srfm::backtest::PerformanceCalculator Class Reference

#include <backtest.hpp>

Static Public Member Functions
static std::optional< double >	sharpe (std::span< const double > returns, double risk_free_rate=constants::DEFAULT_RISK_FREE_RATE, double annualisation=constants::ANNUALISATION_FACTOR) noexcept
static std::optional< double >	sortino (std::span< const double > returns, double risk_free_rate=constants::DEFAULT_RISK_FREE_RATE, double annualisation=constants::ANNUALISATION_FACTOR) noexcept
static std::optional< double >	max_drawdown (std::span< const double > returns) noexcept
static std::optional< double >	gamma_weighted_ir (std::span< const double > strategy_returns, std::span< const double > benchmark_returns, std::span< const double > gamma_factors) noexcept

Detailed Description

Stateless utility for computing financial performance metrics.

All methods are static and operate on std::span<const double> for zero-copy access to any contiguous container.

Definition at line 127 of file backtest.hpp.

Member Function Documentation

gamma_weighted_ir()

std::optional< double > srfm::backtest::PerformanceCalculator::gamma_weighted_ir ( std::span< const double >  strategy_returns, std::span< const double >  benchmark_returns, std::span< const double >  gamma_factors  )

staticnoexcept

Compute γ-weighted information ratio.

Formula

IR_γ = (mean(active_ret) × mean(γ)) / σ(active_ret) where active_ret_t = strategy_ret_t − benchmark_ret_t

The γ factor up-weights the mean active return when signals were generated in high-velocity (high-γ) market regimes.

Returns

nullopt if inputs are mismatched in length, too short, or numerically degenerate.

Definition at line 148 of file performance_metrics.cpp.

max_drawdown()

std::optional< double > srfm::backtest::PerformanceCalculator::max_drawdown ( std::span< const double >  returns )

staticnoexcept

Compute maximum drawdown of an equity curve.

Formula

MDD = max over t of { (peak_t − trough_t) / peak_t } where peak_t = max_{s ≤ t} equity_curve[s]

The equity curve is constructed by cumulative-summing the return series.

Returns

Maximum drawdown in [0, 1]. Returns nullopt on empty input.

Definition at line 124 of file performance_metrics.cpp.

sharpe()

std::optional< double > srfm::backtest::PerformanceCalculator::sharpe ( std::span< const double >  returns, double  risk_free_rate = constants::DEFAULT_RISK_FREE_RATE, double  annualisation = constants::ANNUALISATION_FACTOR  )

staticnoexcept

Compute annualised Sharpe ratio.

Formula

Sharpe = (mean(R) − r_f) / σ(R) × √ann

Returns

nullopt if series has fewer than 2 elements, σ = 0, or any NaN/Inf.

Definition at line 81 of file performance_metrics.cpp.

sortino()

std::optional< double > srfm::backtest::PerformanceCalculator::sortino ( std::span< const double >  returns, double  risk_free_rate = constants::DEFAULT_RISK_FREE_RATE, double  annualisation = constants::ANNUALISATION_FACTOR  )

staticnoexcept

Compute annualised Sortino ratio (downside-deviation denominator).

Formula

Sortino = (mean(R) − r_f) / σ_down(R) × √ann

where σ_down is the standard deviation of returns below r_f.

Returns

nullopt if series is too short, downside-vol is zero, or any NaN/Inf.

Definition at line 103 of file performance_metrics.cpp.

---

The documentation for this class was generated from the following files:

• include/srfm/backtest.hpp
• src/backtest/performance_metrics.cpp