# pragma pylint: disable=missing-docstring, invalid-name, pointless-string-statement # flake8: noqa: F401 # isort: skip_file # --- Do not remove these imports --- import numpy as np import pandas as pd from datetime import datetime, timedelta, timezone from pandas import DataFrame from typing import Dict, Optional, Union, Tuple, List from freqtrade.optimize.space import Categorical, Dimension, Integer, SKDecimal import logging from functools import reduce logger = logging.getLogger(__name__) from freqtrade.strategy import ( IStrategy, Trade, Order, PairLocks, informative, # @informative decorator # Hyperopt Parameters BooleanParameter, CategoricalParameter, DecimalParameter, IntParameter, RealParameter, # timeframe helpers timeframe_to_minutes, timeframe_to_next_date, timeframe_to_prev_date, # Strategy helper functions merge_informative_pair, stoploss_from_absolute, stoploss_from_open, ) # -------------------------------- # Add your lib to import here import talib.abstract as ta import pandas_ta as pta from technical import qtpylib # ========================================== # ALMA + MACD Strategy: Why This Works Better (Real Backtest Results) # https://youtu.be/x34p8VNitkY # ========================================== # ================================ # Freqtrade Version # ================================ """ freqtrade -V Operating System: Linux-6.6.87.2-microsoft-standard-WSL2-x86_64-with-glibc2.36 Python Version: Python 3.13.8 CCXT Version: 4.5.20 Freqtrade Version: freqtrade 2025.11 """ # ================================ # Download Historical Data # ================================ """ freqtrade download-data \ -c user_data/binance_futures_ALMA_MACD.json \ --timerange 20230101- \ -t 1m 5m 15m 30m 1h 2h 4h 1d """ # ================================ # Hyperopt Optimization # ================================ """ freqtrade hyperopt \ --strategy ALMA_MACD \ --config user_data/binance_futures_ALMA_MACD.json \ --timeframe 1h \ --timerange 20241101-20250701 \ --hyperopt-loss MultiMetricHyperOptLoss \ --spaces buy \ -e 100 \ --job-workers 10 \ --random-state 9319 \ --min-trades 10 \ --max-open-trades 1 \ -p BCH/USDT:USDT """ # ================================ # Backtesting # ================================ """ freqtrade backtesting \ --strategy ALMA_MACD \ --timeframe 1h \ --timerange 20241101-20251101 \ --breakdown month \ -c user_data/binance_futures_ALMA_MACD.json \ --max-open-trades 1 \ --cache none \ --timeframe-detail 5m \ -p BCH/USDT:USDT """ # ================================ # Start FreqUI Web Interface # ================================ """ freqtrade webserver \ --config user_data/binance_futures_ALMA_MACD.json """ class ALMA_MACD(IStrategy): # Strategy interface version - allow new iterations of the strategy interface. # Check the documentation or the Sample strategy to get the latest version. INTERFACE_VERSION = 3 # Optimal timeframe for the strategy. timeframe = "1h" # Can this strategy go short? can_short: bool = False # Minimal ROI designed for the strategy. # This attribute will be overridden if the config file contains "minimal_roi". minimal_roi = {} # Optimal stoploss designed for the strategy. # This attribute will be overridden if the config file contains "stoploss". stoploss = -0.25 # Trailing stoploss trailing_stop = False # trailing_only_offset_is_reached = False # trailing_stop_positive = 0.01 # trailing_stop_positive_offset = 0.0 # Disabled / not configured # Run "populate_indicators()" only for new candle. process_only_new_candles = True # These values can be overridden in the config. use_exit_signal = True exit_profit_only = False ignore_roi_if_entry_signal = False # Number of candles the strategy requires before producing valid signals startup_candle_count: int = 250 alma_window = CategoricalParameter([100, 150, 200], default=200, space="buy", optimize=True) alma_sigma = CategoricalParameter([10, 15, 20], default=10, space="buy", optimize=True) alma_offset = CategoricalParameter([0.7, 0.75, 0.8, 0.85], default=0.8, space="buy", optimize=True) alma_cross_window = CategoricalParameter([5, 10, 15], default=10, space="buy") macd_cross_window = CategoricalParameter([5, 10, 15], default=10, space="buy") risk_ratio = CategoricalParameter([1.5, 2, 2.5, 3], default=2, space="buy") atr_mult = CategoricalParameter([1.5, 2, 2.5, 3], default=2, space="buy") @property def plot_config(self): plot_config = { "main_plot": { f"alma{self.alma_window.value}_{self.alma_sigma.value}_{self.alma_offset.value}": { "color": "#FF9800", "type": "line" } }, "subplots": { "MACD": { "macd": {"color": "#2962ff", "fill_to": "macdhist"}, "macdsignal": {"color": "#ff6d00"}, "macdhist": {"type": "bar", "plotly": {"opacity": 0.9}} } } } return plot_config def informative_pairs(self): """ Define additional, informative pair/interval combinations to be cached from the exchange. These pair/interval combinations are non-tradeable, unless they are part of the whitelist as well. For more information, please consult the documentation :return: List of tuples in the format (pair, interval) Sample: return [("ETH/USDT", "5m"), ("BTC/USDT", "15m"), ] """ return [] def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame: for window_val in self.alma_window.range: for sigma_val in self.alma_sigma.range: for offset_val in self.alma_offset.range: dataframe[f"alma{window_val}_{sigma_val}_{offset_val}"] = pta.alma(dataframe['close'], length=window_val, sigma=sigma_val, distribution_offset=offset_val, offset=0) macd = ta.MACD(dataframe) dataframe["macd"] = macd["macd"] dataframe["macdsignal"] = macd["macdsignal"] dataframe["macdhist"] = macd["macdhist"] dataframe["atr"] = ta.ATR(dataframe) return dataframe def populate_entry_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame: long_conditions = [] long_conditions.append( (dataframe["close"] > dataframe[f"alma{self.alma_window.value}_{self.alma_sigma.value}_{self.alma_offset.value}"]) & (dataframe["close"].rolling(window=self.alma_cross_window.value).apply( lambda x: any(qtpylib.crossed_above(x, dataframe[f"alma{self.alma_window.value}_{self.alma_sigma.value}_{self.alma_offset.value}"].iloc[x.index[0]:x.index[-1]+1])) )) & (dataframe["macd"] > dataframe["macdsignal"]) & (dataframe["macd"].rolling(window=self.macd_cross_window.value).apply( lambda x: any(qtpylib.crossed_above(x, dataframe["macdsignal"].iloc[x.index[0]:x.index[-1]+1])) )) & (dataframe["volume"] > 0) ) if long_conditions: dataframe.loc[ reduce(lambda x, y: x & y, long_conditions), "enter_long"] = 1 return dataframe def populate_exit_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame: dataframe.loc[:, "exit_long"] = 0 return dataframe def custom_exit(self, pair: str, trade: Trade, current_time: datetime, current_rate: float, current_profit: float, **kwargs): side = -1 if trade.is_short else 1 # Retrieve TP and SL from custom data take_profit = trade.get_custom_data('take_profit') stop_loss = trade.get_custom_data('stop_loss') # If TP or SL is not set, initialize them if take_profit is None or stop_loss is None: dataframe, _ = self.dp.get_analyzed_dataframe(pair, self.timeframe) # Get the date just before trade opened trade_date = timeframe_to_prev_date(self.timeframe, trade.open_date_utc) # Filter dataframe to candles before the trade opened signal_data = dataframe.loc[dataframe["date"] < trade_date] if signal_data.empty: logger.warning(f"[{pair}] No signal candle found. Skip setting TP/SL.") return None signal_candle = signal_data.iloc[-1] # Calculate TP and SL atr = signal_candle["atr"] close = signal_candle["close"] take_profit = close + side * self.atr_mult.value * atr * self.risk_ratio.value stop_loss = close - side * self.atr_mult.value * atr # Save to trade's custom data trade.set_custom_data('take_profit', take_profit) trade.set_custom_data('stop_loss', stop_loss) # logger.info(f"[{pair}] TP/SL set. TP: {take_profit:.2f}, SL: {stop_loss:.2f}") # Get the current close price dataframe, _ = self.dp.get_analyzed_dataframe(pair, self.timeframe) current_close = dataframe.iloc[-1]['close'] # Check exit conditions if (trade.is_short and current_close <= take_profit) or \ (not trade.is_short and current_close >= take_profit): # logger.info(f"[{pair}] Take Profit hit! Close: {current_close:.2f}, TP: {take_profit:.2f}") return "take_profit_achieved" if (trade.is_short and current_close >= stop_loss) or \ (not trade.is_short and current_close <= stop_loss): # logger.info(f"[{pair}] Stop Loss hit! Close: {current_close:.2f}, SL: {stop_loss:.2f}") return "stop_loss_achieved" return None