stable/freqtrade/optimize/optimize_reports.py

940 lines
42 KiB
Python
Raw Normal View History

import logging
from copy import deepcopy
2020-06-26 18:08:45 +00:00
from datetime import datetime, timedelta, timezone
from pathlib import Path
from typing import Any, Dict, List, Union
from pandas import DataFrame, to_datetime
from tabulate import tabulate
2022-09-18 11:31:52 +00:00
from freqtrade.constants import (DATETIME_PRINT_FORMAT, LAST_BT_RESULT_FN, UNLIMITED_STAKE_AMOUNT,
Config)
2022-12-25 20:29:37 +00:00
from freqtrade.data.metrics import (calculate_cagr, calculate_calmar, calculate_csum,
calculate_expectancy, calculate_market_change,
calculate_max_drawdown, calculate_sharpe, calculate_sortino)
from freqtrade.misc import decimals_per_coin, file_dump_joblib, file_dump_json, round_coin_value
from freqtrade.optimize.backtest_caching import get_backtest_metadata_filename
2020-09-28 17:39:41 +00:00
logger = logging.getLogger(__name__)
def store_backtest_stats(
recordfilename: Path, stats: Dict[str, DataFrame], dtappendix: str) -> None:
2020-08-18 14:15:24 +00:00
"""
Stores backtest results
:param recordfilename: Path object, which can either be a filename or a directory.
Filenames will be appended with a timestamp right before the suffix
2021-06-25 13:45:49 +00:00
while for directories, <directory>/backtest-result-<datetime>.json will be used as filename
2020-08-18 14:15:24 +00:00
:param stats: Dataframe containing the backtesting statistics
:param dtappendix: Datetime to use for the filename
2020-08-18 14:15:24 +00:00
"""
if recordfilename.is_dir():
filename = (recordfilename / f'backtest-result-{dtappendix}.json')
else:
filename = Path.joinpath(
recordfilename.parent, f'{recordfilename.stem}-{dtappendix}'
2021-08-06 22:19:36 +00:00
).with_suffix(recordfilename.suffix)
# Store metadata separately.
file_dump_json(get_backtest_metadata_filename(filename), stats['metadata'])
del stats['metadata']
2020-06-26 05:46:59 +00:00
file_dump_json(filename, stats)
latest_filename = Path.joinpath(filename.parent, LAST_BT_RESULT_FN)
2020-06-26 05:46:59 +00:00
file_dump_json(latest_filename, {'latest_backtest': str(filename.name)})
def store_backtest_signal_candles(
recordfilename: Path, candles: Dict[str, Dict], dtappendix: str) -> Path:
"""
Stores backtest trade signal candles
:param recordfilename: Path object, which can either be a filename or a directory.
Filenames will be appended with a timestamp right before the suffix
2022-04-16 15:32:04 +00:00
while for directories, <directory>/backtest-result-<datetime>_signals.pkl will be used
as filename
:param stats: Dict containing the backtesting signal candles
:param dtappendix: Datetime to use for the filename
"""
if recordfilename.is_dir():
filename = (recordfilename / f'backtest-result-{dtappendix}_signals.pkl')
else:
filename = Path.joinpath(
recordfilename.parent, f'{recordfilename.stem}-{dtappendix}_signals.pkl'
2022-04-19 12:00:09 +00:00
)
file_dump_joblib(filename, candles)
return filename
2020-06-26 05:46:59 +00:00
2022-04-16 15:15:04 +00:00
def _get_line_floatfmt(stake_currency: str) -> List[str]:
2020-05-25 17:18:53 +00:00
"""
Generate floatformat (goes in line with _generate_result_line())
"""
return ['s', 'd', '.2f', '.2f', f'.{decimals_per_coin(stake_currency)}f',
'.2f', 'd', 's', 's']
2020-05-25 17:18:53 +00:00
def _get_line_header(first_column: str, stake_currency: str,
direction: str = 'Entries') -> List[str]:
"""
Generate header lines (goes in line with _generate_result_line())
"""
return [first_column, direction, 'Avg Profit %', 'Cum Profit %',
2021-10-13 22:34:30 +00:00
f'Tot Profit {stake_currency}', 'Tot Profit %', 'Avg Duration',
'Win Draw Loss Win%']
def generate_wins_draws_losses(wins, draws, losses):
if wins > 0 and losses == 0:
wl_ratio = '100'
elif wins == 0:
wl_ratio = '0'
else:
wl_ratio = f'{100.0 / (wins + draws + losses) * wins:.1f}' if losses > 0 else '100'
return f'{wins:>4} {draws:>4} {losses:>4} {wl_ratio:>4}'
def _generate_result_line(result: DataFrame, starting_balance: int, first_column: str) -> Dict:
2020-05-25 17:18:53 +00:00
"""
Generate one result dict, with "first_column" as key.
