I think for the most part it has to be far off... I haven't had UW for a few months now, but I dont think the 50% percentile (equiv to a 218) is at a 56% correct... I may be over estimating the people out there, but considering that, this conversion can't be correct. I know I was at 76% complete, which was the 80th percentile... 80th percentile should be about 1SD from national average, or close to 240, yet the formula estimates me at 266... actual is only 12 points lower than that, but when the range of scores is only about 60-80 points, that puts this formula off pretty high and a pretty poor predictor. Most of you are either A) falling for the anecdotal evidence trap that gets people in trouble all the time or B) not realizing that if something is +/- 5-10 points in predicting, especially in the middle range of numbers, its not a really good predictor (realistically a 254 vs a 266 doesn't make that much of a difference, but a 235 vs a 223 for some programs is the difference between being considered for an interview and having your application dropped in the garbage before its even looked at)... i guess its better than not having anything, especially with the variability of the exams (I'm sure with the correct question set I might have been able to get that 266, or could have gotten a 242, or 212) so something that helps you gauge and ease your mind is fine, but looking at it after the fact is not all that helpful.
