Learning goal: Euler–Maruyama adds drift×h and noise×√h. Brownian increment scaling preserves the intended process as the step changes.