"""
profit_sum = result['profit_ratio'].sum()
# (end-capital - starting capital) / starting capital
profit_total = result['profit_abs'].sum() / starting_balance
2020-05-25 17:18:53 +00:00
return {
'key': first_column,
'trades': len(result),
'profit_mean': result['profit_ratio'].mean() if len(result) > 0 else 0.0,
'profit_mean_pct': result['profit_ratio'].mean() * 100.0 if len(result) > 0 else 0.0,
'profit_sum': profit_sum,
'profit_sum_pct': round(profit_sum * 100.0, 2),
'profit_total_abs': result['profit_abs'].sum(),
'profit_total': profit_total,
2021-10-13 22:18:16 +00:00
'profit_total_pct': round(profit_total * 100.0, 2),
2020-05-25 17:18:53 +00:00
'duration_avg': str(timedelta(
minutes=round(result['trade_duration'].mean()))
2020-05-25 17:18:53 +00:00
) if not result.empty else '0:00',
# 'duration_max': str(timedelta(
# minutes=round(result['trade_duration'].max()))
2020-05-25 17:18:53 +00:00
# ) if not result.empty else '0:00',
# 'duration_min': str(timedelta(
# minutes=round(result['trade_duration'].min()))
2020-05-25 17:18:53 +00:00
# ) if not result.empty else '0:00',
'wins': len(result[result['profit_abs'] > 0]),
'draws': len(result[result['profit_abs'] == 0]),
'losses': len(result[result['profit_abs'] < 0]),
2020-05-25 17:18:53 +00:00
}
def generate_pair_metrics(pairlist: List[str], stake_currency: str, starting_balance: int,
2020-05-25 17:50:09 +00:00
results: DataFrame, skip_nan: bool = False) -> List[Dict]:
"""
Generates and returns a list for the given backtest data and the results dataframe
:param pairlist: Pairlist used
2020-01-02 08:37:54 +00:00
:param stake_currency: stake-currency - used to correctly name headers
:param starting_balance: Starting balance
2020-01-02 08:37:54 +00:00
:param results: Dataframe containing the backtest results
:param skip_nan: Print "left open" open trades
:return: List of Dicts containing the metrics per pair
"""
tabular_data = []
2020-05-25 17:18:53 +00:00
for pair in pairlist:
result = results[results['pair'] == pair]
if skip_nan and result['profit_abs'].isnull().all():
continue
tabular_data.append(_generate_result_line(result, starting_balance, pair))
2021-04-03 22:19:38 +00:00
# Sort by total profit %:
tabular_data = sorted(tabular_data, key=lambda k: k['profit_total_abs'], reverse=True)
# Append Total
tabular_data.append(_generate_result_line(results, starting_balance, 'TOTAL'))
2020-05-25 17:18:53 +00:00
return tabular_data
def generate_tag_metrics(tag_type: str,
starting_balance: int,
results: DataFrame,
skip_nan: bool = False) -> List[Dict]:
"""
Generates and returns a list of metrics for the given tag trades and the results dataframe
:param starting_balance: Starting balance
:param results: Dataframe containing the backtest results
:param skip_nan: Print "left open" open trades
:return: List of Dicts containing the metrics per pair
"""
tabular_data = []
2021-10-21 14:25:38 +00:00
if tag_type in results.columns:
for tag, count in results[tag_type].value_counts().items():
2021-10-21 14:25:38 +00:00
result = results[results[tag_type] == tag]
if skip_nan and result['profit_abs'].isnull().all():
continue
tabular_data.append(_generate_result_line(result, starting_balance, tag))
2021-10-21 14:25:38 +00:00
# Sort by total profit %:
tabular_data = sorted(tabular_data, key=lambda k: k['profit_total_abs'], reverse=True)
2021-10-21 14:25:38 +00:00
# Append Total
tabular_data.append(_generate_result_line(results, starting_balance, 'TOTAL'))
return tabular_data
else:
2021-10-24 13:18:29 +00:00
return []
def generate_exit_reason_stats(max_open_trades: int | float, results: DataFrame) -> List[Dict]:
"""
Generate small table outlining Backtest results
:param max_open_trades: Max_open_trades parameter
:param results: Dataframe containing the backtest result for one strategy
:return: List of Dicts containing the metrics per Sell reason
"""
tabular_data = []
for reason, count in results['exit_reason'].value_counts().items():
2022-03-24 19:33:47 +00:00
result = results.loc[results['exit_reason'] == reason]
profit_mean = result['profit_ratio'].mean()
profit_sum = result['profit_ratio'].sum()
profit_total = profit_sum / max_open_trades
tabular_data.append(
{
2022-03-24 19:33:47 +00:00
'exit_reason': reason,
'trades': count,
'wins': len(result[result['profit_abs'] > 0]),
'draws': len(result[result['profit_abs'] == 0]),
'losses': len(result[result['profit_abs'] < 0]),
'profit_mean': profit_mean,
'profit_mean_pct': round(profit_mean * 100, 2),
'profit_sum': profit_sum,
'profit_sum_pct': round(profit_sum * 100, 2),
'profit_total_abs': result['profit_abs'].sum(),
'profit_total': profit_total,
'profit_total_pct': round(profit_total * 100, 2),
}
)
return tabular_data
def generate_strategy_comparison(bt_stats: Dict) -> List[Dict]:
"""
Generate summary per strategy
2022-01-06 08:15:30 +00:00
:param bt_stats: Dict of <Strategyname: DataFrame> containing results for all strategies
:return: List of Dicts containing the metrics per Strategy
"""
tabular_data = []
for strategy, result in bt_stats.items():
tabular_data.append(deepcopy(result['results_per_pair'][-1]))
# Update "key" to strategy (results_per_pair has it as "Total").
tabular_data[-1]['key'] = strategy
tabular_data[-1]['max_drawdown_account'] = result['max_drawdown_account']
tabular_data[-1]['max_drawdown_abs'] = round_coin_value(
result['max_drawdown_abs'], result['stake_currency'], False)
2020-05-25 17:18:53 +00:00
return tabular_data
def generate_edge_table(results: dict) -> str:
floatfmt = ('s', '.10g', '.2f', '.2f', '.2f', '.2f', 'd', 'd', 'd')
tabular_data = []
headers = ['Pair', 'Stoploss', 'Win Rate', 'Risk Reward Ratio',
'Required Risk Reward', 'Expectancy', 'Total Number of Trades',
'Average Duration (min)']
for result in results.items():
if result[1].nb_trades > 0:
tabular_data.append([
result[0],
result[1].stoploss,
result[1].winrate,
result[1].risk_reward_ratio,
result[1].required_risk_reward,
result[1].expectancy,
result[1].nb_trades,
round(result[1].avg_trade_duration)
])
# Ignore type as floatfmt does allow tuples but mypy does not know that
return tabulate(tabular_data, headers=headers,
2022-04-23 09:31:12 +00:00
floatfmt=floatfmt, tablefmt="orgtbl", stralign="right")
2021-10-21 04:58:40 +00:00
def _get_resample_from_period(period: str) -> str:
if period == 'day':
return '1d'
if period == 'week':
return '1w'
if period == 'month':
2021-10-21 05:09:17 +00:00
return '1M'
2021-10-21 04:58:40 +00:00
raise ValueError(f"Period {period} is not supported.")
