Docs / quant / com.nablatensor.lattice

final class

LatticeGreeks

Greeks from a binomial tree. delta and gamma are read straight off the first two backward-induction slices — they cost nothing beyond the price. vega, rho and theta rebuild the tree with a bumped parameter (two or three extra trees), which for an O(n^2) lattice is still cheap.

This is the tree engine's answer to the adjoint sweep: there is no tape, so the sensitivities come from the lattice geometry and small rebuilds instead.

Methods

static LatticeGreeksBuilder of()
double price()
double delta()
double gamma()
double vega()
double rho()
double theta()
static LatticeGreeks vanilla(double spot, double rate, double dividend, double vol, double maturity, int steps, BinomialTree.MethodEnum method, OptionType type, double strike, ExerciseSchedule schedule)
spot
spot
rate
risk-free rate
dividend
continuous dividend yield
vol
volatility
maturity
years to expiry
steps
tree steps
method
lattice parameterisation
type
call or put
strike
strike
schedule
European / American / Bermudan exercise