Exploding Penguin -> RE: Bludgeon Balance (10/7/2013 23:40:35)
|
quote:
you would get a straight line every time. No curves, no limits. An average simply allows people to look at a set of data as 1 comprehensible number. But you should be getting a straight line considering the skill point progression (excluding massacre cause that has wonky damage increases). However, thanks to ED's strange rounding system, it becomes curved and really loses that nice correlation/r-value of 1. But my main point is that there are really strange inconsistencies in damage boosts and ratio increases because the LSRL (least squares regression line) has little significance since the rounding system is quite strange. Actually, now that I look at it, finding the mean is really not a great idea for comparisons. However, I know what's perfect for comparing data side-by-side, and that is not the value at max skill point investment because you are only giving a piece of the puzzle that will lead others into thinking what you've stated is entirely true. Not that I'm accusing you, I'm just using an analogy with such connotation. Rather than finding the mean, the true best way to do side-by-side comparisons would be to find the mean ratio of all the skill points, then use that to standardize every single skill point investment. Once you've found the z-scores for everything, comparing things will be much more comprehensible and easier to make clear judgments on. Of course, there's still going to be inevitable bias of how much that 1 turn less of cooldown is worth, and how much that warm-up period is worth. Reply to deleted content removed. ~~Therril Oreb
|
|
|
|