I think you should arrange your dice more neatly...
Yeah. If they were my regular dice set, it would have all the maximum values face up so all the weight in the dice settle at the bottom. Gamers and their silly dice rituals and superstitions, right?
Personally, I've been told to alternately boil or freeze my dice when I get them to show them who's boss. My current personal theory is the Dice Fairy can only inhabit one set of dice at a time, and I have to guess which set that is. But having multiple sets for her (naturally it's a her) to live in makes her happier, which makes my rolls better.
No, I don't think it makes sense, either.
It doesn't make any sense. Gamers seem weirdly prone to these strange superstitions about luck, karma, dice hate/love, etc... I find it really funny given that most dice at a table aren't rolled often enough for them to roll "average". I played with my various D20s to see if they were off from the expected/variance values and it took over 400 rolls of each die to get the power of the stats high enough to find that they were rolling pretty average (less than .01% deviation from expected norms after 400 rolls each). So, given that amount of rolling vs the, say 20 rounds of combat, means that I use the dice less than 100 rolls a game (skill checks, attacks, random BSing) and so I'd expect to see "streaks" of low or high values, bounces to extremes (rolling 1s and 20s more), sudden falling to the "middle", etc... as they aren't being rolled enough to be "average" each night.
My group hates me for all the "crits" that I roll in games (currently a Ranger in a Pathfinder game), but then, I roll a d20 8-10 times each combat round right now, when I bother to roll at all. So, one roll a round coming up 19 or 20... falls right on the "average" curve (10% chance of rolling a 19 or 20). I'm also the biggest fumbler in the party (roll lots of 1s when I roll) for the same reason.
For games like EotE or D6 Star Wars, rolls tend to be a bit more towards the "middle" as more dice are added to a pool as the variances from the mean for each die starts to counteract each other.