In my experience you can have wildly different results on unofficial practice tests. I had a tutoring student once score a 590 an unofficial practice test and then a few days later over 700 on the actual GMAT.

Unofficial tests - this means anything that is not a product of the company that makes the GMAT - are not to be used to determine how you will do on test day. The only reliable predictors of test day performance are the official practice tests that are found at MBA.com.

At Veritas we offer 15 practice tests to our students, but we still let students know that the official practice tests are the best predictor of scores. The other tests can help you work on many things, timing, stamina, using the computer instead of paper, and so forth, but are not great predictors of score.

If you have taken both of the official practice tests, take each of them a second time. You may get a few questions that repeat, and if you do make sure that you allow the clock to run the appropriate amount of time before you answer - in other words don't answer a repeated problem solving question in 20 seconds - more like 2 minutes and 20 seconds.

Yet, even with 3 or 4 questions repeated, taking an official test a second time is a much better predictor than an unofficial test. Over the course of four years I have taken four official practice tests and taken the real thing twice. All of my scores on these six tests have been 760 or 770. That is pretty reliable.

Good luck!

David