The Test Bench › indicator
Does ADX (14) actually work?
DEADNo edge left once tested on a fair footing.
The claim
ADX (14) signals when to be in or out of the market — J. Welles Wilder
What we did
Tested against Buy & hold SPY, SPY 1993-2026, thresholds swept both directions, OOS>=2014, null-floor controlled.
The result
Result: -0.04 dSharpe (vs buy-hold), OOS
OOS edge -0.04 vs buy-hold; data-mining floor +0.35 (below it = noise)
Verdict
As a mechanical buy/sell timing signal on the index, ADX (14) scored -0.04 out-of-sample -- under the +0.35 data-mining floor. It measures trend STRENGTH, not direction -- it times nothing on its own.
What the test did and did not do
- survivorship-free universe
- net of real costs
- tested out of sample
- point-in-time data
- checked for look-ahead
Almost nobody publishing a backtest says which of these they did. We list the ones we failed as well as the ones we passed.
Tested 2026-06-23. The same treatment, applied to every other system, is on the test bench.
Written with AI assistance from our own test database. The numbers are computed, not composed.
General market commentary for education, not personalised investment advice. Backtested results are hypothetical. Trading carries a real risk of loss.