We Rebuild Viral Indicators and Test Them: What The Lab Has Learned

The Viral Indicator Economy Has a Testing Problem

Every week, another indicator goes viral. The video is always the same shape: a trader pulls up a chart, draws a few signals, shows a handful of winning trades, and collects views. What it never shows is a structured out-of-sample test across hundreds of assets and timeframes — because that kind of test is slow, expensive, and usually delivers an uncomfortable answer.

That is what this lab exists to do. Across 660,005 backtests covering 903 assets and 382 indicators, tested on the 1-Hour, 4-Hour, Daily, and Weekly timeframes, the results are now large enough to say something honest about what the viral crowd keeps promoting.

The Headline Number That Should Reset Your Expectations

Before you get to any specific indicator, start here: only 26% of indicator-and-asset combinations beat a passive buy-and-hold in out-of-sample testing. Read that again. Three quarters of the combinations you would find on a YouTube thumbnail failed to clear the most basic bar.

Even among assets where at least one indicator worked, only 63% of the 903 assets tested had any winning indicator at all. That leaves more than a third of the asset universe where the data says to sit on your hands. Short signals had edge in just 17.4% of cases. The 'trade both sides' pitch you hear constantly from viral content creators is not supported by these numbers.

The median best Sharpe ratio across all winning combinations was 0.62 — decent, not spectacular, and that is the top-performer figure, not the average.

The Win-Rate Trap: Where Most Viral Indicators Live

The most dangerous pattern in the data is not an indicator that fails obviously. It is one that looks like it works because it wins most of its trades, while quietly underperforming a passive position in the same asset.

Holy Grail Confluence — a staple of 'confluence is king' content — posted a 73.3% median win rate. It beat buy-and-hold in only 8% of tested assets. SMC: Liquidity Sweep printed a 71.2% win rate and beat buy-and-hold in 8% of assets. Murrey Math Lines looked better still on the surface — 74.3% wins — but cleared the bar in only 11% of assets.

Others in this trap: RSI Mean-Reversion (71.7% win rate, 10% beat rate), Money Flow Index (72.2% win rate, 9% beat rate), CCI and DeMarker (both 71.0% win rate, 9–10% beat rate), Ultimate Oscillator (72.7% win rate, 11% beat rate). A high win rate in an oscillator usually means it is catching small mean-reversion moves inside a trend — the winners are small, the rare losers are large, and buy-and-hold quietly laps the whole thing.

On Smart Money Concepts broadly: across every SMC variant tested, none beat buy-and-hold consistently. That result held regardless of asset class or timeframe.

What Actually Survived the Lab

The indicators that do survive tend to share two things: they are not viral, and they fit the specific character of their asset class rather than claiming to work everywhere on everything.

In forex, the Fisher Transform is the clearest outlier — it topped the leaderboard for 17 currency pairs, more than any other single indicator in any asset class. In stocks, Fibonacci Pivots led with 22 assets, followed by Projection Bands, Intraday Momentum Index, and Camarilla Pivots at 16 each. Crypto favors MA Envelope (5 assets), Fibonacci Pivots (4), and Delta Volume / CVD proxy (4). In commodities, Keltner Mean-Reversion leads with 3 assets.

None of these were in the most-viewed indicator videos when this project started. The gap between what gets promoted and what the data supports is precisely what the lab is here to measure. You can explore the full ranked results by asset at /assets.

How to Use This Before You Trade Any Strategy

If a viral indicator is not in the survivors above — or it is, but only works on a small fraction of assets — that is information, not a reason to dismiss it entirely. Check the indicator list to see if it was tested, find your specific asset, and look at whether it actually cleared buy-and-hold on the timeframes we cover: 1-Hour, 4-Hour, Daily, and Weekly. If it was not tested, treat it as unverified until you can run a structured out-of-sample check yourself.

The short version: win rate alone tells you almost nothing. The only question worth asking is whether the strategy beat a passive position after costs. Everything else is noise dressed up as a setup.