stuff = -29.8744 + 0.13653*LSAT + 2.5727*GPA
probability = exp(stuff)/(1+exp(stuff))
LSAT percentiles (25,50, 75) = 162.8, 165, 168.2
GPA percentiles (25,50, 75) = 3.408, 3.640, 3.792
*Cornell tends to waitlist people rather than reject them, so the data available is heavily skewed towards acceptances
** traditional splitters will not benefit from this model as a high LSAT score will significantly alter the probability outcomes
*** the same could be said about reverse splitters, although it seems that low lsat high gpas tend to be admitted if an URM.