def generate_periodic_breakdown_stats(trade_list: List, period: str) -> List[Dict[str, Any]]:
results = DataFrame.from_records(trade_list)
2021-10-21 05:09:17 +00:00
if len(results) == 0:
return []
results['close_date'] = to_datetime(results['close_date'], utc=True)
2021-10-21 05:42:19 +00:00
resample_period = _get_resample_from_period(period)
resampled = results.resample(resample_period, on='close_date')
2021-10-21 04:58:40 +00:00
stats = []
2021-10-21 05:42:19 +00:00
for name, day in resampled:
profit_abs = day['profit_abs'].sum().round(10)
wins = sum(day['profit_abs'] > 0)
draws = sum(day['profit_abs'] == 0)
loses = sum(day['profit_abs'] < 0)
2021-10-21 04:58:40 +00:00
stats.append(
{
'date': name.strftime('%d/%m/%Y'),
'profit_abs': profit_abs,
'wins': wins,
'draws': draws,
'loses': loses
}
)
2021-10-21 04:58:40 +00:00
return stats
def generate_trading_stats(results: DataFrame) -> Dict[str, Any]:
""" Generate overall trade statistics """
if len(results) == 0:
return {
'wins': 0,
'losses': 0,
'draws': 0,
'holding_avg': timedelta(),
'winner_holding_avg': timedelta(),
'loser_holding_avg': timedelta(),
}
winning_trades = results.loc[results['profit_ratio'] > 0]
draw_trades = results.loc[results['profit_ratio'] == 0]
losing_trades = results.loc[results['profit_ratio'] < 0]
holding_avg = (timedelta(minutes=round(results['trade_duration'].mean()))
if not results.empty else timedelta())
winner_holding_avg = (timedelta(minutes=round(winning_trades['trade_duration'].mean()))
if not winning_trades.empty else timedelta())
loser_holding_avg = (timedelta(minutes=round(losing_trades['trade_duration'].mean()))
if not losing_trades.empty else timedelta())
return {
'wins': len(winning_trades),
'losses': len(losing_trades),
'draws': len(draw_trades),
'holding_avg': holding_avg,
'holding_avg_s': holding_avg.total_seconds(),
'winner_holding_avg': winner_holding_avg,
'winner_holding_avg_s': winner_holding_avg.total_seconds(),
'loser_holding_avg': loser_holding_avg,
'loser_holding_avg_s': loser_holding_avg.total_seconds(),
}
2020-07-03 17:45:45 +00:00
def generate_daily_stats(results: DataFrame) -> Dict[str, Any]:
""" Generate daily statistics """
if len(results) == 0:
return {
'backtest_best_day': 0,
'backtest_worst_day': 0,
2021-03-05 18:21:09 +00:00
'backtest_best_day_abs': 0,
'backtest_worst_day_abs': 0,
'winning_days': 0,
'draw_days': 0,
'losing_days': 0,
'daily_profit_list': [],
}
2021-03-05 18:21:09 +00:00
daily_profit_rel = results.resample('1d', on='close_date')['profit_ratio'].sum()
daily_profit = results.resample('1d', on='close_date')['profit_abs'].sum().round(10)
worst_rel = min(daily_profit_rel)
best_rel = max(daily_profit_rel)
2020-07-03 17:45:45 +00:00
worst = min(daily_profit)
best = max(daily_profit)
winning_days = sum(daily_profit > 0)
draw_days = sum(daily_profit == 0)
losing_days = sum(daily_profit < 0)
daily_profit_list = [(str(idx.date()), val) for idx, val in daily_profit.items()]
2020-07-03 17:45:45 +00:00
return {
2021-03-05 18:21:09 +00:00
'backtest_best_day': best_rel,
'backtest_worst_day': worst_rel,
'backtest_best_day_abs': best,
'backtest_worst_day_abs': worst,
2020-07-03 17:45:45 +00:00
'winning_days': winning_days,
'draw_days': draw_days,
'losing_days': losing_days,
'daily_profit': daily_profit_list,
2020-07-03 17:45:45 +00:00
}
def generate_strategy_stats(pairlist: List[str],
strategy: str,
content: Dict[str, Any],
min_date: datetime, max_date: datetime,
market_change: float
) -> Dict[str, Any]:
"""
:param pairlist: List of pairs to backtest
:param strategy: Strategy name
:param content: Backtest result data in the format:
{'results: results, 'config: config}}.
:param min_date: Backtest start date
:param max_date: Backtest end date
:param market_change: float indicating the market change
2021-06-25 13:45:49 +00:00
:return: Dictionary containing results per strategy and a strategy summary.
