Study

What the data says about US stocks — measured, not assumed

Measured Sep 24, 2026 · outcomes through Aug 21, 2026

Every popular signal makes a claim about what a stock does next. We measured those claims on the recomputed data behind Opulence Alpha: about 1,300 US stocks, every week since 1995, against the same day's median stock. Most signals are context, not a forecast. A few carry a small, steady edge. Here is which is which.

Five things the data says

Chart patterns are at chance.Against the median stock on the same day, none of 15 formations beat a coin flip over the next month — whatever the volume, the stage or the analyst context.
Short-run reversal is the strongest relation.Stocks that beat their industry most over three sessions lagged over the next week, in 32 of 32 years. Small per stock, steady across the market.
Volume and liquidity carry a small, steady edge.Unusually heavy volume preceded slightly better weeks; illiquid names earned more over three months — and cost more to trade.
Many famous signals are context, not forecasts.Insider buying, ownership filings, macro sensitivities and days to cover described the stock without telling which way it went next.
Every relation is small.A tilt that holds across hundreds of stocks can be real and still be a nudge for any single name. The pages say so, with the numbers.
Do chart patterns predict stock returns?Chart patternsNot on their own. Across 1,354,468 weekly-sampled detections of 15 chart patterns on US stocks since 1995, a stock showing a pattern went on to beat the median stock over the next month 47 to 51 times in 100 — the range a coin flip produces. How often a pattern reaches its drawn target depends mostly on how far away the target is.Do RSI, MACD and other indicators predict returns?RSI, MACD and indicatorsOnly weakly — and mostly the opposite way to how they are usually read. Over the next week, US stocks with RSI(14) below 30 beat the median stock 51.6% of the time and those above 70 only 48.5%; a close above the upper Bollinger band 48.6%, a new 10-day high 48.7%. That is short-run reversal, not momentum. MACD and golden crosses showed no reliable edge.Do stocks that just jumped keep winning?Short-term reversalNot on average: they tend to lag the next week. Since 1995, the fifth of US stocks that beat their industry most over three sessions beat the median stock the following week 48 times in 100; the fifth that lagged most, 51 times. A coin flip gives 50. The relation pointed the same way in 32 of 32 years: steady across hundreds of stocks, a nudge for any one.Does trading volume or liquidity predict stock returns?Volume and liquidityA little. Across US stocks since 1995, the fifth trading on the heaviest volume for its own past year beat the median stock over the next week 51 times in 100, the quietest fifth 49. Over three months the least liquid fifth beat it 52 times in 100, the most liquid 48: a liquidity premium, before the higher cost of trading illiquid stocks. Small tilts, not forecasts.Does short interest predict stock returns?Short interestA little, through short-sale volume, not days to cover. Since 2009, the fifth of US stocks with the heaviest share of volume sold short beat the median stock over the next month 49 times in 100, against 51 for the lightest, the same way in 15 of 18 years: a tilt across hundreds of stocks, a nudge for any one. Days to cover (since 2021) showed no measurable edge.Do insider buys or analyst revisions predict stock returns?Insiders and analystsMostly no. Across 22 measures of Form 4 insider buying and selling on US stocks since 2003, none showed a reliable edge within three months; analyst upgrades and price targets, at most a faint tilt. The one exception is analyst coverage, measured since 2012: over three months, the least-covered fifth of stocks beat the median stock 52 times in 100, the most-covered fifth 49. A small tilt, not a forecast.Which popular stock signals don't predict returns?Signals without an edgeMacro sensitivities, statistical factor loadings, wavelet cycles, crowding scores, 13D/13G filings, anomaly intensity and chart patterns combined with context. Of 69 such features measured on US stocks as far back as 1995, 0 cleared the bar for a measured tendency and 14 showed a faint tilt. Sorted into fifths on 4 of them, each fifth beat the median stock over the next month 49 to 50 times in 100.

How this was measured

Every Wednesday since 1995, the stocks in our universe are compared on the signal and followed for 1, 5, 21 and 63 trading sessions. A reading counts when the stock beat the median stock on the same day, so chance is always 50 in 100 and a market-wide rally cannot flatter the result. The strength of a relation is a rank correlation (IC) between the signal and the forward return; its t-statistic is taken on non-overlapping dates, and a relation is called proven only when |t| is at least 3, it held in at least 60% of years and it is large enough to matter. With hundreds of tests, a looser bar would pass about one in twenty by luck. Studies run Sep 24, 2026; outcomes measured through Aug 21, 2026.

Research, not advice. A measured tendency across hundreds of stocks is a nudge for any one of them, never a forecast of its price.