There are 9 experimental questions, and 28 questions that count in Quant. Q50 and Q51 level test takers will answer all or almost all of their experimental questions correctly, because experimental questions will span all difficulty levels (so will mostly be easy to a test taker at the Q50 level), while the questions that count for the Q50-Q51 test taker will mostly be high-level problems, and might be challenging even for a top level test taker.
As you correctly point out, it's hard to translate # of incorrect answers to a score on an adaptive test. This is a complicated subject to discuss in brief, but question difficulty isn't perfectly predictable, and will vary from test to test. Some GMATs are full of very hard questions when you perform well, and then you'll have more latitude to make mistakes; you'll be able to get a Q51 with more wrong answers. Some GMATs have fewer hard questions even when you do well, and then you'll need fewer errors to get a Q51 (but it will be easier to answer each question correctly). Whether your test is easy or hard shouldn't affect your score, but it will affect the number of wrong answers you can afford to have to score a Q51. Dabral, who is a reputable GMAT expert, did some studies with GMATPrep, and found that it's sometimes possible to get a Q51 with as many as 6 incorrect answers. But the free GMATPrep tests have a very large supply of hard questions to draw from, and you sometimes might not see quite as many hard questions on a real test, depending on a few factors. It sounds like that was true on your test, since two wrong answers most of the time will produce a Q51 score.
It's also true that not every Q50 is the same, because of rounding. With only two wrong answers, I'd bet your Q50 is really more like a Q50.4, and it's the unrounded score that is used to produce your score out of 800. If you google your Q/V score combination, and find that other test takers with the same Q/V split sometimes have a total score that is 10 points lower than yours, it's likely your Q score was just a small fraction of a point away from Q51.
Normally I wouldn't agree with the post above, that says there's a thin line between Q50 and Q51; since you often need to be almost perfect to get a Q51, but can afford a comparatively large number of mistakes to get a Q50, I'd normally describe the improvement from Q50 to Q51 as the biggest one-point jump you can make on the GMAT in either section. But in your case, it sounds like you're either very close to the Q51 level or are already there, so it seems to be true in your situation. A Q50 is a great score regardless, and raising it to a Q51 isn't likely to make any difference if you're applying to MBA programs.