What are the objects of stats from bt.run()?

[erdult]

What are the the different objects from stats from bt.run()?

I found _trades very useful, what is in _values? what are other useful objects to debug the trades and strategy actions

[initial-d]

stats is a pandas.Series. Most entries are ordinary scalar performance metrics, and a few underscore-prefixed entries contain richer objects from the run.

The easiest way to inspect what is available is:

stats = bt.run()
print(stats.index.tolist())
print(stats)

The most useful ones for debugging are usually:

trades = stats['_trades']          # DataFrame: one row per closed trade
equity = stats['_equity_curve']    # DataFrame: equity/drawdown over time
strategy = stats['_strategy']      # the strategy instance used for the run

_trades is usually the first place to look when debugging entries, exits, SL/TP behavior, duration, P/L, etc. _equity_curve is useful when you want to understand drawdown timing or compare the portfolio path against the trade list.

I would not treat stats._values as a Backtesting.py-specific result object. Since stats is a pandas Series, _values is mainly pandas' internal/raw array of the Series values. It is not as stable or self-describing as selecting entries by name, e.g. stats['Return [%]'], stats['_trades'], or stats['_equity_curve'].

For quick diagnostics, I often use something like:

stats = bt.run()

print(stats[['Return [%]', 'Sharpe Ratio', 'Max. Drawdown [%]', '# Trades']])
print(stats['_trades'].tail())
print(stats['_equity_curve'].tail())

If you need custom per-bar debugging, another practical pattern is to store your own values on self inside the strategy and then inspect stats['_strategy'] after the run.