We tested 108 technical and psychology formulas on S&P 500 stocks since 2005. The winner: stocks in a strong, steady trend that nobody is excited about yet. Every number, including what didn't work.
Most 'winning formula' claims rest on examples picked in hindsight. We tested it the hard way: every stock that was actually in the S&P 500 in each month since 2005 (including those that later left it), 108 variants of technical and psychology formulas, and what holds up over 20 years.
To avoid fooling ourselves, the formula was chosen on 2005-2015 only. 2016 to today was kept aside as a blind test that played no part in the choice. Total returns with dividends, monthly rebalancing, and a 0.1% cost per side of every trade.
Step 1: rank every index stock by trend quality — a strong rise over time with relatively low volatility (quality momentum). Keep the top 30%.
Step 2: from those, take the 30 stocks with the lowest FOMO score — rising, but without euphoria: no sharp jump in recent weeks, no abnormal volume, no extreme distance from their averages. In plain words: strong stocks the crowd hasn't discovered yet.
The portfolio is always invested, equal-weighted, and updated monthly. That's all — no forecasts, no macro calls, no human judgment in the loop.
13.0% a year versus 9.5% for the S&P 500. $10,000 invested in 2005 would have grown to about $70,180 by the end of 2020, versus about $42,678 in the index.
It beat the index in 13 of 16 years. Its biggest strength is in bad years: in 2008 it fell 24.7% while the index fell 34.3%. Its weakness: in sharp, narrow rallies (2016, 2020) it lagged — 'calm' stocks miss the phase when everyone piles into the same names.
Importantly, on the blind test (2016 to September 2026), which played no part in choosing the formula, it returned about 16.8% a year versus 15.6% for the index, beating it in 7 of 11 years. The edge is smaller than in the selection period — exactly what an honest test should show.
None of the 108 variants beat the index 10 years in a row. Anyone promising that is worth asking which period they tested.
On small caps (S&P 600) the formula did not work; on mid caps (S&P 400) it did. Six further improvements we tested (a market-trend filter, a more concentrated portfolio, volatility weighting and others) did not improve the blind-test result.
Risk: the portfolio is always in the market, so in a crash it falls with it — the largest drawdown was 45% in 2008-2009. It also trades a lot: at high trading costs (0.3% per trade) the 2005-2020 return drops to 9.1% a year, about the same as the index.
Data limitation: 306 of the 997 stocks that were in the index over the years have no price history in our data source, mostly delisted or acquired companies. That may bias the results slightly upward, so we present them with caution.
This portfolio runs on the site as a live model portfolio since October 2026 — the current 30 stocks, performance versus the index, and every monthly rebalance — under 'Buy the Pullback — No Filter' (stockiqai.com/en/portfolios/pullback-no-filter).
This is historical research, not investment advice: past performance does not guarantee future returns.
Want to see this mechanism on a real stock? Try X-Ray or analyze a stock now
More research
This research is intended for general educational purposes only, does not constitute investment advice, and is not a prediction. Full details on the disclaimer page.