"""
results: Dict[str, DataFrame] = content['results']
if not isinstance(results, DataFrame):
return {}
config = content['config']
max_open_trades = min(config['max_open_trades'], len(pairlist))
start_balance = config['dry_run_wallet']
stake_currency = config['stake_currency']
pair_results = generate_pair_metrics(pairlist, stake_currency=stake_currency,
starting_balance=start_balance,
results=results, skip_nan=False)
2021-10-12 22:22:53 +00:00
enter_tag_results = generate_tag_metrics("enter_tag", starting_balance=start_balance,
results=results, skip_nan=False)
2022-03-24 19:33:47 +00:00
exit_reason_stats = generate_exit_reason_stats(max_open_trades=max_open_trades,
results=results)
left_open_results = generate_pair_metrics(
pairlist, stake_currency=stake_currency, starting_balance=start_balance,
results=results.loc[results['exit_reason'] == 'force_exit'], skip_nan=True)
daily_stats = generate_daily_stats(results)
trade_stats = generate_trading_stats(results)
best_pair = max([pair for pair in pair_results if pair['key'] != 'TOTAL'],
key=lambda x: x['profit_sum']) if len(pair_results) > 1 else None
worst_pair = min([pair for pair in pair_results if pair['key'] != 'TOTAL'],
key=lambda x: x['profit_sum']) if len(pair_results) > 1 else None
2022-06-18 14:27:43 +00:00
winning_profit = results.loc[results['profit_abs'] > 0, 'profit_abs'].sum()
losing_profit = results.loc[results['profit_abs'] < 0, 'profit_abs'].sum()
profit_factor = winning_profit / abs(losing_profit) if losing_profit else 0.0
2021-10-20 17:39:37 +00:00
backtest_days = (max_date - min_date).days or 1
strat_stats = {
'trades': results.to_dict(orient='records'),
'locks': [lock.to_json() for lock in content['locks']],
'best_pair': best_pair,
'worst_pair': worst_pair,
'results_per_pair': pair_results,
'results_per_enter_tag': enter_tag_results,
2022-03-24 19:33:47 +00:00
'exit_reason_summary': exit_reason_stats,
'left_open_trades': left_open_results,
# 'days_breakdown_stats': days_breakdown_stats,
'total_trades': len(results),
2021-11-18 19:34:59 +00:00
'trade_count_long': len(results.loc[~results['is_short']]),
'trade_count_short': len(results.loc[results['is_short']]),
'total_volume': float(results['stake_amount'].sum()),
'avg_stake_amount': results['stake_amount'].mean() if len(results) > 0 else 0,
'profit_mean': results['profit_ratio'].mean() if len(results) > 0 else 0,
'profit_median': results['profit_ratio'].median() if len(results) > 0 else 0,
'profit_total': results['profit_abs'].sum() / start_balance,
'profit_total_long': results.loc[~results['is_short'], 'profit_abs'].sum() / start_balance,
'profit_total_short': results.loc[results['is_short'], 'profit_abs'].sum() / start_balance,
'profit_total_abs': results['profit_abs'].sum(),
'profit_total_long_abs': results.loc[~results['is_short'], 'profit_abs'].sum(),
'profit_total_short_abs': results.loc[results['is_short'], 'profit_abs'].sum(),
2022-04-25 08:51:11 +00:00
'cagr': calculate_cagr(backtest_days, start_balance, content['final_balance']),
2022-12-25 20:29:37 +00:00
'expectancy': calculate_expectancy(results),
2022-12-25 23:19:51 +00:00
'sortino': calculate_sortino(results, min_date, max_date, start_balance),
'sharpe': calculate_sharpe(results, min_date, max_date, start_balance),
2022-12-25 20:29:37 +00:00
'calmar': calculate_calmar(results, min_date, max_date, start_balance),
2022-06-18 14:27:43 +00:00
'profit_factor': profit_factor,
'backtest_start': min_date.strftime(DATETIME_PRINT_FORMAT),
'backtest_start_ts': int(min_date.timestamp() * 1000),
'backtest_end': max_date.strftime(DATETIME_PRINT_FORMAT),
'backtest_end_ts': int(max_date.timestamp() * 1000),
'backtest_days': backtest_days,
'backtest_run_start_ts': content['backtest_start_time'],
'backtest_run_end_ts': content['backtest_end_time'],
2021-10-20 17:39:37 +00:00
'trades_per_day': round(len(results) / backtest_days, 2),
'market_change': market_change,
'pairlist': pairlist,
'stake_amount': config['stake_amount'],
'stake_currency': config['stake_currency'],
'stake_currency_decimals': decimals_per_coin(config['stake_currency']),
'starting_balance': start_balance,
'dry_run_wallet': start_balance,
'final_balance': content['final_balance'],
2021-05-23 07:46:51 +00:00
'rejected_signals': content['rejected_signals'],
2022-02-07 17:49:30 +00:00
'timedout_entry_orders': content['timedout_entry_orders'],
'timedout_exit_orders': content['timedout_exit_orders'],
2022-05-16 22:41:31 +00:00
'canceled_trade_entries': content['canceled_trade_entries'],
'canceled_entry_orders': content['canceled_entry_orders'],
'replaced_entry_orders': content['replaced_entry_orders'],
'max_open_trades': max_open_trades,
'max_open_trades_setting': (config['max_open_trades']
if config['max_open_trades'] != float('inf') else -1),
'timeframe': config['timeframe'],
'timeframe_detail': config.get('timeframe_detail', ''),
'timerange': config.get('timerange', ''),
'enable_protections': config.get('enable_protections', False),
'strategy_name': strategy,
# Parameters relevant for backtesting
'stoploss': config['stoploss'],
'trailing_stop': config.get('trailing_stop', False),
'trailing_stop_positive': config.get('trailing_stop_positive'),
'trailing_stop_positive_offset': config.get('trailing_stop_positive_offset', 0.0),
