Research guides
How to read a calendar-rule result, explained with the figures of the published studies. Start with how the research pages are computed.
Average winners and losers: what a hit rate leaves out
Two SPY calendar windows lost more in an average losing year than they gained in a winning one (payoff ratios 0.62× and 0.75×). How often each lost made the difference. The ratios, defined.
How much can a round trip cost before a calendar rule stops paying?
Break-even arithmetic on two SPY rules: the turn of the month's lead over holding is 9.0 basis points a trade, twelve times a year; November's is 173 basis points, once a year. Illustrative round-trip costs from 0 to 20 basis points.
Before the cutoff, after it, after publication: three kinds of seasonal result
SPY's turn-of-the-month and FOMC-day rules completed 16 trades after the 31 December 2025 research cutoff, all before the rules were published on 17 October 2026. Which results are history, which are after the cutoff, and which test a fixed rule.
Same rule, three ETFs: how one extra year reorders a November comparison
On each fund's own history QQQ's November excess (+1.87 pp) tops SPY's (+1.73 pp); on the shared 2000–2025 years SPY leads, +1.74 pp to +1.48 pp. How to compare backtests on the same years.
Mean versus median seasonal returns: what the gap reveals
SPY's October 31 window: median year +1.38 pp over holding, mean +0.12 pp. November: +1.83 pp and +1.73 pp. What the gap between the two says, with every year shown.
Why a high seasonal win rate can still mean little advantage
SPY's October 31 window rose in 69.2% of years and November in 85.2%, yet their average excess over holding was +0.12 pp and +1.73 pp. How to read a win rate, with four real rules.