What exactly we tested
We took 12 indicators every MT5 terminal has: EMA20/50, EMA200, MACD in two settings, DI, Parabolic SAR, Ichimoku, the Bollinger middle band, RSI50, Stochastic, CCI and Momentum. We tested them alone, in pairs and in triples, on M5 and M15 candles, with and without an ADX > 25 trend filter — 1,144 rules in total. Entry: the moment the indicators newly line up in one direction. Exit: a target and a stop at 0.12% of price (about 5 USD on gold).
Result
Not one rule made money after costs. The median result before costs was −0.14 bp per trade — practically zero, even slightly negative: after indicators line up, price more often pulls back a little than keeps going. We checked the ten best rules from the build period on 2022–2024: all of them lost there.
Why
An indicator is past price, recalculated. When it is visible in every terminal in the world, the information in it is already in the price. On short moves (a few dollars on gold) the trading cost is several times larger than any edge you can squeeze out of that information.
What it means for a trader
- A moving-average cross or an "oversold" RSI is not an edge on its own. If someone sells a strategy built only on indicators, ask for the result after costs over several years of data.
- Indicators can help describe the market (trend, volatility), but the entry decision needs information that not everyone has.
- For short-term strategies, trading costs decide the result more than the choice of indicator.
Limits
We tested one exit type (fixed target and stop), two timeframes and one instrument. Other exits and markets are covered in separate lab studies.