'trailing_only_offset_is_reached': config.get('trailing_only_offset_is_reached', False),
'use_custom_stoploss': config.get('use_custom_stoploss', False),
'minimal_roi': config['minimal_roi'],
2022-04-05 18:07:58 +00:00
'use_exit_signal': config['use_exit_signal'],
'exit_profit_only': config['exit_profit_only'],
'exit_profit_offset': config['exit_profit_offset'],
'ignore_roi_if_entry_signal': config['ignore_roi_if_entry_signal'],
**daily_stats,
**trade_stats
}
try:
max_drawdown_legacy, _, _, _, _, _ = calculate_max_drawdown(
results, value_col='profit_ratio')
(drawdown_abs, drawdown_start, drawdown_end, high_val, low_val,
max_drawdown) = calculate_max_drawdown(
2022-01-07 09:09:17 +00:00
results, value_col='profit_abs', starting_balance=start_balance)
# max_relative_drawdown = Underwater
2022-04-10 15:53:47 +00:00
(_, _, _, _, _, max_relative_drawdown) = calculate_max_drawdown(
results, value_col='profit_abs', starting_balance=start_balance, relative=True)
strat_stats.update({
'max_drawdown': max_drawdown_legacy, # Deprecated - do not use
'max_drawdown_account': max_drawdown,
2022-04-10 15:53:47 +00:00
'max_relative_drawdown': max_relative_drawdown,
'max_drawdown_abs': drawdown_abs,
'drawdown_start': drawdown_start.strftime(DATETIME_PRINT_FORMAT),
'drawdown_start_ts': drawdown_start.timestamp() * 1000,
'drawdown_end': drawdown_end.strftime(DATETIME_PRINT_FORMAT),
'drawdown_end_ts': drawdown_end.timestamp() * 1000,
'max_drawdown_low': low_val,
'max_drawdown_high': high_val,
})
csum_min, csum_max = calculate_csum(results, start_balance)
strat_stats.update({
'csum_min': csum_min,
'csum_max': csum_max
})
except ValueError:
strat_stats.update({
'max_drawdown': 0.0,
'max_drawdown_account': 0.0,
2022-04-10 15:53:47 +00:00
'max_relative_drawdown': 0.0,
'max_drawdown_abs': 0.0,
'max_drawdown_low': 0.0,
'max_drawdown_high': 0.0,
'drawdown_start': datetime(1970, 1, 1, tzinfo=timezone.utc),
'drawdown_start_ts': 0,
'drawdown_end': datetime(1970, 1, 1, tzinfo=timezone.utc),
'drawdown_end_ts': 0,
'csum_min': 0,
'csum_max': 0
})
return strat_stats
def generate_backtest_stats(btdata: Dict[str, DataFrame],
all_results: Dict[str, Dict[str, Union[DataFrame, Dict]]],
min_date: datetime, max_date: datetime
2020-06-09 06:00:35 +00:00
) -> Dict[str, Any]:
2020-06-07 12:32:01 +00:00
"""
:param btdata: Backtest data
:param all_results: backtest result - dictionary in the form:
{ Strategy: {'results: results, 'config: config}}.
2020-06-09 06:00:35 +00:00
:param min_date: Backtest start date
:param max_date: Backtest end date
2021-06-25 13:45:49 +00:00
:return: Dictionary containing results per strategy and a strategy summary.
2020-06-07 12:32:01 +00:00
"""
result: Dict[str, Any] = {
'metadata': {},
'strategy': {},
'strategy_comparison': [],
}
market_change = calculate_market_change(btdata, 'close')
metadata = {}
pairlist = list(btdata.keys())
for strategy, content in all_results.items():
strat_stats = generate_strategy_stats(pairlist, strategy, content,
min_date, max_date, market_change=market_change)
metadata[strategy] = {
'run_id': content['run_id'],
'backtest_start_time': content['backtest_start_time'],
}
result['strategy'][strategy] = strat_stats
strategy_results = generate_strategy_comparison(bt_stats=result['strategy'])
result['metadata'] = metadata
result['strategy_comparison'] = strategy_results
return result
2020-06-07 09:29:14 +00:00
###
# Start output section
###
2020-06-07 09:35:02 +00:00
def text_table_bt_results(pair_results: List[Dict[str, Any]], stake_currency: str) -> str:
2020-06-07 09:29:14 +00:00
"""
Generates and returns a text table for the given backtest data and the results dataframe
:param pair_results: List of Dictionaries - one entry per pair + final TOTAL row
:param stake_currency: stake-currency - used to correctly name headers
:return: pretty printed table with tabulate as string
"""
headers = _get_line_header('Pair', stake_currency)
floatfmt = _get_line_floatfmt(stake_currency)
2020-06-07 09:29:14 +00:00
output = [[
t['key'], t['trades'], t['profit_mean_pct'], t['profit_sum_pct'], t['profit_total_abs'],
t['profit_total_pct'], t['duration_avg'],
generate_wins_draws_losses(t['wins'], t['draws'], t['losses'])
2020-06-07 09:29:14 +00:00
] for t in pair_results]
# Ignore type as floatfmt does allow tuples but mypy does not know that
return tabulate(output, headers=headers,
floatfmt=floatfmt, tablefmt="orgtbl", stralign="right")
2022-03-24 19:33:47 +00:00
def text_table_exit_reason(exit_reason_stats: List[Dict[str, Any]], stake_currency: str) -> str:
2020-06-07 09:29:14 +00:00
"""
Generate small table outlining Backtest results
:param sell_reason_stats: Exit reason metrics
2020-06-07 09:29:14 +00:00
:param stake_currency: Stakecurrency used
:return: pretty printed table with tabulate as string
"""
headers = [
'Exit Reason',
'Exits',
'Win Draws Loss Win%',
2020-06-07 09:29:14 +00:00
'Avg Profit %',
'Cum Profit %',
f'Tot Profit {stake_currency}',
'Tot Profit %',
]
output = [[
2022-03-24 19:53:22 +00:00
t.get('exit_reason', t.get('sell_reason')), t['trades'],
generate_wins_draws_losses(t['wins'], t['draws'], t['losses']),
t['profit_mean_pct'], t['profit_sum_pct'],
round_coin_value(t['profit_total_abs'], stake_currency, False),
t['profit_total_pct'],
2022-03-24 19:33:47 +00:00
