Testing the GMATPrep scoring algorithm for Quant section

tutorphd wrote: You are comparing bananas to apples here. When the test is very hard like GMAT, getting a few questions wrong will lead to a smaller decrease in percentiles because very few people get those questions right.

no, you've got it backward.

the gmat population is much, much more competitive and well-prepared, because the only people taking the gmat are those who are actually motivated to attend a graduate b-school program. by contrast, the sat is taken by some huge fraction of the US high-school population -- basically, everybody who is entertaining any thoughts of attending a four-year college.
as a result, the SAT percentiles are much more forgiving.

SAT is way lower difficulty than GMAT so comparing them to prove a paper test is more prone to errors than adaptive test is not very sound.

the SAT was your example, not mine.

The problem I see in the adaptive test is the fact it is less adaptive in the middle and the end of the test, while a paper test is continuously scanning the full range in every sub-topic.

to scan "the full range in every subtopic", you would need hundreds and hundreds of problems.
instead, a paper test generally has one or two problems in each subtopic. as for which topics get to be the "hard" problems, well, that's just sort of random -- and that's the problem with paper tests that progress in difficulty.
for instance, let's say you're stronger in geometry, but you have the bad luck to take a paper test on which the hardest problems aren't geometry. then bad news bears.
on an adaptive test, by contrast, you'll get tough geometry questions right along with tough questions in the other topics.