1. 和值振幅与标准差回归算法
STD-REG-V3利用历史和值的波动率，结合布林带变体计算和值的均值回归概率。当短期波动偏离标准差极值时，触发反向修正信号。
数学模型与切比雪夫不等式:
Z = (X - μ) / σ, P(|Z| > k) ≤ 1 / k²
std_regression.py
code
def calc_std_regression(sums_series, window=20):
rolling_mean = sums_series.rolling(window).mean()
rolling_std = sums_series.rolling(window).std()
z_score = (sums_series.iloc[-1] - rolling_mean.iloc[-1]) / rolling_std.iloc[-1]
prob_reversal = 1.0 - (1.0 / (z_score ** 2)) if abs(z_score) > 1 else 0.5
return {'z_score': round(z_score, 4), 'prob': round(max(prob_reversal, 0), 4)}