rahul16singh28 wrote:
broall wrote:
First, I have to confirm that the score algorithms in GMATPrep (old or new one) and the real test are the same. The only different is because of the input to the adaptive system to calculate score. For example:
- In GMATPrep, there are no experimental questions but there are some experimental questions.
- The question bank in GMATPrep is far less than the question bank in the real test. (So there are some significant results between those flatforms)
- The statistic information for each question is updated regularly in the real test, but not in GMATPrep (old or new one). ie a question could change the difficulty from hard to easy because of the improvement of test takers over time. However, this event will never happen in GMATPrep because GMAC no longer update statistic information of those questions in GMATPrep.
After all, there is no difference in the heart of the score algorithm. GMAC's claim is totally true.
However, adaptive system is really complex and we don't know what the exact result will be since the overall score depends on variety of reasons even that the score algorithm remains unchanged. Let's put it in this way: GMAT is the exam that wants to test the real abilities of test takers. Thus GMAT requires that those abilities must be consistent. Thus, the more consistent the performance during the test is, the higher score will be.
Ah, there is no way to know which questions are experimental. If anyone knew, I'm sure that GMAC would change the algorithm immediately.
Hi
broallI feel GMAC (Even though they deny it) has changed their scoring algorithm in the new format otherwise Q50 with 1 Wong (That too in last section) isn't possible. In GMAT Prep Mock, 7 wrong is fetching Q50. So, obviously GMAT Mock scorinng algorithm and Real GMAT scoring algorithm is not same.
I don't think that GMAC change its scoring algorithm. The score algorithm doesn't care about the number of question answered correctly. It cares about the difficulty of each question.
For example, theoretically, we could still earn V50 if we just answer 50% questions correctly. In the real life, this case will never happen because the question bank doesn't have enough very hard questions for those test takers. Thus, could could see that, in
ESR file of Ron Purewal, who has scored perfect 800, the difficulty comes down. The difficulty should go up right?
So, you can't simply conclude that the score algorithm has changed.
P/S: I think that GMAC does make some changes, but that's called system tuning (or calibration). This one shouldn't be called big changes in the core of system.
As I understand GMAC scoring algorithm (Please correct me if I am wrong) - as you answer the questions correctly the difficulty level goes up. I can't have 7 consecutive answers wrong in First 10 questions and then the GMAC throwing 8th question as 700 Level question and evaluating only the 8th Question correct answer for Scoring.