thats not the case, if, as is increasingly speculated, the lambda is not constant over time. you figure two parameters for linear and three for a quadratic experience