# source: https://raw.githubusercontent.com/tradingstrategy-ai/gmx-ccxt-freqtrade/809240211df5d5b41603d97e4296a9adf95fde8b/user_data/strategies/IchiV2_LS_Optimised.py # Github_tradingstrategy_ai_gmx_ccxt_freqtrade__IchiV2_LS_Optimised__20260312_125037 - Dual Long/Short Ichimoku Strategy (Hyperopt-Optimised) # Based on: IchiV2_LS_Backtest.py # Description: Uses hyperopt-optimised parameters from 100-epoch SharpeHyperOptLossDaily run # on 28 chainlink token pairs (all available data). from freqtrade.strategy.interface import IStrategy from pandas import DataFrame import talib.abstract as ta import pandas as pd pd.options.mode.chained_assignment = None import technical.indicators as ftt import freqtrade.vendor.qtpylib.indicators as qtpylib from datetime import datetime from freqtrade.persistence import Trade from freqtrade.strategy import informative, DecimalParameter, IntParameter from freqtrade.strategy.strategy_helper import stoploss_from_absolute from freqtrade.exchange import timeframe_to_prev_date import logging class Github_tradingstrategy_ai_gmx_ccxt_freqtrade__IchiV2_LS_Optimised__20260312_125037(IStrategy): """ Github_tradingstrategy_ai_gmx_ccxt_freqtrade__IchiV2_LS_Optimised__20260312_125037 - hyperopt-optimised version of IchiV2_LS_Backtest. Dual long/short Ichimoku strategy on 1h timeframe. Parameters optimised via SharpeHyperOptLossDaily on 28 chainlink pairs. Optimised results (100 epochs): - 2624 trades, 715W/0D/1909L (27.3% win rate) - Avg profit 0.77%, total profit 508.62% - Avg duration 1 day 19:24 Key optimisation findings: - Wider lookback window (54 vs 24) for short entries - Much tighter ATR stop (1.5x vs 3.5x) for shorts - Faster velocity detection (11 vs 20 candle window) Long entries (checked at 4h boundary): - 4h close crosses above senkou_a_4h - Cloud is green (senkou_a_4h > senkou_b_4h) - SAR is below close (sar_4h < close_4h) Long exits: - 1h close crosses below sar_4h (exit on SAR cross) - Cloud flip to bearish - Base stoploss at -0.129 as fallback Short entries (checked on any 1h candle): - Cloud is bearish (senkou_a < senkou_b) - Price retested senkou_b (high >= senkou_b) within 54-candle lookback - Current price breaks down below senkou_b Short exits: - RSI 4h oversold (< 27) - OR price velocity exhaustion (< -12%) - Cloud flip to bullish - ATR-based stop loss (1.5x ATR from entry, locked at entry) """ INTERFACE_VERSION = 3 can_short = True # Hyperopt-optimised parameters buy_params = { 'retest_lookback_window': 54, } sell_params = { 'atr_stop_multiplier': 1.5, 'atr_timeperiod': 20, 'price_velocity_window': 11, 'rsi_tp_threshold': 27, 'use_rsi_exit': 1, 'velocity_tp_threshold': -0.12, } # Short entry parameters retest_lookback_window = IntParameter(4, 96, default=54, space='buy', optimize=False) # Short exit parameters rsi_tp_threshold = IntParameter(15, 35, default=27, space='sell', optimize=False) use_rsi_exit = IntParameter(0, 1, default=1, space='sell', optimize=False) # ATR stop loss for shorts atr_stop_multiplier = DecimalParameter(1.5, 4.0, default=1.5, decimals=1, space='sell', optimize=False) atr_timeperiod = IntParameter(2, 20, default=20, space='sell', optimize=False) # Price velocity window velocity_tp_threshold = DecimalParameter(-0.13, -0.06, default=-0.12, decimals=2, space='sell', optimize=False) price_velocity_window = IntParameter(4, 20, default=11, space='sell', optimize=False) custom_info = {} # ROI table - minimal minimal_roi = { "0": 10.0 # Disabled } # Stop loss stoploss = -0.129 use_custom_stoploss = True # Timeframe timeframe = '1h' startup_candle_count = 121 process_only_new_candles = True # Trailing stop - DISABLED trailing_stop = False trailing_stop_positive = None trailing_stop_positive_offset = None trailing_only_offset_is_reached = False # Exit settings use_exit_signal = True exit_profit_only = False ignore_roi_if_entry_signal = True ICHI_PARAMS = dict( conversion_line_period=20, base_line_periods=60, laggin_span=120, displacement=1, ) plot_config = { 'main_plot': { 'sar_4h': {'color': 'purple', 'marker': '.', 'linestyle': 'none'}, 'senkou_a_4h': {'color': 'lightgreen', 'linestyle': '--'}, 'senkou_b_4h': {'color': 'lightcoral', 'linestyle': '--'}, }, 'subplots': { 'RSI 4h': { 'rsi_4h': {'color': 'purple'}, }, 'Price Velocity': { 'price_velocity': {'color': 'magenta'}, }, }, } def __init__(self, config): super().