ADX thresholds: 20, 25, or something else?

[Fairman]

ADX thresholds: 20, 25, or something else?

ADX measures trend strength, not direction. A common rule is that ADX above 25 means a trending market and below 20 means a range. Lots of people take those numbers as fixed, but they're just conventions.

On lower timeframes, ADX values behave differently than on daily charts where those thresholds came from. On M5, ADX can jump above 25 during a short burst and drop back quickly. It might rarely stay above 25 for long. The thresholds that make sense depend on the pair, the timeframe and the period setting.

Another thing that's easy to miss: ADX lags. It's a smoothed average of smoothed values. By the time ADX rises above 25, a good part of the move has often happened. When it starts falling from a high level, the trend might be fading, or it might just be pausing.

The direction of ADX can be more useful than the level. Rising ADX means the trend, whichever way it goes, is getting stronger. Falling ADX means it's losing strength.

I'd look at ADX on the pairs and timeframe you trade over a few weeks and see what values typically appear in clear trends and clear ranges. Then set thresholds from that instead of using 25 because it's the default.

[Fairman]

The period also changes things a lot. The default is 14, but on M5 some people use shorter periods to react faster, or longer ones to filter out short bursts. A shorter period makes ADX jumpy and quicker to cross thresholds, so the 25 rule triggers more often. Changing the period without adjusting thresholds can make a filter almost meaningless. If you change one, revisit the other. It helps to test them together rather than separately. Changing both without testing is how filters end up doing nothing at all. Retest after any change.

[Fairman]

The DI lines that come with ADX matter too. Plus DI above minus DI suggests upward pressure, and the reverse for down. Combined with ADX, it gives direction and strength together. The crosses of the DI lines can be noisy on lower timeframes, so I wouldn't use them as entry signals. But checking which DI is on top when ADX is rising gives a quick read on which side has the momentum at that moment. It's context rather than a signal, which is probably the best way to use the whole indicator on lower timeframes anyway.

[FXS Prop Desk]

On M1 and M5, ADX values tend to be noisier, and the 25 threshold may be crossed many times in a session. Some traders use a higher threshold on lower timeframes, or look at the direction of ADX rather than the level.

A rising ADX from a low base often signals the start of a trend better than a fixed number. Testing different thresholds on your own pair and timeframe is the only reliable way to choose.

[Fairman]

ADX is slow to drop after a trend ends, which catches people out. A strong trend pushes ADX up, and when the trend stops and price starts ranging, ADX takes a while to fall below the threshold. In that time, a strategy using ADX above 25 as a trend filter might keep taking trend trades in what is really a range. Watching the direction of ADX, not just its level, helps. A falling ADX above 25 often means the trend is fading, even though the level says it's still strong.

[Fairman]

Something I've found useful is comparing ADX across timeframes. If M5 ADX is high but H1 ADX is low, the M5 trend might just be a short burst within a bigger range. If both are high, the trend is more established. Using the higher timeframe ADX as a filter and the lower one for timing can reduce the number of trend trades taken in ranging markets. It's simple to set up and doesn't need any new indicators, just the same one on two charts. The cost is a few missed trend trades early in new moves, which seems a fair trade.