The calculator is basically a simulator, it simulates a real result each time. In a simplified way, imagine if you played 4 games right now.  You would either lose all four, win one, win two, win three or win all four.  There are only 5 possible results.  Now, if you are a good player, it is a bit more likely that you would win three or four than it is that you would lose three or four.Now look at the variance calculator, when you put in a winrate that reflects on how often you see a positive or negative sample.  So while you may see two losing results in a row, if you have a positive expectation, if you calculated 1,000 times on our variance calculator, you would be likely to see over 500 winning examples (and depending on your winrate and sample perhaps a lot more).Another way to look at it is that you have a coin that lands on your side 60% of the time and on your opponent's side 40% of the time (this represents your winrate, and that of a good winner in many husng formats).  If you did 100 coinflips (or played 100 games) you're not going to necessarily win 60 and lose 40, there's variance (or uncertainty) in results versus expectation.  That's what we're trying to give you an idea of with the calculator, what kinds of swings one can expect with a various edge.  The buyin level and games just serve to put it into identifiable numbers and game length for each individual user.Let me know if that makes sense.

Exactly, not even close.  Hokie and I used to mess with it more often, and I remember discussing some 5% type ROIs in buyin levels like $220s, and even over 10k games you see results can vary a ton.  Just goes to show you how much variance can impact results, but if you're good enough, it means a 60k year if you run like shit instead of the 100k expectation, and most people will accept that (not like it, but accept it).