Calendar Effects
Intra-month and day-of-week seasonal tendencies — the short-timeframe complement to the annual seasonality tool.
Today’s edge · S&P 500 (10y of data · for Fri 30 Oct)
For S&P 500 on Fri 30 Oct: Fridays: median +0.079% (55% up, n=526); trading day 22: median -0.046% (n=39); inside the Turn of month window (median +0.574%/instance, 67% positive, n=126). Historical tendency, not a promise — size accordingly.
Average return by day of week
Average return by trading day of month
Monthly seasonality · S&P 500
■ median■ mean· win rate below · current month highlighted
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Win rate | 70% | 40% | 50% | 70% | 90% | 80% | 91% | 64% | 45% | 50% | 90% | 60% |
| 2026 | +1.4 | -0.9 | -5.1 | +10.4 | +5.1 | -1.1 | -0.1 | +2.6 | -0.5 | -0.7 | – | – |
| 2025 | +2.7 | -1.4 | -5.8 | -0.8 | +6.2 | +5 | +2.2 | +1.9 | +3.5 | +2.3 | +0.1 | -0.1 |
| 2024 | +1.6 | +5.2 | +3.1 | -4.2 | +4.8 | +3.5 | +1.1 | +2.3 | +2 | -1 | +5.7 | -2.5 |
| 2023 | +6.2 | -2.6 | +3.5 | +1.5 | +0.2 | +6.5 | +3.1 | -1.8 | -4.9 | -2.2 | +8.9 | +4.4 |
| 2022 | -5.3 | -3.1 | +3.6 | -8.8 | +0 | -8.4 | +9.1 | -4.2 | -9.3 | +8 | +5.4 | -5.9 |
| 2021 | -1.1 | +2.6 | +4.2 | +5.2 | +0.5 | +2.2 | +2.3 | +2.9 | -4.8 | +6.9 | -0.8 | +4.4 |
| 2020 | -0.2 | -8.4 | -12.5 | +12.7 | +4.5 | +1.8 | +5.5 | +7 | -3.9 | -2.8 | +10.8 | +3.7 |
| 2019 | +7.9 | +3 | +1.8 | +3.9 | -6.6 | +6.9 | +1.3 | -1.8 | +1.7 | +2 | +3.4 | +2.9 |
| 2018 | +5.6 | -3.9 | -2.7 | +0.3 | +2.2 | +0.5 | +3.6 | +3 | +0.4 | -6.9 | +1.8 | -9.2 |
| 2017 | +1.8 | +3.7 | +0 | +0.9 | +1.2 | +0.5 | +1.9 | +0.1 | +1.9 | +2.2 | +2.8 | +1 |
| 2016 | – | – | – | – | – | – | +3.6 | -0.1 | -0.1 | -1.9 | +3.4 | +1.8 |
Monthly returns, %. Win rate and the median/mean bars exclude the current (partial) month; the year rows include it as month-to-date.
The documented calendar windows · measured on S&P 500
| Window | Median | Mean | % pos% positive | n | Status |
|---|---|---|---|---|---|
| Turn of month | +0.574% | +0.468% | 67% | 126 | active now |
| OPEX week | +0.02% | -0.05% | 50% | 125 | — |
| Mid-month lull | +0.269% | +0.066% | 58% | 125 | — |
Per-instance cumulative returns. “Active” refers to the session in the Today’s-edge card (Fri 30 Oct); the current month’s last-two-sessions test uses the US exchange calendar (approximate), not the bars printed so far.
The 60-second version
The annual Seasonality tool answers “which months favour this asset”. This one answers the day trader’s version: which days. Markets have documented micro-rhythms — the turn-of-month institutional bid, the mid-month lull, options- expiry pinning, and persistent day-of-week tilts (crypto trades all seven). Each is measured here from ~10 years of daily bars for eight majors, and the search box computes the same profile for any Yahoo ticker on demand — your exact stock, ETF or coin, not a proxy.
How a short-timeframe trader uses it
- Start with Today's edge — the card at the top pre-combines the weekday, trading-day-of-month and every active named window into one sentence — the calendar's read on the next session (its date is on the card; weekends roll to Monday), refreshed hourly. It leads with the median, and says when a few outlier days have dragged the mean away from it.
- Time entries with the month's shape — the trading-day chart shows where in the month the asset historically gets its bid. Buying planned exposure into the mid-month lull and holding through the turn-of-month has been the structural pattern; the windows table quantifies it with hit rates.
- Respect the weekday tilt at the margin — day-of-week effects are small per instance but persistent across hundreds of observations. They're tie-breakers for WHEN to execute a decision already made — never the reason for the decision.
- Check your own ticker — single names have their own rhythms (earnings drift, index-rebalance flows). Type the ticker, get the same three panels computed on its actual history with sample sizes shown.
Methodology & honesty
All effects are computed from raw Yahoo daily closes (~10 years, seven-day weeks for crypto): day-of-week and trading-day-of-month show the median and the mean with hit rates and sample sizes on hover; the named windows measure per-instance cumulative returns (turn-of-month = last 2 + first 3 sessions, OPEX week = the week containing the 3rd Friday, mid-month = trading days 10-15). For the month in progress, “last 2 sessions” and today’s trading-day number come from the US exchange calendar (regular holidays; one-off closures aren’t modelled), not from the bars printed so far. Tickers with under ~2 years of history are flagged and charted by median — a single listing-day spike can otherwise pass for a weekday edge. These edges are real but SMALL — a few basis points per day on average — so they pay as execution-timing improvements on trades you were already making, not as standalone strategies. Costs and slippage would eat them traded naked.
In the MTS framework
Calendar effects are the finest-grained seasonal layer — beneath annual seasonality and the election cycle, which all sit in Space. Use them last: after Motion picks the trade, Time confirms the window and the bigger seasonal tools frame it, this page picks the day. A planned entry that can wait two sessions for the turn-of-month bid, or an exit moved ahead of a historically weak weekday, is the entire practical application.