Wednesday, December 29, 2010

The Expected Build

I've done a little testing of the Giant Attacks list--but our breakthrough on Total AP Cost (TAP) characteristics has had me scrambling. Our formulation of it effects everything from Fast Company to Stretching and beyond (that was how the testing started: I wanted to know what -4 DM was /worth/).

This does bring up one thing though: I am testing with 4 test characters (Mixed DP and Armor, Full DP, Force Field, Full Armor). It's unlikely that we'd see some of these builds in conjunction with some of the TAP abilities. Who would take "Doesn't Take Hurt Condition" and leave themselves at only 14 DP on 64 AP? No one I can think of. That isn't to say it's horrible: the FULL ARMOR rating is usually the best in most of these common tests but when you start adding in a lot of TAP defenses you'll often see FULL DP or Force Field come out the highest victory.

Here's my thought: Some of these abilities -- and this goes for all abilities -- are going to show up where the player thinks they'll be the most effective. While it turns out that within a fairly narrow range of balanced battles Full Armor and Doesn't Take Hurt Condition are within tolerances for reasonable choices we're far more likely to see that on DP-heavy characters.

Does that mean we need to change our testing line up?

The obvious answer is that even if it /did/ I don't think we could--not easily--but the fact is that looking over these various builds it appears to me that the distribution of wins (the best mix of defenses) is, although somewhat skewed towards full armor (half one's DP spent on armor) mixed /enough/ that while there may be one optimal grouping for a given defensive strategy most of the choices a player can make are okay.


1 comment:

  1. I think we should take the optimal PoV into account for the costing. If a power wins 80% when mixed with DP and 65% on average, it should cost as if it wins somewhere around 70%+ to account for the assumption that most people who take it will have the DP to make it count.

    I wouldn't change the testing lineup or the over-all methodology, but I would, in effect, factor the highly unlikely builds out of the average.

    Note: If this changes the cost less than 1/16th of the total points (e.g. less than 1 AP at 16 TAP or less than 4 AP at 64 TAP), then I wouldn't bother -- but if the difference is substantial I think it's worth factoring that in.