] for t in exit_reason_stats]
2020-06-07 09:29:14 +00:00
return tabulate(output, headers=headers, tablefmt="orgtbl", stralign="right")
def text_table_tags(tag_type: str, tag_results: List[Dict[str, Any]], stake_currency: str) -> str:
"""
Generates and returns a text table for the given backtest data and the results dataframe
:param pair_results: List of Dictionaries - one entry per pair + final TOTAL row
:param stake_currency: stake-currency - used to correctly name headers
:return: pretty printed table with tabulate as string
"""
2022-08-01 04:43:59 +00:00
if (tag_type == "enter_tag"):
2021-10-13 22:34:30 +00:00
headers = _get_line_header("TAG", stake_currency)
else:
headers = _get_line_header("TAG", stake_currency, 'Exits')
floatfmt = _get_line_floatfmt(stake_currency)
output = [
[
t['key'] if t['key'] is not None and len(
t['key']) > 0 else "OTHER",
t['trades'],
t['profit_mean_pct'],
t['profit_sum_pct'],
t['profit_total_abs'],
t['profit_total_pct'],
t['duration_avg'],
generate_wins_draws_losses(
t['wins'],
t['draws'],
t['losses'])] for t in tag_results]
# Ignore type as floatfmt does allow tuples but mypy does not know that
return tabulate(output, headers=headers,
floatfmt=floatfmt, tablefmt="orgtbl", stralign="right")
2021-10-21 04:58:40 +00:00
def text_table_periodic_breakdown(days_breakdown_stats: List[Dict[str, Any]],
stake_currency: str, period: str) -> str:
"""
Generate small table with Backtest results by days
:param days_breakdown_stats: Days breakdown metrics
:param stake_currency: Stakecurrency used
:return: pretty printed table with tabulate as string
"""
headers = [
2021-10-21 04:58:40 +00:00
period.capitalize(),
f'Tot Profit {stake_currency}',
'Wins',
'Draws',
'Losses',
]
output = [[
2021-10-20 18:01:31 +00:00
d['date'], round_coin_value(d['profit_abs'], stake_currency, False),
d['wins'], d['draws'], d['loses'],
] for d in days_breakdown_stats]
return tabulate(output, headers=headers, tablefmt="orgtbl", stralign="right")
def text_table_strategy(strategy_results, stake_currency: str) -> str:
2020-06-07 09:29:14 +00:00
"""
Generate summary table per strategy
2021-06-25 17:13:31 +00:00
:param strategy_results: Dict of <Strategyname: DataFrame> containing results for all strategies
2020-06-07 09:29:14 +00:00
:param stake_currency: stake-currency - used to correctly name headers
:return: pretty printed table with tabulate as string
"""
floatfmt = _get_line_floatfmt(stake_currency)
2020-06-07 09:29:14 +00:00
headers = _get_line_header('Strategy', stake_currency)
# _get_line_header() is also used for per-pair summary. Per-pair drawdown is mostly useless
# therefore we slip this column in only for strategy summary here.
headers.append('Drawdown')
# Align drawdown string on the center two space separator.
if 'max_drawdown_account' in strategy_results[0]:
drawdown = [f'{t["max_drawdown_account"] * 100:.2f}' for t in strategy_results]
else:
# Support for prior backtest results
drawdown = [f'{t["max_drawdown_per"]:.2f}' for t in strategy_results]
dd_pad_abs = max([len(t['max_drawdown_abs']) for t in strategy_results])
dd_pad_per = max([len(dd) for dd in drawdown])
drawdown = [f'{t["max_drawdown_abs"]:>{dd_pad_abs}} {stake_currency} {dd:>{dd_pad_per}}%'
for t, dd in zip(strategy_results, drawdown)]
2020-06-07 09:29:14 +00:00
output = [[
t['key'], t['trades'], t['profit_mean_pct'], t['profit_sum_pct'], t['profit_total_abs'],
t['profit_total_pct'], t['duration_avg'],
generate_wins_draws_losses(t['wins'], t['draws'], t['losses']), drawdown]
for t, drawdown in zip(strategy_results, drawdown)]
2020-06-07 09:29:14 +00:00
# Ignore type as floatfmt does allow tuples but mypy does not know that
return tabulate(output, headers=headers,
floatfmt=floatfmt, tablefmt="orgtbl", stralign="right")
2020-06-26 07:22:50 +00:00
def text_table_add_metrics(strat_results: Dict) -> str:
if len(strat_results['trades']) > 0:
best_trade = max(strat_results['trades'], key=lambda x: x['profit_ratio'])
worst_trade = min(strat_results['trades'], key=lambda x: x['profit_ratio'])
short_metrics = [
('', ''), # Empty line to improve readability
('Long / Short',
f"{strat_results.get('trade_count_long', 'total_trades')} / "
f"{strat_results.get('trade_count_short', 0)}"),
('Total profit Long %', f"{strat_results['profit_total_long']:.2%}"),
('Total profit Short %', f"{strat_results['profit_total_short']:.2%}"),
('Absolute profit Long', round_coin_value(strat_results['profit_total_long_abs'],
strat_results['stake_currency'])),
('Absolute profit Short', round_coin_value(strat_results['profit_total_short_abs'],
strat_results['stake_currency'])),
] if strat_results.get('trade_count_short', 0) > 0 else []
drawdown_metrics = []
if 'max_relative_drawdown' in strat_results:
# Compatibility to show old hyperopt results
drawdown_metrics.append(
('Max % of account underwater', f"{strat_results['max_relative_drawdown']:.2%}")
)
drawdown_metrics.extend([
('Absolute Drawdown (Account)', f"{strat_results['max_drawdown_account']:.2%}")
if 'max_drawdown_account' in strat_results else (
'Drawdown', f"{strat_results['max_drawdown']:.2%}"),
('Absolute Drawdown', round_coin_value(strat_results['max_drawdown_abs'],
strat_results['stake_currency'])),
('Drawdown high', round_coin_value(strat_results['max_drawdown_high'],
