Logistic Price Forecasting with Volume Imbalance

Summary

This proposed long-only strategy uses price rate of change to separate volume associated with rising and falling prices, then forms a net-volume measure. It standardizes that measure using its moving average and standard deviation, and combines the resulting z-score with the closing price and its 20-period standard deviation in a logistic-style function to estimate the next price. A long entry is triggered when the estimate exceeds the current price by 0.5%; the position closes when it falls below the current price by 0.5%.

The document provides a short BTC/USDT futures backtest configuration covering November to December 2023, but reports no performance results. Its description claims the volume imbalance may indicate sentiment, while acknowledging lag, forecast error, and the absence of a stop loss. The published code also appears inconsistent with the stated accumulation method: it repeatedly assigns the current volume to positive or negative variables rather than accumulating separate volume totals. No evidence establishes predictive accuracy or profitability.

Key ideas

• The strategy combines a volume imbalance statistic with price volatility in a logistic-style price estimate.
• It enters long when the forecast clears a threshold above the current price and closes below a corresponding lower threshold.
• The published configuration specifies a BTC/USDT futures backtest, but gives no results.
• The code's volume calculations do not clearly implement the accumulation described in the prose.
• The proposal acknowledges forecast uncertainty and has no stop loss in its described rules.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.