DOC: Explain backtest statistics formulas

[Stamina9]

Closes #1238.

Summary

• Document the formulas and assumptions behind the less self-explanatory values returned by Backtest.run().
• Explain why Buy & Hold Return [%] begins after indicator warm-up and can therefore change with an indicator lookback.
• Record annualization conventions for intraday through yearly data and the distinction between annualized return and CAGR.
• Clarify the definitions of Sharpe, Sortino, Calmar, alpha, beta, profit factor, expectancy, SQN, and Kelly Criterion.
• State when annualized statistics return NaN.

Testing

• python -m flake8 backtesting/backtesting.py
• python -m unittest backtesting.test._test.TestDocs.test_backtest_run_docstring_contains_stats_keys
• git diff --check

Scope

Documentation only; no runtime behavior changes.