stable/freqtrade/templates/sample_hyperopt_loss.py

50 lines
1.7 KiB
Python
Raw Normal View History

2019-07-17 18:51:44 +00:00
from datetime import datetime
2020-09-28 17:39:41 +00:00
from math import exp
2019-07-17 18:51:44 +00:00
from pandas import DataFrame
from freqtrade.optimize.hyperopt import IHyperOptLoss
2020-09-28 17:39:41 +00:00
2019-07-17 18:51:44 +00:00
# Define some constants:
# set TARGET_TRADES to suit your number concurrent trades so its realistic
# to the number of days
TARGET_TRADES = 600
# This is assumed to be expected avg profit * expected trade count.
# For example, for 0.35% avg per trade (or 0.0035 as ratio) and 1100 trades,
# self.expected_max_profit = 3.85
# Check that the reported Σ% values do not exceed this!
# Note, this is ratio. 3.85 stated above means 385Σ%.
EXPECTED_MAX_PROFIT = 3.0
# max average trade duration in minutes
# if eval ends with higher value, we consider it a failed eval
MAX_ACCEPTED_TRADE_DURATION = 300
class SampleHyperOptLoss(IHyperOptLoss):
"""
Defines the default loss function for hyperopt
2019-07-18 04:31:44 +00:00
This is intended to give you some inspiration for your own loss function.
2019-07-17 18:51:44 +00:00
2020-01-21 16:14:19 +00:00
The Function needs to return a number (float) - which becomes smaller for better backtest
results.
2019-07-17 18:51:44 +00:00
"""
@staticmethod
def hyperopt_loss_function(results: DataFrame, trade_count: int,
min_date: datetime, max_date: datetime,
*args, **kwargs) -> float:
"""
Objective function, returns smaller number for better results
"""
total_profit = results['profit_ratio'].sum()
trade_duration = results['trade_duration'].mean()
2019-07-17 18:51:44 +00:00
trade_loss = 1 - 0.25 * exp(-(trade_count - TARGET_TRADES) ** 2 / 10 ** 5.8)
profit_loss = max(0, 1 - total_profit / EXPECTED_MAX_PROFIT)
duration_loss = 0.4 * min(trade_duration / MAX_ACCEPTED_TRADE_DURATION, 1)
result = trade_loss + profit_loss + duration_loss
return result