__init__(config) self.logger = logging.getLogger(__name__) def custom_stoploss(self, pair: str, trade: 'Trade', current_time: datetime, current_rate: float, current_profit: float, after_fill: bool, **kwargs) -> float | None: """ Dynamic stoploss comparing base stoploss vs strategy-specific stops. :param pair: Trading pair. :param trade: Current trade object. :param current_time: Current datetime. :param current_rate: Current market rate. :param current_profit: Current profit ratio. :param after_fill: Whether this is called after a fill. :return: Stoploss value or None. LONGS: Compare base stoploss vs SAR level, choose the TIGHTER (higher) stop. SHORTS: Compare base stoploss vs ATR stop, choose the TIGHTER (lower) stop. """ if not trade.is_short: # === LONG POSITIONS: Compare base stoploss vs SAR === dataframe, _ = self.dp.get_analyzed_dataframe(pair, self.timeframe) if dataframe.empty: return None try: last_candle = dataframe.iloc[-1] sar_level = last_candle.get('sar_4h') if pd.isna(sar_level): return None entry_rate = trade.open_rate base_stop_price = entry_rate * (1 + self.stoploss) final_stop_price = max(base_stop_price, sar_level) return stoploss_from_absolute( stop_rate=final_stop_price, current_rate=current_rate, is_short=False, leverage=trade.leverage ) except (KeyError, IndexError) as e: self.logger.warning(f"Error calculating SAR stoploss for {pair}: {e}") return None else: # === SHORT POSITIONS: Compare base stoploss vs ATR stop === entry_atr = trade.get_custom_data(key='entry_atr') if entry_atr is None: dataframe, _ = self.dp.get_analyzed_dataframe(pair, self.timeframe) if not dataframe.empty: try: trade_candle_time = timeframe_to_prev_date(self.timeframe, trade.open_date_utc) entry_candles = dataframe[dataframe['date'] <= trade_candle_time] if not entry_candles.empty: entry_atr = entry_candles.iloc[-1]['atr'] if not pd.isna(entry_atr): trade.set_custom_data(key='entry_atr', value=float(entry_atr)) except (KeyError, IndexError) as e: self.logger.warning(f"Failed to store entry ATR for {pair}: {e}") if entry_atr is None: return None entry_rate = trade.open_rate atr_distance = self.atr_stop_multiplier.value * float(entry_atr) atr_stop_price = entry_rate + atr_distance base_stop_price = entry_rate * (1 + abs(self.stoploss)) final_stop_price = min(atr_stop_price, base_stop_price) return stoploss_from_absolute( stop_rate=final_stop_price, current_rate=current_rate, is_short=True, leverage=trade.leverage ) def custom_exit(self, pair: str, trade: 'Trade', current_time: datetime, current_rate: float, current_profit: float, **kwargs) -> str | bool | None: """ Exit logic - separate for longs vs shorts. :param pair: Trading pair. :param trade: Current trade object. :param current_time: Current datetime. :param current_rate: Current market rate. :param current_profit: Current profit ratio. :return: Exit reason string, or None. LONGS: SAR exit (1h close crosses below sar_4h), cloud flip to bearish. SHORTS: RSI 4h oversold, velocity exhaustion, cloud flip to bullish. """ dataframe, _ = self.dp.get_analyzed_dataframe(pair, self.timeframe) if dataframe.empty: return None try: last_row = dataframe.iloc[-1] last_close = last_row['close'] last_sar = last_row.get('sar_4h') # LONG EXIT if not trade.is_short: # Cloud flip check cloud_flip = last_row['senkou_a_4h'] < last_row['senkou_b_4h'] if cloud_flip: return 'senkou_a_4h_exit_long' # SAR cross down sar_cross_down = not pd.isna(last_sar) and last_close <= last_sar if sar_cross_down: return 'sar_exit_long' # SHORT EXIT if trade.is_short: # Cloud flip check cloud_flip = last_row['senkou_a_4h'] > last_row['senkou_b_4h'] if cloud_flip: return 'senkou_a_4h_exit_short' # RSI oversold last_rsi_4h = last_row.get('rsi_4h') rsi_oversold = False if self.use_rsi_exit.value == 1: rsi_oversold = not pd.isna(last_rsi_4h) and last_rsi_4h < self.rsi_tp_threshold.value if rsi_oversold: return 'tp_rsi_4h_oversold' # Velocity exhaustion last_velocity = last_row.get('price_velocity') velocity_exhausted = not pd.isna(last_velocity) and last_velocity < self.velocity_tp_threshold.value if velocity_exhausted: return 'tp_velocity_achieved' except (KeyError, IndexError) as e: self.logger.warning(f"Error in custom_exit for {pair}: {e}") return None def confirm_trade_entry(self, pair: str, order_type: str, amount: float, rate: float, time_in_force: str, current_time: datetime, entry_tag: str | None, side: str, **kwargs) -> bool: """ Store ATR at entry time for shorts. :param pair: Trading pair. :param order_type: Order type. :param amount: Trade amount. :param rate: Entry rate. :param time_in_force: Time in force. :param current_time: Current datetime. :param entry_tag: Entry tag. :param side: Trade side ('long' or 'short'). :return: True to confirm entry. """ trade = kwargs.get('trade') if trade is not None and side == 'short': dataframe, _ = self.dp.get_analyzed_dataframe(pair, self.timeframe) if not dataframe.empty: try: entry_candle = dataframe.iloc[-1] entry_atr = entry_candle.get('atr') if not pd.isna(entry_atr): trade.set_custom_data(key='entry_atr', value=float(entry_atr)) except (KeyError, IndexError) as e: self.logger.warning(f"Failed to store entry ATR for {pair}: {e}") return True def _add_ichimoku(self, df: DataFrame) -> None: """ Add Ichimoku indicator columns to the dataframe. :param df: Input dataframe with OHLC data. """ ichi = ftt.ichimoku(df, **self.ICHI_PARAMS) df['senkou_a'] = ichi['senkou_span_a'] df['senkou_b'] = ichi['senkou_span_b'] df['chikou_span'] = ichi['chikou_span'] df['tenkan_sen'] = ichi['tenkan_sen'] df['kijun_sen'] = ichi['kijun_sen'] df['cloud_green'] = ichi['cloud_green'] df['cloud_red'] = ichi['cloud_red'] @informative('4h') def populate_indicators_4h(self, dataframe: DataFrame, metadata: dict) -> DataFrame: """Calculate 4h Ichimoku, SAR, and RSI.""" self._add_ichimoku(dataframe) dataframe['sar'] = ta.SAR(dataframe) dataframe['rsi'] = ta.RSI(dataframe['close'], timeperiod=14) return dataframe def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame: """Calculate 1h indicators.""" # Mark 4h boundaries for long entries dataframe['is_4h_boundary'] = dataframe['date'].dt.hour % 4 == 0 # Cloud status for short entries dataframe['cloud_bearish'] = dataframe['senkou_a_4h'] < dataframe['senkou_b_4h'] # Price velocity for short exits dataframe['price_velocity'] = ( (dataframe['close'] - dataframe['close'].shift(self.price_velocity_window.value)) / dataframe['close'].shift(self.price_velocity_window.value) ) # ATR for short stop loss dataframe['atr'] = ta.ATR(dataframe, timeperiod=14) # Store for custom functions self.custom_info[metadata['pair']] = dataframe[[ 'date', 'close', 'sar_4h', 'rsi_4h', 'price_velocity', 'atr', 'senkou_a_4h', 'senkou_b_4h' ]].copy().set_index('date') return dataframe def populate_entry_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame: """ Dual entry system: 1. LONG: Ichimoku cloud crossover at 4h boundaries 2. SHORT: Senkou B retest breakdown """ # Initialize dataframe['enter_long'] = 0 dataframe['enter_short'] = 0 dataframe['buy'] = 0 dataframe['enter_tag'] = '' # === LONG ENTRY === at_4h_boundary = dataframe['is_4h_boundary'] effective_close = dataframe['close_4h'] prev_close_4h = effective_close.shift(4) prev_senkou_a = dataframe['senkou_a_4h'].shift(4) prev_below_cloud = prev_close_4h < prev_senkou_a curr_above_cloud = effective_close > dataframe['senkou_a_4h'] cloud_green = dataframe['senkou_a_4h'] > dataframe['senkou_b_4h'] sar_below_close = dataframe['sar_4h'] < effective_close long_entry = at_4h_boundary & prev_below_cloud & curr_above_cloud & cloud_green & sar_below_close # === SHORT ENTRY === window = self.retest_lookback_window.value dataframe['retested_senkou_b'] = dataframe['high'] >= dataframe['senkou_b_4h'] dataframe['had_senkou_b_retest'] = dataframe['retested_senkou_b'].rolling( window=window, min_periods=1 ).max().astype(bool) dataframe['cross_down_senkou_b'] = ( (dataframe['close'].shift(1) >= dataframe['senkou_b_4h'].shift(1)) & (dataframe['close'] < dataframe['senkou_b_4h']) ) short_entry = ( dataframe['cloud_bearish'] & dataframe['had_senkou_b_retest'] & dataframe['cross_down_senkou_b'] ) # Apply signals dataframe.loc[long_entry, 'enter_long'] = 1 dataframe.loc[long_entry, 'enter_tag'] = 'long_entry' dataframe.loc[short_entry, 'enter_short'] = 1 dataframe.loc[short_entry, 'enter_tag'] = 'short_entry' dataframe.loc[long_entry | short_entry, 'buy'] = 1 return dataframe def populate_exit_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame: """Exit signals disabled - using custom_exit.""" dataframe['exit_long'] = 0 dataframe['exit_short'] = 0 dataframe['exit_tag'] = '' dataframe['sell'] = 0 return dataframe