strat_results['stake_currency'])),
('Drawdown low', round_coin_value(strat_results['max_drawdown_low'],
strat_results['stake_currency'])),
('Drawdown Start', strat_results['drawdown_start']),
('Drawdown End', strat_results['drawdown_end']),
])
entry_adjustment_metrics = [
('Canceled Trade Entries', strat_results.get('canceled_trade_entries', 'N/A')),
('Canceled Entry Orders', strat_results.get('canceled_entry_orders', 'N/A')),
('Replaced Entry Orders', strat_results.get('replaced_entry_orders', 'N/A')),
] if strat_results.get('canceled_entry_orders', 0) > 0 else []
# Newly added fields should be ignored if they are missing in strat_results. hyperopt-show
# command stores these results and newer version of freqtrade must be able to handle old
# results with missing new fields.
2020-06-08 04:38:29 +00:00
metrics = [
('Backtesting from', strat_results['backtest_start']),
('Backtesting to', strat_results['backtest_end']),
2020-11-24 05:57:11 +00:00
('Max open trades', strat_results['max_open_trades']),
('', ''), # Empty line to improve readability
('Total/Daily Avg Trades',
f"{strat_results['total_trades']} / {strat_results['trades_per_day']}"),
2021-02-17 19:19:03 +00:00
('Starting balance', round_coin_value(strat_results['starting_balance'],
strat_results['stake_currency'])),
2021-02-17 19:19:03 +00:00
('Final balance', round_coin_value(strat_results['final_balance'],
strat_results['stake_currency'])),
('Absolute profit ', round_coin_value(strat_results['profit_total_abs'],
strat_results['stake_currency'])),
2021-11-11 12:55:55 +00:00
('Total profit %', f"{strat_results['profit_total']:.2%}"),
2022-04-25 15:37:25 +00:00
('CAGR %', f"{strat_results['cagr']:.2%}" if 'cagr' in strat_results else 'N/A'),
2022-12-25 20:29:37 +00:00
('Sortino', f"{strat_results['sortino']:.2f}" if 'sortino' in strat_results else 'N/A'),
('Sharpe', f"{strat_results['sharpe']:.2f}" if 'sharpe' in strat_results else 'N/A'),
('Calmar', f"{strat_results['calmar']:.2f}" if 'calmar' in strat_results else 'N/A'),
2022-06-18 14:27:43 +00:00
('Profit factor', f'{strat_results["profit_factor"]:.2f}' if 'profit_factor'
in strat_results else 'N/A'),
2022-12-25 23:19:51 +00:00
('Expectancy', f"{strat_results['expectancy']:.2f}" if 'expectancy'
in strat_results else 'N/A'),
2020-06-26 07:22:50 +00:00
('Trades per day', strat_results['trades_per_day']),
('Avg. daily profit %',
2021-11-11 14:58:30 +00:00
f"{(strat_results['profit_total'] / strat_results['backtest_days']):.2%}"),
2021-02-17 19:19:03 +00:00
('Avg. stake amount', round_coin_value(strat_results['avg_stake_amount'],
strat_results['stake_currency'])),
2021-02-14 12:08:49 +00:00
('Total trade volume', round_coin_value(strat_results['total_volume'],
strat_results['stake_currency'])),
*short_metrics,
('', ''), # Empty line to improve readability
2020-11-28 16:52:29 +00:00
('Best Pair', f"{strat_results['best_pair']['key']} "
2021-11-11 14:58:30 +00:00
f"{strat_results['best_pair']['profit_sum']:.2%}"),
2020-11-28 16:52:29 +00:00
('Worst Pair', f"{strat_results['worst_pair']['key']} "
2021-11-11 14:58:30 +00:00
f"{strat_results['worst_pair']['profit_sum']:.2%}"),
2021-11-11 12:55:55 +00:00
('Best trade', f"{best_trade['pair']} {best_trade['profit_ratio']:.2%}"),
2020-11-28 16:52:29 +00:00
('Worst trade', f"{worst_trade['pair']} "
2021-11-11 12:55:55 +00:00
f"{worst_trade['profit_ratio']:.2%}"),
2020-11-28 16:45:56 +00:00
2021-03-05 18:21:09 +00:00
('Best day', round_coin_value(strat_results['backtest_best_day_abs'],
strat_results['stake_currency'])),
('Worst day', round_coin_value(strat_results['backtest_worst_day_abs'],
strat_results['stake_currency'])),
2020-07-03 17:45:45 +00:00
('Days win/draw/lose', f"{strat_results['winning_days']} / "
f"{strat_results['draw_days']} / {strat_results['losing_days']}"),
2020-07-03 17:58:02 +00:00
('Avg. Duration Winners', f"{strat_results['winner_holding_avg']}"),
('Avg. Duration Loser', f"{strat_results['loser_holding_avg']}"),
2022-03-22 05:45:36 +00:00
('Rejected Entry signals', strat_results.get('rejected_signals', 'N/A')),
2022-02-07 17:49:30 +00:00
('Entry/Exit Timeouts',
f"{strat_results.get('timedout_entry_orders', 'N/A')} / "
f"{strat_results.get('timedout_exit_orders', 'N/A')}"),
*entry_adjustment_metrics,
2020-06-09 06:00:35 +00:00
('', ''), # Empty line to improve readability
2021-02-17 19:19:03 +00:00
('Min balance', round_coin_value(strat_results['csum_min'],
2021-02-20 19:21:30 +00:00
strat_results['stake_currency'])),
2021-02-17 19:19:03 +00:00
('Max balance', round_coin_value(strat_results['csum_max'],
2021-02-20 19:21:30 +00:00
strat_results['stake_currency'])),
*drawdown_metrics,
2021-11-11 12:55:55 +00:00
('Market change', f"{strat_results['market_change']:.2%}"),
2020-06-08 04:38:29 +00:00
]
return tabulate(metrics, headers=["Metric", "Value"], tablefmt="orgtbl")
else:
start_balance = round_coin_value(strat_results['starting_balance'],
strat_results['stake_currency'])
stake_amount = round_coin_value(
strat_results['stake_amount'], strat_results['stake_currency']
2021-08-06 22:19:36 +00:00
) if strat_results['stake_amount'] != UNLIMITED_STAKE_AMOUNT else 'unlimited'
message = ("No trades made. "
f"Your starting balance was {start_balance}, "
f"and your stake was {stake_amount}."
