forex-scalping.com

What a Regime Filter Does to the Totals

A strategy took 100 trades. The 60 trades taken when ADX was above 25 averaged plus 0.30R. The 40 trades taken when ADX was below 25 averaged minus 0.20R.

Compute the unfiltered total. High-ADX: 60 times 0.30 is plus 18R. Low-ADX: 40 times minus 0.20 is minus 8R. Total: plus 10R, an average of 0.10R per trade.

Apply the filter, skipping trades when ADX is below 25. Total: plus 18R over 60 trades, an average of 0.30R per trade. The filter raised the total by 8R, and the per-trade average tripled.

What is the cost? Forty fewer trades means fewer opportunities, and if costs per trade are 0.05R, the filter also saves 2R in costs.

Check reliability. Is the difference between the two groups meaningful? With standard deviations of about 1.5R, the standard error for 40 trades is 0.24R, and for 60 trades it is 0.19R. The gap between 0.30 and minus 0.20 is 0.50R, compared with a combined standard error of about 0.31R. That is 1.6 standard errors, suggestive but not conclusive.

Also note the arbitrary threshold. If ADX 22 or 28 gave similar results, the idea is robust. If only 25 works, suspect fitting.

Validate on new data, and keep the filter simple.

Exercise: split your trades by one condition, such as trend strength, and compare the groups.

The rate of incoming ticks says something about the market's state. A pair that normally produces 3 ticks per second and suddenly produces 15 is telling a story, even before price moves.

Rising tick rate usually accompanies rising volatility. A scalping strategy can use this as a filter: only trade when the tick rate is above a threshold, which implies enough activity for a move to develop. Or only trade when it's below, for mean-reversion setups that need calm.

Because it comes straight from the feed, the measure needs no indicator or price calculation. A simple counter over a rolling 10-second window does the job.

The filter also helps to avoid the dead hours. A tick rate that falls below a minimum signals that the market isn't worth trading.

The calibration is broker-specific. Thresholds that work on one feed may not translate to another, so each setup needs its own baseline, measured over weeks.

It's one of the cheaper regime indicators, and it relies on something that most platforms already deliver.

What threshold, if any, made ticks per second a useful trading filter?