ABSTRACT
The topic of Logit modeling itself and its usefulness in quantitative finance is important enough to merit a further direct discussion. Logit statistics and regression statistics are alternative ways to characterize and parameterize quantative relationships; and so it would, perhaps, be easy to lump them into the same broad category of "modeling"—but in truth, they differ fundamentally. A regression model is a formula for some figure of interest. The name Logit statistics derives from the logistic curve, which is a very convenient and tractable S curve. The number of such propositions and corresponding Logit models is truly unlimited, but a caution is in order. Logit statistics, by the way, are very old, having been used before regression was popularized by Francis Galton in the middle of the nineteenth century. As a result, a large lore of estimation techniques had developed before maximum likelihood methods were available.
