We do take these sorts of concerns seriously. Thanks to Zulghinlour Industries, Inc., we have ways to track this sort of thing, and ask "Are APs getting screwed?" in a scientific way rather than an anecdotal one. I fished in the data, and threw roughly three months of AP kills into a spreadsheet. The results:

Given the toughness of opponents, the expected number of charges handed out over this time period would have been 543.66. The actual number of charges handed out was 539, which is within 1% of the expected value. Given the size of the data set, this is within statistical noise.

Ancrhuljin (whose entire career is inside the above data set) should have expected 101.98 charges over his lifetime, given who he killed. He collected 99. A little low, but also within statistical noise, given the smaller data set.

The slight discrepancy is even in the direction you'd expect. If an AP player convinces himself to delete after his first real bad streak, you'd expect a slight lowering of the observed average. (Random variation with a guaranteed negative streak at the end will result in a slightly low average.)

In other words, stop blaming the random number generator. It's doing its job.

Valguarnera, Number Cruncher
valguarnera@carrionfields.com