Skip to main content

limen.backtest

Evaluate aligned predictions as trading outcomes through either a vectorized snapshot or a sequential simulation.

Canonical docs​

What this package owns​

Owns Limen's two backtest modes and the assumptions they encode. Does not own signal generation, experiment logging, or raw portfolio bookkeeping beyond what it delegates to limen.trading.

Key entry points​

Entry pointUse it whenNotes
backtest_snapshotYou want the fast, vectorized evaluation used across many roundsImport from limen.backtest.backtest_snapshot
BacktestSequentialYou want bar-by-bar account-state simulationImport from limen.backtest.backtest_sequential

Adjacent modules​

  • limen.log uses backtest_snapshot() to summarize experiment permutations.
  • limen.trading provides the Account ledger used by the sequential path.
  • limen.experiment and limen.sfd sit upstream by producing the predictions that backtests consume.

Quick orientation​

backtest/
├── backtest_snapshot.py     # Vectorized snapshot evaluator
└── backtest_sequential.py   # Stateful account-based simulation

Things to know​

  • The package root currently does not re-export the backtest helpers, so import from the module paths directly.
  • backtest_snapshot() is the common analysis path for experiment sweeps because it is simple and fast.
  • BacktestSequential is the right tool when you need a ledger and bar-by-bar account transitions.
  • The underlying Account supports long and short bookkeeping, but the current sequential backtest loop only executes the long-side buy and sell path.
  • Snapshot outputs are reported in percent units, while fee and slippage inputs are specified in basis points.