Monthly DCA: $1,000 a month into last month’s best S&P 500 stock

Method

• Universe: S&P 500 members on the last trading day of the signal month, from the fja05680/sp500 historical components (the list runs through ; later months use that list).
• Signal: total return over the prior calendar month, from Yahoo Finance dividend- and split-adjusted closes (last trading day of the month vs the month before). Ties go to the first ticker alphabetically.
• Buy: at the close of the first trading day of the next month, after the signal is known, so there is no lookahead. Luo buys at that day’s open, which is slightly different.
• Strategy: $1,000 every month into the winner, never sold and never rebalanced. Repeat winners get another $1,000. Extra lines: last month’s top 5 at $200 each, and the stock with the best trailing 12-month return at $1,000.
• Benchmark: $1,000 into the S&P 500 Total Return index (^SP500TR) at the same close each month. Money-weighted return (XIRR) uses the actual dated deposits.
• Data screens: a stock is skipped in a month if (a) Yahoo has no prices for it, (b) its Yahoo history starts after it joined the index, which means the symbol was later reused by another company, (c) more than 30% of its daily closes over the past year were unchanged, a sign of an illiquid stand-in rather than an S&P 500 company, or (d) it is a known recycled symbol (El Paso “EP”, MCI “MCIC”, Compuware “CPWR”, Sonat “SNT”, and Johnbrick123/sp500-data’s list). Missing returns are never filled in.
• Delisted holdings: none of our picks stopped trading, because the screens leave only stocks with prices through today. That is exactly the survivorship problem described below.

Why our number is higher than his, and why neither is the true answer

Data

• dca_lots.csv (every lot, all strategies) · dca_coverage.csv (members, ranked, screened and missing tickers per month) · dca.json · rebuilt with scripts/build_dca.py.
• This is not investment advice.