Backtest evidence

Is there one stochastic setting that works on every timeframe?

Best Stochastic Settings: 2,700 Tests of K/D and Smoothing

Finding

This test did not find one. K/D crosses were profitable in both 2024 and 2025 for 1 of 75 settings on the 15-minute chart, 43 on the 1-hour and 56 on the 4-hour. On the 1-hour chart, all 15 K/D period combinations with smoothing set to 20 were profitable in both years, which does not establish future results. The common 14/3/3 cross returned +686.6 pips over 1,525 trades in 2024 and −49.7 pips over 1,510 trades in 2025 on the 1-hour chart, so it was not profitable in both years. Win rate was no guide: in 2025 the 1-hour 14/3/3 long fade won 60.38% of 106 trades and lost 574.9 pips, with an average win of 46.83 pips against an average loss of 85.05. A level-only rule never reads D, so changing D does not change that trading condition.

Key results

K/D crosses profitable in both years
1 of 75 on 15-minute, 43 on 1-hour, 56 on 4-hour
The common 14/3/3 cross on the 1-hour chart
+686.6 pips over 1,525 trades in 2024; −49.7 pips over 1,510 trades in 2025
The 14/3/3 long fade on the 1-hour chart in 2025
106 trades, 60.38% win rate, −574.9 pips

Scope

  • USD/JPY: 2025, 2024 and the two halves of 2025, each run independently
  • 15-minute, 1-hour and 4-hour charts
  • 5 K lookbacks x 5 smoothing periods, with 3 D periods for crossovers and 3 level pairs for fades (D held at 3): 75 settings per rule, 3 rules x 3 timeframes x 4 periods = 2,700 runs
  • Candle-close fills, fixed 0.3 pip spread, zero slippage, 0.1 lot

Method

Measured with Formiq's backtester. Crossovers enter on a K/D cross and exit on the opposite cross; fades enter at a level and exit at the opposite level. Positions still open at the end of each period are closed. Pre-start candles are provided for the indicator calculation. No stop or target in the main sweep.

Limitations

  • The sample covers USD/JPY in 2024 and 2025, with candle-close fills and fixed spreads.
  • Variable spreads, execution differences and swap charges are excluded.
  • Pre-start candles are provided for indicator calculation, and positions still open at each period's end are closed.
  • Beyond the bars needed to compute K, smoothing and D, past bars are needed to detect a cross.