So if you throw a bunch of different snippets at this algorithm, it gives wildly divergent results between one type of scene and the next, but within a given type, it was 100% consistent. I'm pretty sure this says more about the algorithm than about my writing...
Type of Scene / # of results / Author Match
Character introductions, scene-setting / x3 / Ernst Hemmingway
Dialogue-driven /x3 / Agatha Christi
Action, heightened tension / x2 / Arthur C. Clark
Internal character narrative / x2 / Chuck Palahniuk (Author of Fight Club, Choke)