a couple of comments.
first, you are bound to experience some variation in your scores. the test actually has a fairly large "standard error" (approximately the same thing as a standard deviation), so that 70% of your scores will be within 30 points of your "true" score. that's a fair amount of variation as it is - that's a 60-point-wide score band already - but think about the corollary: the other 30% of your scores will fall
outside the score band. that means that, 30% of the time, you can expect a score that's either MORE than 30 points above your "true" score, or MORE than 30 points below your "true" score.
so, don't freak out about a couple scores that are lower than normal. when you have a large sample of practice tests, you will have outliers.
or:
some days you get all the lights red.
--
a couple of responses to the points you've made in your post:
* TIME MANAGEMENT IS YOUR PRIMARY ISSUE at this point.
when you take your next mgmat practice test, you should
set the quant section to a per-problem time limit, such as 2:30 or 2:35 per problem. (it's ok to go above 2:00, which is the
average time per problem, because you will solve a fair # of problems in
under two minutes.)
one major flaw i found in the mgmat tests was that the xm quickly goes onto the 700+ level questions right after the first couple of qs...
our algorithm mirrors the real gmat algorithm fairly closely, so this isn't a "major flaw".
there is, however, one important discrepancy between our practice tests and the official test: our tests (currently) don't contain experimental questions. this is a hugely important factor, because experimental questions are of random difficulty levels, and so will average around 500-600 (intermediate) level. this means that if you're a 700+ student, the experimentals will, on average, be substantially easier than the problems that "count". the resulting difference will save you some time on the real test - but nowhere near 10-12 minutes (the equivalent of 5-6 questions).
and since it takes longer to solve these tougher qs one is bound to fall short on time at the end by atleast 5 or 6 qs...
as stated above, there will be somewhat more of a time crunch on these tests than on the actual gmat, mostly as a result of the lack of experimental questions.
BUT
5-6 questions is a LOT. if you're leaving 5-6 questions blank at the end of the test, then the problem lies mostly in poor time management, not in the design of the exam. it is therefore imperative on you to
improve your time management.
here are some specific ways in which you can do that:
* use a hierarchy of strategies in problem solvingnamely:
PLAN A = try to work out the "textbook solution" to the problem.
if you can't figure out the textbook solution, or if you hit an impasse of some sort, then IMMEDIATELY BAIL and go to plan b:
PLAN B = see if you can use a number-plugging technique.
there are 3 kinds of these: plugging in a
selection of numbers (for data sufficiency problems), plugging in
your own numbers for undetermined quantities (the VIC method) (for problem solving problems), and
plugging in the answer choices themselves, if the answer choices are numbers (also for problem solving problems).
if that doesn't apply, then you should IMMEDIATELY BAIL and go to plan c:
PLAN C = process of elimination / estimation / intuition / educated guessing
if that doesn't work, then just guess randomly, and MOVE ON.
you should pretend that you have a serious attention deficit: if a technique isn't working,
don't continue hammering away at it. change up your technique, according to this hierarchy.
* make sure you have OPENERS for problemsat some point, you should run through all the quant problems in the o.g. in a couple of hours - about 20 seconds per problem. in this drill, don't actually
do the problems; just
say how you would start them.
that may sound weird, but you MUST get to the point where you have an "opener" (a way to START the problem) for each and every problem within the first 15 seconds or so.
* never, ever, ever practice problems untimednever.
ever.
you shouldn't even bother working out the solution to a problem once your stopwatch has passed 3 minutes or so: a solution that takes 4-5 minutes isn't any better than no solution at all. remember, this test is not about how many problems you can get right; it's about how many problems you can get right
within the allotted time. those are 2 very different things indeed.
i dont think thats the case with the actual gmat where the difficulty level progresses slower...
the actual test also features precipitous increases / decreases in difficulty at times, especially near the beginning (at which point the test is doing its best to calibrate your ability).
remember: our tests will require your time-management game to be incredibly tight - probably a bit more so than the actual test - but, once you've tuned your time management to the point where you can handle our tests, you should be able to NAIL the real test.