I had given a CAT test from mba.com and scored a 680 around a week back. But I do not consider that as an accurate reflection of my ability as I started that around midnight, and by the time I was on the verge of completion, it was 3 am so I was very tired. Subsequent review of my mistakes the next morning revealed that I had not even registered what the question was asking, since I was so sleepy.
You should never do this. For a few reasons, #1 you do not know if this test gave you an accurate reflection as you were too tired and #2 when you retook the test it was not accurate as you saw repeated questions. #3 unless you typically go to sleep at 3AM you have actually disrupted your sleep pattern and this can throw you off mentally for several days.
That 3rd statement is true. Recent research into sleep patterns has shown that even varying the time you go to sleep OR the time you wake up by 1 hour can show effects on your cognitive performance days later. (It is like the ripple effect in traffic. If there is an accident in the morning there will still be a traffic delay in the same spot hours after the accident is cleared). And there is no such thing as "making up" sleep.
To quote
The Organized Mind by neuroscientist Daniel J. Levitin:
"At one time or another you probably thought that if you could only sleep less you could get more done. Or that you could just borrow time by sleeping one hour less tonight and one hour more tomorrow night. As enticing as these seem, they're not borne out by research. Sleep is among the most critical factors for peak performance, memory, productivity, immune function, and mood regulation."
"Going to bed just one hour late one night, or sleeping in for just an hour or two one morning, can affect your productivity, immune function, and mood significantly for several days after the irregularity."
I wanted to ask you - is the level of GMAT tests on the economist a bit more challenging than the actual GMAT? An answer to that would greatly help me know where I stand.
I am not aware of the pattern for the Economist tests. I am not sure how there scores relate to the actual GMAT. The Veritas Prep practice tests are based on about 8 million actual student responses and are based on Item Response Theory - the same basis for the actual GMAT exam's scoring algorithm. And even then people can score differently.
So you have three exams and none of them can actually give us a way to predict your score other than to say that you will likely score above a 680!
Have you taken the second GMATPrep exam?