How the test was run
The stock list is the 100 largest companies in the world by market value (companiesmarketcap.com, 2026-10-16) whose shares trade on the NYSE or Nasdaq: 79 American companies and 21 foreign ones through their New York listings. Daily prices come from Yahoo Finance, adjusted for splits and dividends, so every return includes dividends. A stock needs 252 sessions of history before its first signal, so 99 of the 100 were tested (SPCX listed too recently). One price-history break (a spin-off recorded as a split) is treated as a data error and the trades across it are left out.
What the results say
The top three by the ranking rule were golden cross, monthly momentum 12-1, all-time-high breakout: profit factors of 3.9 or more, trades held for 120 sessions or longer on average, and a profit in both halves of the test. They did not dodge the bear markets: in 2008 their portfolios lost -27%, -35%, -28% (S&P 500 -37%), and in 2022 -25%, +1%, -12% (S&P 500 -18%). The fast rules (trend Breakout, RSI(2) pullback, 52-week high on volume) made +0.9%, +0.2%, +1.1% per trade after costs. 3 of the 6 (trend Breakout, RSI(2) pullback, 52-week high on volume) ended below the S&P 500's ×8.79, and none beat the equal-weight basket's ×35.6.
Why the basket is the bar
These are today's winners. Testing on them since 2006 adds hindsight to every result: Nvidia, Tesla and Broadcom are in the list because they went up. The same hindsight helps an equal-weight basket of the list, so beating it is the fair test of whether a rule adds anything. None did on risk-adjusted return since 2016. What the best rules offer is a smaller worst fall than the index and a plan for when to get out. This is research, not advice; see the disclaimer.