A new method for computing the largest Lyapunov exponent from chaotic
time series is presented, whereby more robust convergence to the exact
value is obtained. It aims to serve as a bridge between the two widel
y known existing methods: the Wolf and the Jacobian methods.