from freqtrade.strategy import IStrategy, DecimalParameter, IntParameter from freqtrade.persistence import Trade from pandas import DataFrame from datetime import datetime import talib.abstract as ta import numpy as np from freqtrade.strategy import stoploss_from_open from functools import reduce class SMIPullbackStrategy(IStrategy): """ 基于您提供的四个代码片段的策略： 1. 动态分段止损 2. 回调检测 3. SMI 趋势指标 4. Kernel SMI """ INTERFACE_VERSION = 3 can_short = False minimal_roi = { "0": 0.05, "30": 0.03, "60": 0.02, "120": 0.015 } stoploss = -0.10 use_custom_stoploss = True trailing_stop = False timeframe = '5m' startup_candle_count: int = 50 order_types = { 'entry': 'limit', 'exit': 'limit', 'stoploss': 'market', 'stoploss_on_exchange': False } # ==================== 片段1：动态止损参数 ==================== pHSL = DecimalParameter(-0.200, -0.040, default=-0.10, decimals=3, space='sell', optimize=True, load=True) pPF_1 = DecimalParameter(0.008, 0.020, default=0.016, decimals=3, space='sell', optimize=True, load=True) pSL_1 = DecimalParameter(0.008, 0.020, default=0.011, decimals=3, space='sell', optimize=True, load=True) pPF_2 = DecimalParameter(0.040, 0.100, default=0.070, decimals=3, space='sell', optimize=True, load=True) pSL_2 = DecimalParameter(0.020, 0.070, default=0.030, decimals=3, space='sell', optimize=True, load=True) # ==================== 片段2：回调检测参数 ==================== pullback_periods = IntParameter(20, 40, default=30, space='buy', optimize=True, load=True) pullback_method = 'pct_outlier' # 'stdev_outlier', 'pct_outlier', 'candle_body' # ==================== 片段3：SMI 趋势参数 ==================== smi_k_length = IntParameter(5, 15, default=9, space='buy', optimize=True, load=True) smi_d_length = IntParameter(2, 5, default=3, space='buy', optimize=True, load=True) smi_smoothing = IntParameter(5, 20, default=10, space='buy', optimize=True, load=True) # ==================== 片段4：Kernel SMI 参数 ==================== ksmi_K = IntParameter(5, 15, default=10, space='buy', optimize=True, load=True) ksmi_h = DecimalParameter(5.0, 15.0, default=8.0, decimals=1, space='buy', optimize=True, load=True) ksmi_rw = DecimalParameter(5.0, 15.0, default=8.0, decimals=1, space='buy', optimize=True, load=True) ksmi_x0 = IntParameter(3, 8, default=5, space='buy', optimize=True, load=True) ksmi_osint = IntParameter(30, 50, default=40, space='buy', optimize=True, load=True) ksmi_obint = IntParameter(30, 50, default=40, space='buy', optimize=True, load=True) # ==================== 片段2：回调检测 ==================== def detect_pullback(self, df: DataFrame, periods=30, method='pct_outlier'): """回调和异常值检测""" if method == 'stdev_outlier': outlier_threshold = 2.0 df['dif'] = df['close'] - df['close'].shift(1) df['dif_squared_sum'] = (df['dif']**2).rolling(window=periods + 1).sum() df['std'] = np.sqrt((df['dif_squared_sum'] - df['dif'].shift(0)**2) / (periods - 1)) df['z'] = df['dif'] / df['std'] df['pullback_flag'] = np.where(df['z'] >= outlier_threshold, 1, 0) df['pullback_flag'] = np.where(df['z'] <= -outlier_threshold, -1, df['pullback_flag']) elif method == 'pct_outlier': outlier_threshold = 2.0 df["pb_pct_change"] = df["close"].pct_change() mean = df["pb_pct_change"].rolling(window=periods).mean() std = df["pb_pct_change"].rolling(window=periods).std() df['pb_zscore'] = (df["pb_pct_change"] - mean) / std df['pullback_flag'] = np.where(df['pb_zscore'] >= outlier_threshold, 1, 0) df['pullback_flag'] = np.where(df['pb_zscore'] <= -outlier_threshold, -1, df['pullback_flag']) elif method == 'candle_body': pullback_pct = 1.0 df['change'] = df['close'] - df['open'] df['pullback'] = (df['change'] / df['open']) * 100 df['pullback_flag'] = np.where(df['pullback'] >= pullback_pct, 1, 0) df['pullback_flag'] = np.where(df['pullback'] <= -pullback_pct, -1, df['pullback_flag']) return df # ==================== 片段3：SMI 趋势指标 ==================== def smi_trend(self, df: DataFrame, k_length=9, d_length=3, smoothing_type='EMA', smoothing=10): """Stochastic Momentum Index (SMI) 趋势指标""" ll = df['low'].rolling(window=k_length).min() hh = df['high'].rolling(window=k_length).max() diff = hh - ll rdiff = df['close'] - (hh + ll) / 2 avgrel = rdiff.ewm(span=d_length).mean().ewm(span=d_length).mean() avgdiff = diff.ewm(span=d_length).mean().ewm(span=d_length).mean() smi = np.where(avgdiff != 0, (avgrel / (avgdiff / 2) * 100), 0) smi_ma = ta.EMA(smi, timeperiod=smoothing) conditions = [ (np.greater(smi, 0) & np.greater(smi, smi_ma)), # (2) 强烈看涨 (np.less(smi, 0) & np.greater(smi, smi_ma)), # (1) 可能看涨反转 (np.greater(smi, 0) & np.less(smi, smi_ma)), # (-1) 可能看跌反转 (np.less(smi, 0) & np.less(smi, smi_ma)) # (-2) 强烈看跌 ] smi_trend = np.select(conditions, [2, 1, -1, -2]) return smi, smi_ma, smi_trend # ==================== 片段4：Kernel SMI ==================== def calculate_smi_kernel(self, df: DataFrame, _K: int = 10, h: float = 8.0, rw: float = 8.0, x_0: int = 5, osint: int = 40, obint: int = 40, _col: str = 'close'): """Kernel 回归 SMI 计算""" def kernel_regression(_src): if len(_src) < x_0: return 0 weights = [(1 + (i**2 / (h**2 * 2 * rw)))**(-rw) for i in range(len(_src))] weighted_sum = sum([val*weight for val, weight in zip(_src[x_0:], weights)]) return weighted_sum / sum(weights) if sum(weights) > 0 else 0 df['highestHigh'] = df[_col].rolling(window=_K).max() df['lowestLow'] = df[_col].rolling(window=_K).min() df['highestLowestRange'] = df['highestHigh'] - df['lowestLow'] df['relativeRange'] = df[_col] - (df['highestHigh'] + df['lowestLow']) / 2 df['smi_k'] = 200 * ( df['relativeRange'].rolling(window=_K).apply(kernel_regression, raw=True) / df['highestLowestRange'].rolling(window=_K).apply(kernel_regression, raw=True).replace(0, 1) ) df['k_smi'] = df['smi_k'].rolling(window=_K).apply(kernel_regression, raw=True) df['k_smi_down'] = (df['smi_k'] < obint) & (df['smi_k'].shift(1) >= obint) df['k_smi_up'] = (df['smi_k'] > -osint) & (df['smi_k'].shift(1) <= -osint) return df['smi_k'], df['k_smi'], df['k_smi_down'], df['k_smi_up'] def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame: """使用四个片段的所有指标""" # 片段2：回调检测 dataframe = self.detect_pullback( dataframe, periods=self.pullback_periods.value, method=self.pullback_method ) # 片段3：SMI 趋势指标 dataframe['smi'], dataframe['smi_ma'], dataframe['smi_trend'] = self.smi_trend( dataframe, k_length=self.smi_k_length.value, d_length=self.smi_d_length.value, smoothing_type='EMA', smoothing=self.smi_smoothing.value ) # 片段4：Kernel SMI dataframe['smi_k'], dataframe['k_smi'], dataframe['k_smi_down'], dataframe['k_smi_up'] = self.calculate_smi_kernel( dataframe, _K=self.ksmi_K.value, h=self.ksmi_h.value, rw=self.ksmi_rw.value, x_0=self.ksmi_x0.value, osint=self.ksmi_osint.value, obint=self.ksmi_obint.value ) # 辅助指标 dataframe['rsi'] = ta.RSI(dataframe, timeperiod=14) dataframe['volume_mean'] = dataframe['volume'].rolling(window=20).mean() return dataframe def populate_entry_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame: """ 买入逻辑：结合四个片段 1. Kernel SMI 上穿超卖区（片段4） 2. SMI 趋势看涨（片段3） 3. 下跌回调后反弹（片段2） """ conditions = [] # 片段4：Kernel SMI 上穿超卖区 conditions.append(dataframe['k_smi_up'] == True) # 片段3：SMI 趋势至少是可能看涨 conditions.append(dataframe['smi_trend'] >= 1) # 片段2：检测到下跌回调（前1-3根蜡烛） conditions.append( (dataframe['pullback_flag'].shift(1) == -1) | (dataframe['pullback_flag'].shift(2) == -1) | (dataframe['pullback_flag'].shift(3) == -1) ) # 辅助条件：RSI 不超买 conditions.append(dataframe['rsi'] < 70) # 成交量确认 conditions.append(dataframe['volume'] > dataframe['volume_mean'] * 0.5) if conditions: dataframe.loc[ reduce(lambda x, y: x & y, conditions), 'enter_long'] = 1 return dataframe def populate_exit_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame: """ 卖出逻辑：结合四个片段 1. Kernel SMI 下穿超买区（片段4） 2. SMI 趋势看跌（片段3） 3. 上涨异常（片段2） """ conditions = [] # 片段4：Kernel SMI 下穿超买区 conditions.append(dataframe['k_smi_down'] == True) # 片段3：SMI 趋势看跌 conditions.append(dataframe['smi_trend'] <= -1) # 片段2：检测到上涨异常 conditions.append(dataframe['pullback_flag'] == 1) if conditions: dataframe.loc[ reduce(lambda x, y: x & y, conditions), 'exit_long'] = 1 return dataframe # ==================== 片段1：动态止损 ==================== def custom_stoploss(self, pair: str, trade: Trade, current_time: datetime, current_rate: float, current_profit: float, **kwargs) -> float: """ 片段1的原始动态止损逻辑 """ HSL = self.pHSL.value PF_1 = self.pPF_1.value SL_1 = self.pSL_1.value PF_2 = self.pPF_2.value SL_2 = self.pSL_2.value if (current_profit > PF_2): sl_profit = SL_2 + (current_profit - PF_2) elif (current_profit > PF_1): sl_profit = SL_1 + ((current_profit - PF_1) * (SL_2 - SL_1) / (PF_2 - PF_1)) else: sl_profit = HSL if (sl_profit >= current_profit): return -0.99 return stoploss_from_open(sl_profit, current_profit)