A few things:
- you should only really trust your scores on GMATPrep tests. Company tests aren't using real questions or the real scoring algorithm, and some of those tests are very different from the actual GMAT. Unless you need to take a lot of tests, you should only be using GMATPrep tests - buy the additional tests if you need more.
- there's another important reason to rely primarily on GMATPrep, and it's something it sounds like you've noticed during your practice attempts. You can only get realistic pacing practice using GMATPrep, because other tests may not be realistically time-consuming. Some company test questions are far too long. If you use those questions as your pacing benchmark, you'll find yourself rushing unnecessarily through the real test, and possibly making careless mistakes.
- one thing about familiar questions on diagnostic tests - you should try to answer them 'honestly' (ask yourself 'would I get this right if I hadn't seen it before?'). If the answer is yes, answer correctly, and if the answer is no, then either use whatever criteria you would normally use to guess, or use a random number generator to pick a random answer choice (you shouldn't obligate yourself to pick a wrong answer every time - that will unfairly deflate your score). You should also let a fair amount of time elapse before you confirm your answer, so you don't get a misleading impression about how well you're pacing yourself. It's hard to say if those familiar questions on your diagnostics have inflated your score. If they were all questions you'd normally get right, and you let some time go by, your score shouldn't be inflated at all, but if not, your scores might be slightly higher than they should be.
- I would not agree with the advice above to try to rush through easy questions, unless you know you don't make careless errors. In my experience, easy questions usually take longer than hard ones, once you see what to do - the logic of an easy question is more immediately obvious, but the question designers balance out the time these questions take by making any calculation more awkward. One important principle to understand about the scoring algorithm is that wrong answers on easy questions are very damaging to your score. So you don't want to run the risk of careless errors on those questions.
- some tests start easy, some start hard. If your test starts easy, wrong answers early on will hurt you quite a lot. But if your test starts hard, wrong answers early on won't hurt you much at all. Don't panic if you need to guess at some early questions, especially if they seem difficult. There is no 'rule' about the first ten questions; it all depends on how hard those questions are.
- you'll learn by taking additional GMATPrep tests just how much of an issue (or hopefully non-issue!) pacing is for you. If you're having trouble reaching the end of the test without guessing in order to finish, you will want to find spots to save time earlier in the test. Bear in mind that wrong answers on very hard questions don't really matter (no matter where the question appears), unless you need a perfect Q51. So if you've fallen behind an ideal pace in the middle of a test, and 30 to 60 seconds into a question you feel you don't understand it or you have no idea what to do, you're probably looking at a very difficult problem. Those are the best spots to guess if you know you need to save time.
- I would not have advised you save the
OG to the end of your prep - those are the best questions to practice from, and I'd suggest you start using them as soon as you can. You'll want to revisit those questions that you had trouble with the first time, after you review the relevant concepts.
Good luck!