P-Signal Entries and Exits Using a Gaussian Error Function
Summary
The P-Signal method applies a Gaussian error function to a normalized statistic built from the simple moving average and standard deviation of price changes. The indicator is smoothed over a lookback window. It opens a long position when P-Signal is below zero and turning upward, then closes when it is above zero and turning downward. The provided default lookback is nine bars. This pairs a threshold crossing with the direction of the indicator, aiming to capture a recovery from negative readings and exit as positive readings roll over.
The document includes a BTC_USDT futures backtest configuration using daily bars from January 2023 to January 2024, but gives no results to assess performance. Its discussion warns that noisy or ranging markets may create repeated false signals, while transaction costs and slippage can weigh on a high-turnover implementation. The explanation's interpretation of the statistic is not fully reliable: the code applies the calculation to changes in OHLC4, and a positive mean-to-standard-deviation ratio maps toward positive values rather than indicating that fluctuations exceed the mean. Parameter sensitivity and the assumed return distribution also warrant testing.
Key ideas
- P-Signal maps a normalized moving-average-to-standard-deviation statistic through a Gaussian error function.
- The strategy opens long when P-Signal is negative and rising, and exits when it is positive and falling.
- The default observation length is nine bars, with the indicator smoothed over its lookback window.
- The code calculates the signal from changes in OHLC4, which qualifies the article's explanation of the statistic.
- No backtest results are provided, and false signals, turnover, fees, and slippage remain unresolved risks.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.