Once upon a time, predictions for the accuracy of inference on gravitational-wave signals relied on computationally inexpensive but often inaccurate techniques. Recently, the approach has shifted to actual inference on noisy signals with complex stoc
hastic Bayesian methods, at the expense of significant computational cost. Here, we argue that it is often possible to have the best of both worlds: a Bayesian approach that incorporates prior information and correctly marginalizes over uninteresting parameters, providing accurate posterior probability distribution functions, but carried out on a simple grid at a low computational cost, comparable to the inexpensive predictive techniques.
