@quantpad: Turn of the month effect backtest: the last trading day plus the first three, tested on S&P 500 and Nasdaq futures 2011-2025 McConnell and Xu (Financial Analysts Journal, 2008) found that from 1926 to 2005, US stocks paid investors nothing except on those four days around the turn of each month. I replicated it on every month since 2011 with one-minute E-mini S&P and Nasdaq futures data, measuring each day close to close from the 3:59 p.m. bar, on one contract. It did not hold. On the S&P, the turn-of-the-month days averaged +0.96 points against +1.67 on every other day, so 18% of the days carried 11.5% of the points. 673 of 1,000 random four-day windows did better. The Nasdaq said the same: +3.90 against +5.68. As a trade (buy the close two days before month end, sell the close of day 3, $16.50 a round turn), one ES contract made $5,518 in 15 years over 158 trades, against $275,608 for holding every day. All 20 nearby entry and exit windows on both markets made less than holding, and at $33 a trade the study's window made $2,911. Walked forward (expanding yearly, window picked on past years only, 2014-2025), the S&P made $63,510 out of sample, mostly by holding eight days instead of four. Robustness: filling in holiday early closes instead of skipping them made the four days look worse, not better (+0.53 vs +1.69). What this does not show: 22 of 180 months had no clean entry or exit price (mostly holiday early closes), and the first two days of the month did average more than a normal day, but not significantly (t = 0.52). The takeaway: don't sit out the month for four magic days. I built this in QuantPad. It has every trade on the CME since 2010. Link in bio. #QuantPad #backtesting #futurestrading #quanttrading #tradingstrategy

QuantPad

Open In TikTok:

Region: US

Sunday 01 November 2026 05:04:48 GMT