Why do most 15-minute strategies fail?

Three reasons: overfitting to a single period, dependence on one market regime, and conclusions drawn from too few trades. Out-of-sample testing exposes all three.

Which strategy survived the 8-year test?

On NQ, the walk-forward-validated ICT Silver Bullet was the only setup profitable both in-sample and out-of-sample.

Why Most 15-Minute Strategies Fail on ES (and What Survives 8 Years)

We put 15 well-known setups — EMA pullbacks, Donchian breakouts, RSI exhaustion, VWAP fades, ICT setups and more — through 8 years of ES 15-minute data. The result is a useful gut-check: almost none had a durable edge.

The three killers

Overfitting. A setup that prints PF 4 on 5 years can collapse to a loser on 8. Our ICT OTE test went from PF 4.14 to 0.65 once we added more history.
Regime dependence. Many setups only work in one market character (trend vs range) and bleed out the rest of the time.
Sample size. "Great" results on 5–20 trades are noise, not signal.

What survives

The traits that held up: a large trade sample, profitability in both in-sample and out-of-sample windows, and a regime filter so the setup only fires where it has an edge. On NQ, only the walk-forward-validated ICT Silver Bullet made the cut.

Test before you trade

Do not take any strategy — ours or a guru's — on faith. Run it on real data first with BacktestAI, and get the daily market context in our free "What Happened Today" recap.

Educational content only — not financial advice. Trading futures involves substantial risk of loss. Past backtested performance does not guarantee future results.