A Normalized EMA Difference for MACD-Based Crypto Trading
Code Freqtrade
Summary
This Freqtrade strategy computes a normalized MACD-like measure as the ratio of the 12-period exponential moving average to the 26-period exponential moving average, minus one. On a five-minute timeframe, it enters long when this value falls within configurable buy bounds and exits when it falls within separate configurable sell bounds. The parameter ranges are exposed for optimization, and the strategy does not enable short selling.
The document lists a set of reported backtest statistics, including trade counts, win and loss counts, average and median profit, total profit, and average duration. However, it gives no dataset, test period, fee assumptions, or out-of-sample validation, so the figures cannot establish robustness. Its return targets and stop loss are part of the configuration, and the code does not explain why the thresholds should generalize across assets or market conditions.
Key ideas
- The indicator is the 12-period EMA divided by the 26-period EMA, minus one.
- Long entries and exits use separate configurable ranges for that indicator.
- The example uses five-minute candles and does not open short positions.
- The reported backtest figures lack details about the test period and validation method.
Tags
Full text
# UniversalMACD.py
```py
# pragma pylint: disable=missing-docstring, invalid-name, pointless-string-statement
# flake8: noqa: F401
# isort: skip_file
# --- Do not remove these libs ---
import numpy as np
import pandas as pd
from pandas import DataFrame
from datetime import datetime
from typing import Optional, Union
from freqtrade.strategy import (BooleanParameter, CategoricalParameter, DecimalParameter,
IntParameter, IStrategy, merge_informative_pair)
# --------------------------------
# Add your lib to import here
import talib.abstract as ta
import pandas_ta as pta
from technical import qtpylib
class UniversalMACD(IStrategy):
# By: Masoud Azizi (@mablue)
# Tradingview Page: https://www.tradingview.com/script/xNEWcB8s-Universal-Moving-Average-Convergence-Divergence/
# 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 = '5m'
# Can this strategy go short?
can_short: bool = False
# $ freqtrade hyperopt -s UniversalMACD --hyperopt-loss SharpeHyperOptLossDaily
# "max_open_trades": 1,
# "stake_currency": "USDT",
# "stake_amount": 990,
# "dry_run_wallet": 1000,
# "trading_mode": "spot",
# "XMR/USDT","ATOM/USDT","FTM/USDT","CHR/USDT","BNB/USDT","ALGO/USDT","XEM/USDT","XTZ/USDT","ZEC/USDT","ADA/USDT",
# "CHZ/USDT","BTT/USDT","LUNA/USDT","VRA/USDT","KSM/USDT","DASH/USDT","COMP/USDT","CRO/USDT","WAVES/USDT","MKR/USDT",
# "DIA/USDT","LINK/USDT","DOT/USDT","YFI/USDT","UNI/USDT","FIL/USDT","AAVE/USDT","KCS/USDT","LTC/USDT","BSV/USDT",
# "XLM/USDT","ETC/USDT","ETH/USDT","BTC/USDT","XRP/USDT","TRX/USDT","VET/USDT","NEO/USDT","EOS/USDT","BCH/USDT",
# "CRV/USDT","SUSHI/USDT","KLV/USDT","DOGE/USDT","CAKE/USDT","AVAX/USDT","MANA/USDT","SAND/USDT","SHIB/USDT",
# "KDA/USDT","ICP/USDT","MATIC/USDT","ELON/USDT","NFT/USDT","ARRR/USDT","NEAR/USDT","CLV/USDT","SOL/USDT","SLP/USDT",
# "XPR/USDT","DYDX/USDT","FTT/USDT","KAVA/USDT","XEC/USDT"
# "method": "StaticPairList"
# *16 / 100: 40 trades.
# 31 / 9 / 0 Wins / Draws / Losses.
# Avg profit 2.34 %.
# Median profit 3.00 %.
# Total profit 928.95036811 USDT(92.90 %).
# Avg duration 3: 13:00 min.\
# Objective: -11.63412
# ROI table:
minimal_roi = {
"0": 0.213,
"27": 0.099,
"60": 0.03,
"164": 0
}
# Stoploss:
stoploss = -0.318
# Trailing stop:
trailing_stop = False # value loaded from strategy
trailing_stop_positive = None # value loaded from strategy
trailing_stop_positive_offset = 0.0 # value loaded from strategy
trailing_only_offset_is_reached = False # value loaded from strategy
# Number of candles the strategy requires before producing valid signals
startup_candle_count: int = 30
# Strategy parameters
buy_umacd_max = DecimalParameter(-0.05, 0.05, decimals=5, default=-0.01176, space="buy")
buy_umacd_min = DecimalParameter(-0.05, 0.05, decimals=5, default=-0.01416, space="buy")
sell_umacd_max = DecimalParameter(-0.05, 0.05, decimals=5, default=-0.02323, space="sell")
sell_umacd_min = DecimalParameter(-0.05, 0.05, decimals=5, default=-0.00707, space="sell")
def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
dataframe['ma12'] = ta.EMA(dataframe, timeperiod=12)
dataframe['ma26'] = ta.EMA(dataframe, timeperiod=26)
dataframe['umacd'] = (dataframe['ma12'] / dataframe['ma26']) - 1
# Just for show user the min and max of indicator in different coins to set inside hyperoptable variables.cuz
# in different timeframes should change the min and max in hyperoptable variables.
# print(dataframe['umacd'].min(), dataframe['umacd'].max())
return dataframe
def populate_entry_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
dataframe.loc[
(
(dataframe['umacd'].between(self.buy_umacd_min.value, self.buy_umacd_max.value))
),
'enter_long'] = 1
return dataframe
def populate_exit_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
dataframe.loc[
(
(dataframe['umacd'].between(self.sell_umacd_min.value, self.sell_umacd_max.value))
),
'exit_long'] = 1
return dataframe
```Shown in full with attribution under the source's licence. Licence: GPL-3.0
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.