)
return message
2020-06-08 04:37:30 +00:00
def show_backtest_result(strategy: str, results: Dict[str, Any], stake_currency: str,
2021-10-21 04:58:40 +00:00
backtest_breakdown=[]):
"""
Print results for one strategy
"""
# Print results
print(f"Result for strategy {strategy}")
table = text_table_bt_results(results['results_per_pair'], stake_currency=stake_currency)
if isinstance(table, str):
print(' BACKTESTING REPORT '.center(len(table.splitlines()[0]), '='))
print(table)
if (results.get('results_per_enter_tag') is not None
or results.get('results_per_buy_tag') is not None):
# results_per_buy_tag is deprecated and should be removed 2 versions after short golive.
2021-10-21 14:25:38 +00:00
table = text_table_tags(
"enter_tag",
results.get('results_per_enter_tag', results.get('results_per_buy_tag')),
2021-10-21 14:25:38 +00:00
stake_currency=stake_currency)
2021-10-21 14:25:38 +00:00
if isinstance(table, str) and len(table) > 0:
2022-03-22 05:45:36 +00:00
print(' ENTER TAG STATS '.center(len(table.splitlines()[0]), '='))
2021-10-21 14:25:38 +00:00
print(table)
exit_reasons = results.get('exit_reason_summary', results.get('sell_reason_summary'))
table = text_table_exit_reason(exit_reason_stats=exit_reasons,
stake_currency=stake_currency)
if isinstance(table, str) and len(table) > 0:
print(' EXIT REASON STATS '.center(len(table.splitlines()[0]), '='))
print(table)
table = text_table_bt_results(results['left_open_trades'], stake_currency=stake_currency)
if isinstance(table, str) and len(table) > 0:
print(' LEFT OPEN TRADES REPORT '.center(len(table.splitlines()[0]), '='))
print(table)
2021-10-21 04:58:40 +00:00
for period in backtest_breakdown:
days_breakdown_stats = generate_periodic_breakdown_stats(
trade_list=results['trades'], period=period)
table = text_table_periodic_breakdown(days_breakdown_stats=days_breakdown_stats,
stake_currency=stake_currency, period=period)
2020-06-26 04:47:04 +00:00
if isinstance(table, str) and len(table) > 0:
2021-10-21 04:58:40 +00:00
print(f' {period.upper()} BREAKDOWN '.center(len(table.splitlines()[0]), '='))
print(table)
table = text_table_add_metrics(results)
if isinstance(table, str) and len(table) > 0:
print(' SUMMARY METRICS '.center(len(table.splitlines()[0]), '='))
print(table)
if isinstance(table, str) and len(table) > 0:
print('=' * len(table.splitlines()[0]))
print()
2022-09-18 11:31:52 +00:00
def show_backtest_results(config: Config, backtest_stats: Dict):
stake_currency = config['stake_currency']
for strategy, results in backtest_stats['strategy'].items():
2021-10-21 04:58:40 +00:00
show_backtest_result(
strategy, results, stake_currency,
config.get('backtest_breakdown', []))
if len(backtest_stats['strategy']) > 1:
# Print Strategy summary table
table = text_table_strategy(backtest_stats['strategy_comparison'], stake_currency)
print(f"{results['backtest_start']} -> {results['backtest_end']} |"
f" Max open trades : {results['max_open_trades']}")
print(' STRATEGY SUMMARY '.center(len(table.splitlines()[0]), '='))
print(table)
print('=' * len(table.splitlines()[0]))
print('\nFor more details, please look at the detail tables above')
2022-09-18 11:31:52 +00:00
def show_sorted_pairlist(config: Config, backtest_stats: Dict):
if config.get('backtest_show_pair_list', False):
for strategy, results in backtest_stats['strategy'].items():
2021-10-30 14:53:48 +00:00
print(f"Pairs for Strategy {strategy}: \n[")
for result in results['results_per_pair']:
if result["key"] != 'TOTAL':
2021-11-11 14:58:30 +00:00
print(f'"{result["key"]}", // {result["profit_mean"]:.2%}')
print("]")