Experiments · Opulence Alpha Research

Popular beliefs, tested.

Name: Opulence Alpha evidence snapshot — what signals did next on US stocks
Creator: Opulence Alpha
License: https://creativecommons.org/licenses/by/4.0/
Keywords: stock market evidence, do technical indicators work, chart patterns backtest, technical analysis backtest, trading myths, short interest, insider buying, analyst coverage, factor research, evidence-based investing, US equities

24 ideas repeated on every trading forum, each measured the same way: on the same 1,767 US stocks, every week since 1995, against the same day’s median stock. None showed a proven edge in the direction traders believe; 4 went measurably the other way.

Evidence Drop · 25 Sep 2026 · data through 21 Aug 2026

Proven edge in the direction traders believe0of 24 beliefs20 show no edge · 4 went measurably the other way

What each group of beliefs found

One cell per belief, with its score out of 100 against the same day’s median stock. Hover a cell for its belief; select a row to filter the scoreboard.

No edgeWent the other way, provenScored on reaching a drawn target

No proven edge20of 24

Went the other way4of 24, proven

Tests across all studies1,470about 4 would pass |t| ≥ 3 by luck

One universe1,767US stocks, 1995–2026

The scoreboard

Every belief, its verdict and its headline number

Select a row for the full paper. The grey band is the range chance alone produces.

BeliefHorizon455055Of 100Verdict

Chart patterns8 beliefs

Head and shouldersDoes the head and shoulders pattern predict a fall?1 month50No edge Double bottomDoes a double bottom predict a rebound?1 month50No edge Bull flagDoes a bull flag lead to the next leg higher?1 month51No edge Ascending triangleIs the ascending triangle really a bullish pattern?1 month50No edge Rounding bottomDoes a rounding bottom signal a rally?1 week48Went the other way Falling wedgeIs the falling wedge a bullish pattern?1 month51No edge Volume confirmationDoes high volume confirm a chart pattern?—Reached the drawn target · Volume below normal → Volume 1.5× normal or more23 → 16No edge Reward-to-riskIs a 3:1 reward-to-risk setup a better trade?—Reached the drawn target · Reward under 1.5× risk → Reward 3× risk or more30 → 2No edge

Technical indicators6 beliefs

Golden crossDoes the golden cross predict a rally?1 month50No edge MACD crossoverDoes the MACD crossover predict a rally?1 week49No edge Breakout to new highsDo stocks that break out to a new 10-day high keep running?1 week49Went the other way Bollinger band breakoutDoes a Bollinger band breakout lead to more gains?1 week49Went the other way ADX strong trendDo stocks with an ADX above 25 go on to beat the rest?1 month50No edge Gap upDo stocks that gap up keep rising?1 week50No edge

Momentum, risk and style4 beliefs

Momentum winnersDo momentum winners keep winning?1 month50No edge Near the 52-week highDo stocks near their 52-week high keep outperforming?3 months49No edge Beaten-down stocksDo beaten-down stocks bounce back?3 months49No edge Value vs growthDo stocks that trade like value stocks beat growth stocks?3 months50No edge

Analysts, insiders and short sellers6 beliefs

Analyst upgradesDo analyst upgrades predict which stocks do better?1 month50No edge CEO and insider buyingWhen a CEO or CFO buys the stock, does it do better?1 month50No edge Activist stakesDo stocks beat the rest once an activist's 13D is public?1 month49No edge Short squeezeDo heavily shorted stocks squeeze higher?1 month49No edge Crowded tradesDo crowded trades end badly?1 month50No edge Most-covered stocksDo the stocks analysts follow most do better?3 months49Went the other way

The research papers

Whole signal families, each a short paper

Proven, higher returnsProven, lower returnsNo edge

Studies of whole signal families, each a short paper with its data, method and robustness. The bar is the lead relation’s information coefficient; years shows how many of the years tested it held in.

PaperQuestionInformation coefficienttYears it heldVerdict

Chart patternsDo chart patterns predict stock returns?———No edge RSI, MACD and indicatorsDo RSI, MACD and other indicators predict returns?−0.016−5.027/32Proven Short-term reversalDo stocks that just jumped keep winning?−0.027−12.432/32Proven Volume and liquidityDoes trading volume or liquidity predict stock returns?0.0118.628/32Proven Short interestDoes short interest predict stock returns?−0.017−3.115/18Proven Insiders and analystsDo insider buys or analyst revisions predict stock returns?−0.029−3.414/15Proven Signals without an edgeWhich popular stock signals don't predict returns?0.0070.722/32No edge

Data and methods: the universe, the sample, the statistics and the multiple-testing bar behind every study →

How it is measured

One universe, one method, one bar

The universe1,767 stocks

Every S&P 500 member since 1996, plus large and mid caps chosen in proportion to sector. 422 have since delisted and stay in.

The methodvs the median stock

Every Wednesday since 1995, each signal is scored against the same day’s median stock, so chance is 50 in 100 whatever the market did.

The bar|t| ≥ 3

Proven only at |t| ≥ 3 on non-overlapping dates, in at least 60% of years. Of 1,470 tests, about 4 would clear it by luck; 176 did.

Every Evidence Drop

The research release every figure on these pages quotes

The list, as JSON →

ReleasedData throughFile

25 Sep 202621 Aug 2026The data, as JSON →

Each release is frozen as it shipped and never changed, so a figure quoted from an earlier drop can be checked against the file it came from. Research, not advice. A measured tendency across hundreds of stocks is a nudge for any one of them, never a forecast of its price.