I excluded all 2d6 (and casualty, and pitch invasion) rolls and did this by hand, since, as previously stated, it was a retarded idea to include them, preventing people from easily judging how good there luck is. I mean you could roll a 1, and then get 5 6's on armor, injury, and casualty, and it would look like you had great luck instead of the
4- 16.20% (yay average)
1- 29.91% ! (wtf?)
Average roll- 2.90 (should be 3.5)
Oh and it doesn't end there. Lets look at those 2d6 rolls, the ones you want low numbers on since they are rolls against your own armor and injuries. I excluded fouls because I wasn't sure if those were counted as me or my opponent rolling them.
12- 5.12% (that is fully double expected)
Avergae- 7.93 (expected- 7)
And lastly, block dice, just to show that whatever is wrong with the RNG is universal. I didn't track d8's or casualty rolls.
Block dice (raw)
Pow- 10.48% (How close that is to the d6 result raises a major red flag. We all know through the challenges that 6=pow in the rng, but taken seperately, the chance of getting basically the same unlikely results in two independent trials is astronomical.)
!Pow- 10.01% (Again eerily close)
Push- 38.99% (Much higher than 3+4 were, interestingly)
Both Down- 20.14% (again significantly higher than 2's was)
Skull- 20.38% (Way better than the d6 rolls, but still a good deal higher than the 16.67% expected. interesting.)
I'm not even sure what this means. Except the next time someone tells you bad luck is in your head, you now have the numbers to properly tell them to sod off.
Coming soon, my opponents rolls for the same 150 games. After all, game is still fair if both sides have equally bad dice, right? So these mean nothing without seeing both sides.
Right now, I can't make any conclusions, because these numbers could be accounted for in many ways. What they can't be is "bad luck" anymore. There are just too many dice involved (several thousand) for this to not be beyond bad luck. Also, I find it highly suspicious that the programmers added armor and injury rolls together with 1d6's, because when added together the numbers come out very close to right. This implies the programmers KNEW about the problem, and intentionally tried to mask it. Maybe they are just incompetent statisticians and included them because they are just too dumb to see why numbers you want to roll low don't mix with numbers you want high, but that isn't much better in what it says about them.
Also of note- the d6's are about ~.5 lower and the 2d6's ~1 higher (twice as much deviation thanks to twice as many dice?) This implies the problem exists because of the rng's rounding, though it also implies some rolls are improperly rounded up, others are improperly rounded down. why?
Sorry to the opponents I was slow against recording these numbers by hand, I don't trust the dice tracker, and now I have a good reason not to. And will take a couple days to crunch the numbers on opponent results.