Nikhil -
It is not just that it is a "hard test." A true CAT test like the actual GMAT and like the new Veritas exams is delivered and scored based on item response theory. Basically the test is adapting to you by trying to see what level of question you are comfortable with. Various questions help to distinguish those who are, for example, above the 60th percentile from those who are not above the 60th percentile. Once the computer is satisfied that you are likely above that level, it is not going to be giving you very many questions below the 60th percentile because that will not help to clarify exactly what your score is. Instead the computer is going to work with the idea that you are above the 60th (after your responses have established this) and it will be trying to see exactly where you are above the 60th. So you will be facing questions above that level for most of the test.
With your score of 730 this means that you faced questions above the 80th percentile most of the time on the quant and on the verbal. You could be expected to miss some questions at this level! So missing around 2 of every 5 questions seems about right. What this means is that you impressed the computer and the computer was testing to see how high your score might be.
You said:
So, is the percentile I received calculated on the basis of my score versus others who took this particular test? Or does Veritas use some other way to calculate the score?
It is not just a comparison to other people who have taken "this test" because every test is unique to the test taker. Each question or "item" has an item response curve that shows what level of test taker gets this right and what level does not. So each item is a certain level of difficulty based on a cumulative total of over 1.5 million responses. Your score is the result a very sophisticated algorithm that is based on your responses to various items and the difficulty of those items. This is what is done on the actual GMAT as well.
Here is the bottom line: Someone else could get a Q40 with the same percentage correct (59%) because they answered different levels of questions right and wrong than you did. Or someone could get the same Q49 that you earned, but get 75% of the questions right that they faced. Except that for most of the test they faced lower level questions than you did and so needed to get more of them right to earn their way up to the level that you were at. Does that make sense?
As for the OG - most questions are between the 25th and 75th percentile. There is not a higher percentage of questions in the OG above the 80th percentile, whereas most of the questions you faced on your exam where over the 80th percentile. That accounts for the difference.
Nice job!!
