Struct SimulationStats 

pub struct SimulationStats { /* private fields */ }

Running statistics over the runs of a strategy simulation.

Results are folded in one at a time through SimulationStats::update (or SimulationStats::update_outcome for a bare PathOutcome). Each result is projected onto its PathOutcome through the backtest adapter, From<&SimulationResult> for PathOutcome, so the P&L of a run is PnL::total_pnl, the same figure crate::backtesting::results::SimulationStatsResult reports.

The accumulator counts how each run closed (take profit, stop loss, expiry) and by which exit policy. Every P&L and holding-period figure (mean, median, standard deviation, best, worst, average holding period) comes from SimulationStats::statistics, that is from PathStatistics::from_outcomes over the stored outcomes; no second, running copy of them is kept.

Memory

Every outcome folded in is stored, and every result folded in through update is stored as well, so memory grows linearly with the number of runs. The outcomes cannot be rebuilt from the results on demand: an outcome folded in through update_outcome has no result behind it.

Implementations

impl SimulationStats

pub fn new() -> Self

Creates a new instance of SimulationStats with default values.

Returns

A new SimulationStats instance with all counters set to zero.

pub fn update(&mut self, result: SimulationResult) -> Result<(), BacktestError>

Updates statistics with results from a single simulation run.

The counters are driven by the backtest adapter’s From<&SimulationResult> for PathOutcome view of the result, whose P&L is PnL::total_pnl (realized plus unrealized); the result itself is then stored and read back through SimulationStats::results. A result whose P&L reports a non-zero unrealized leg therefore counts with that leg included, as in crate::backtesting::results::SimulationStatsResult.

Parameters

• result - The simulation result containing all metrics

Errors

Returns BacktestError::CounterOverflow when a run counter overflows. The accumulator is left untouched when that happens: every fallible step is resolved before the first field is written, so a rejected result cannot leave one counter advanced for a run the others never saw.

pub fn update_outcome( &mut self, outcome: &PathOutcome, ) -> Result<(), BacktestError>

Folds one generic path outcome into the counters.

This is the strategy-agnostic half of SimulationStats::update: it advances the outcome counters and the exit-reason distribution, and stores the outcome for SimulationStats::statistics.

Errors

Same contract as SimulationStats::update: the accumulator is left untouched when a counter overflows.

pub fn statistics(&self) -> Result<PathStatistics, BacktestError>

Aggregate statistics over every outcome folded in so far.

The figures are PathStatistics::from_outcomes over the outcomes SimulationStats::update and SimulationStats::update_outcome accepted, so they follow the same formulas, and for results the same P&L, as crate::backtesting::results::SimulationStatsResult. An outcome without a P&L counts as zero, in the mean and in the best and worst figures alike.

Errors

Returns BacktestError::Simulation carrying the error of PathStatistics::from_outcomes, for example when the variance of a P&L set far outside any traded size overflows Decimal.

pub fn total_simulations(&self) -> usize

Number of outcomes folded in.

pub fn profitable_closes(&self) -> usize

Number of runs that closed on their take-profit condition.

pub fn loss_closes(&self) -> usize

Number of runs that closed on their stop-loss condition.

pub fn expired_trades(&self) -> usize

Number of runs that reached expiration without another exit.

pub fn total_pnl(&self) -> Result<Decimal, BacktestError>

Sum of the P&L of every run, in the strategy’s currency.

Summed over the stored outcomes, so it is the same P&L the SimulationStats::statistics mean is taken over; a run without a P&L counts as zero.

Errors

Returns BacktestError::Decimal when the sum leaves the representable Decimal range.

pub fn exit_reasons(&self) -> &HashMap<ExitPolicy, usize>

How many runs each exit policy closed.

pub fn results(&self) -> &[SimulationResult]

The results folded in through SimulationStats::update, in arrival order. Outcomes folded in through SimulationStats::update_outcome alone are not stored.

Trait Implementations

impl Clone for SimulationStats

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for SimulationStats

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for SimulationStats

fn default() -> Self

Returns the “default value” for a type.