(Un) Scientific analysis of gmatprep

Tarun, the scoring algorithm in PowerPrep is completely different from that of the GmatPrep. I got only one question wrong in Q but the PP gave me only 50.
Although PowerPrep is good for studying in an early phase, its questions no longer represent the current GMAT, which is then more similarly represented by the GmatPrep.



Posted from my mobile device

Today, I was revising my 2 GMATPrep tests and I found the following order of CR questions (though approx 80% correct):

A P E W S I W A BF S E = 11 Questions
A=2
P=1
E=2
W=2
S=2
BF=1
I=1

C S W BF BF A W S S A F(A) = 11 Questions
C=1
S=3
W=2
BF=2
A=3

Assumption = A
Paradox = P
Evaluate = E
Weaken = W
Strenthen = S
Inference = I
Bold Face = B
Conclusion = C
Fill In the Blank(Assumption based) = F(A)

So, I think more than majority of the questions is from A, W and S. If you do well in CR, then expect rest from BF, E parts.

And also to build on top of this, the algorithm of GMATprep is not the same as the real GMAT (though very similar in terms of evaluating your abilities but not as through).

Many believe this was done to prevent reverse engineering.



_________________

I do agree with this. I m sure real gmat must be penalizing students who do random guesses with delay < 10sec.

Yes, so pretty much approach the real exam the same. Limit guesses. Finish exam. Answer each question as given without trying to "figure out" the exam.

My experience with GMATPrep Today
780 Q50 V48
On Quant missed 10 (3,5,14,18,19,23,30,32,35,37)
On Verbal missed 1

The last 10 Quant were mainly lucky guesses, when I went back and looked at them again, they seemed much easier than the first few. In fact I am currently researching the correct answer for number 3. I still haven't figured it out.

This thread never ceases even after 2 years.. Unbelievable stuff from bsd_lover and others.. I believe bsd_lover might not be watching this forum anymore.. But still wanna thank him for the experiment

Oh, he is around
Last visited about 10 hrs ago: members/member-33822.html
However, he does not get involved in this thread anymore now that it's been a few years...
_________________

[*] guys,i 've got 36 out of 41 verbal,and only20 out of 37 quants,in powerprep since,am not familiar with inequalities and quadratics as its almost 20years after my schooling,can anyone help me in telling how to tackle this situation at the earliest ?

Guys one very good observation...

funda to get 51 on GMAT (atleast tested on GMATPREP) never to get two continuous questions wrong...even with 5-6 overall wrong one can get a score of 51. Same number of mistakes and at same time with only 1 continuous set wrong will lead to 50

No, I just did GMATPrep #2 (2nd try) and got a Q51 with numbers 28 and 29 wrong (missed 3 out of 37 total). This is the first time I've scored Q51; every other time was Q50.

So far I've found Q50 on my GMAT's every time I get 6-7 quant questions wrong, which is quite a few when you consider that 31/37 is 83.7% correct.

I've done much better on the Verbal parts, even scoring a V51 once (3rd try on GMATPrep #1). V51 resulted from scoring 100%; V48 from answering 1 verbal question (#10) wrong; V47 and V48 from answering 2 verbal questions wrong (both later on, in the 20's-40's), and V42 from answering 6 wrong.

Overall I'm not too worried about the verbal score. Although it is pretty picky, needing to get every single one right to score a 51, it's overall easier imo than the quant section, and the percentile skew is better (V46 is 99%ile). I wish they scored quant and verbal evenly; it seems that you can miss 3-4 per section and still score within the 99th percentile.

Remember the gmat scoring is certainly not linear. Your first experiment where you guess RC and CR and score high on SC, simply shows that your overall score is so bad that the verbal score will mirror that.

I'm not sure that answering correctly the first 10 or so questions is that important.

Wonderful analysis by Hussain.

gmat-prep-analysis-tool-102803.html

I think that statement is more valid for verbal than it is for quant. During my actual GMAT prep, i got 7 wrong in verbal of which i got 2 wrong in a row 3 times (6 out of 7). I also got all questions from 13-35 correct and still scored only a 41. Then again, that may have been because i got 2 consecutive questions wrong twice before # 13.

Verbal definitely has more severe penalties for incorrect answers. I raised this question with a GMAT teacher as to why and his theory was that 'it is a lot harder for the GMAT to determine your competency in verbal than it is for the GMAT to determine your competency in quant'.

Amazing topic! Have been reading very carefully.. I'll try to test it too asa Im ready.. Great job everyone so far

GOod finding. Let us know if u have muchmore finding.. May be that may help us..

i agree with the findings of bsd_lover, but i guess we should also see how the test varies if we get a string of questions wrong!

That will be a very interesting test i believe.

If you are running out of time, which is worse - doing each of the remaining questions to the best of your ability, try to get them all right and leave the rest? Or do quick guesstimates and try to at least attempt all the questions? Which has more severe penalty - an incomplete test or a string of wrong answers at the end?

Never leave questions. Each un-attempted question translates to (roughly) a 3 percentile drop in the score. So the penalty to leave questions unanswered is severe.

please go through this thread as well (if you haven't already) - gmat-prep-software-analysis-and-what-if-scenarios-146146.html

Cheers