after non-red Mondays and to the unconditional baseline. The average Tuesday after a red Monday was only +0.0629%, with the mean-difference test returning p=0.682 — well inside the range of random variation.
You’ll find the full statistics, charts, and resampling details below, but the bottom line is clear: both mean returns and the fraction of positive Tuesdays are statistically indistinguishable from their complements, and the tiny observed gap is small relative to daily volatility. Scroll down for the data and methodology that support this conclusion.
For DIA (the Dow) over the past ~3 years, is 'Turnaround Tuesday' real — after a red Monday, does Tuesday rebound more reliably and more strongly than a typical session, or is it just folklore? Thesis: Tuesday returns following a down Monday come out statistically indistinguishable from the everyday baseline, so the celebrated 'Turnaround Tuesday' bounce is a story traders tell rather than a tradeable edge.
How this was measured
Resampled DIA minute bars to daily closes, computed close-to-close returns, and restricted the window to the last ~36 months (2023-07-03 to 2026-06-30). Flagged Tuesdays whose immediately preceding trading day was a Monday and had a negative return (‘red Monday’). Compared their Tuesday returns against Tuesdays after non-red Mondays (disjoint complement) and against unconditional baselines. Reported mean/median/std, fraction-positive, Welch’s two-sample t-test for mean differences, and a two-proportion z-test for positive-rate differences.
The key numbers
Reading the numbers
Across 58 Tuesdays that followed a down Monday the mean close-to-close return was 0.00062863 versus 0.00118471 for Tuesdays after non-red Mondays; the difference is -0.00055608 and the Welch p-value is 0.6823, so the means are statistically indistinguishable.
The charts
The box plots put Tuesday-after-red-Monday (n=58, mean 0.0006, min -0.013, max 0.0397) next to Tuesday-after-non-red (n=84, mean 0.0012, min -0.0306, max 0.0277) and all days (n=751, mean 0.0007, min -0.0512, max 0.0965). What to notice is the heavy overlap of the boxes and whiskers — medians and ranges sit on top of one another rather than showing a clear separate band for post-red-Monday Tuesdays. That visual overlap means there is no obvious stronger or more consistent rebound on Tuesdays after red Mondays compared with other groups in this sample.
The bar chart of mean returns shows Tuesday-after-red-Monday at 0.0006, Tuesday-after-non-red at 0.0012, All Tuesdays at 0.0008, and All days at 0.0007. Those differences are tiny (on the order of one ten-thousandth to one thousandth) and, given the Welch p-value of 0.6823, the mean advantage for Tuesdays following a down Monday is not statistically clear. In short, the mean-bars reinforce the box-plot story: no reliable Turnaround Tuesday edge in this period.
Return summary by group
| group | N | mean | median | std | frac_positive |
|---|---|---|---|---|---|
| Tue after red Mon | 58 | 0.0006 | 0.0002 | 0.008 | 0.5345 |
| Tue after non-red Mon | 84 | 0.0012 | 0.0006 | 0.0079 | 0.5714 |
| All Tuesdays | 156 | 0.0008 | 0.0003 | 0.0081 | 0.5449 |
| All days | 751 | 0.0007 | 0.0007 | 0.0088 | 0.5486 |
The takeaway
No — over the past ~3 years there’s no evidence that ‘Turnaround Tuesday’ is a reliable effect in DIA. The average Tuesday after a red Monday was only +0.06% (N=58) versus +0.12% for Tuesdays following non-red Mondays (N=84); the fraction of positive Tuesdays was 53.4% vs 57.1%, while the baseline Tuesday mean was +0.08% (all days +0.07%). Statistical tests back this up: the mean-difference test returned p=0.682 and the positive-rate test p=0.663, which means the observed gaps are well within what you’d expect from random chance. Put another way, the tiny mean gap is swamped by daily volatility (std ≈ 0.8% in these groups), so the slight underperformance after red Mondays is not meaningful. Practical takeaway: Turnaround Tuesday looks like folklore here — not a repeatable, tradeable edge over this window.
The fine print
- Only 58 Tuesdays follow a red Monday in this 3‑year window — thin sample weakens power.
- Analysis uses close-to-close returns only; intraday open-to-close rebounds were not tested.
- No controls for volatility regime, macro news, earnings clustering, or other confounders.
- Holidays (Mondays replaced by weekend/holiday rules), ETF microstructure and dividend ex-dates could slightly bias daily returns.