backtest Directory Reference

Files
	array_source.hpp
	ArraySource: zero-copy structure-of-arrays input, the numpy path of the Python bindings (8.5).
	backtest_config.hpp
	BacktestConfig: everything a run needs, assembled from the TOML sections [engine] [[instruments]] [strategy] [strategy.params] [risk] [venues.
	backtest_runner.hpp
	BacktestRunner (8.4): wires Engine<S, SimClock, SimTransport, InlineFeed> to a SimDriver and collects a BacktestResult.
	csv_source.hpp
	CsvSource: hand-rolled parser for the row format documented in data_source.hpp (ts_ns,type,inst,side,price,qty,seq, optional header line, '#' comments).
	data_source.hpp
	DataSource (8.4): a timestamp-ordered stream of normalized market-data events consumed by the simulator.
	fee_model.hpp
	Fees are charged by the simulated venue, so the model lives in fastmm::sim; the backtest exposes it under its own namespace for configuration and reporting.
	journal_source.hpp
	JournalSource: market-data events of an .fmj journal (BookDelta / BookSnapshot / Trade / BookTicker records, in sequence order).
	live_analysis.hpp
	What a live session's fills were worth, from its journal alone (fastmm-data analyze): per instrument and overall, the fills, maker share and traded notional, the spread they captured against the venue mid and their markouts and adverse selection at each horizon (markout.hpp), next to the order traffic.
	markout.hpp
	Post-fill markouts: the metric that separates spread captured from adverse selection.
	metrics.hpp
	Metrics (8.4) computed from the result rows: equity bars (1 s by default), fills and orders.
	replay.hpp
	Deterministic replay (5.11): runs Engine<S, SimClock, ReplayTransport, JournalFeed> over a session journal recorded by fastmm-live, a sim or a backtest run (BacktestConfig::journal_out) and proves that the engine reproduces the recorded outbound stream bit for bit.
	result.hpp
	BacktestResult (8.4): structure-of-arrays rows (numpy-friendly: each column is one contiguous vector) plus the summary Metrics and the raw counters of every component.
	sweep.hpp
	Parameter sweep (8.4/8.5): cartesian grid over strategy parameters run on a std::jthread pool; every worker owns its own data cursor (source_factory) and backend, results come back in grid order (first parameter varies slowest).
	synthetic_source.hpp
	SyntheticSource: a MarketGenerator run against its own MatchingEngine, published through MdAggregator as a plain market-data stream (snapshot, 100 ms depth batches, trades, book tickers).