Splitting entry and exit speed by regime

Name the regime before you set the clocks

Regime-dependence is the design rule that entry and exit lookbacks, thresholds, and even which series is used should change once the tape is classified as rising-price or falling-price. In a rising-price regime, overbought readings can persist while oversold readings tend to last only briefly. For long trades in a rising-price regime, a faster-reacting entry indicator can be paired with a slower-reacting exit indicator. That pairing is reversed for short trades in a falling-price regime.

A two-step Williams %R entry

Williams %R locates the current close inside the highest-high to lowest-low range of a chosen lookback and is sometimes scaled so readings run from 0 to -100. In this workflow it is a lookback oscillator used as a two-step reversal trigger. Dual-threshold-reversal means the rule first records a breach of one level and then requires a later reverse crossing of a second level before acting.

Entry used a two-step Williams %R sequence: a prior move below a threshold, held in a short memory window, then a later move back above a threshold. One worked example started with a pass below -80 followed by a pass back above -80.

A slower clock on the same oscillator

The first exit used the same two-step Williams %R structure with independently chosen lookbacks and thresholds. After stepwise search over ten parameters, one entry-side lookback was 12 periods and a comparable exit-side lookback was 20 periods.

A moving-average close as the slower exit

A second long-only design kept those Williams %R entry rules and exited when the daily close fell below a simple moving average of closes, with the average length searched alongside the entry settings. The moving average is a smoothed close series used as a slower exit: the long is closed when the daily close falls back through the average.

The moving-average exit produced 28 closed trades versus 10 when Williams %R handled both entry and exit, with more turnover during a late-2002 sideways stretch that was not a clean rising-price run.

Breadth as one check of the regime

A Nasdaq 52-week new-highs divided by new-highs-plus-new-lows ratio, often smoothed with a 10-day average, was used as one breadth check of whether a rising-price or falling-price regime was in force. That ratio produced a false start in early 2002 when it rose while prices continued to fall. Yield-curve shape, the level of interest rates, employment figures, and policy actions were listed as additional checks of the same regime question.