Technik4 said:
My question is, given the above arrays, what is the power difference? If you are of the opinion that 3 stats will drive most concepts, the difference between the 36 pt and 25 pt are 16,15,14 and 16,14,14. Or, what point buy total would you equate the 36 pt array with?
Technik
My opinion begins with the 3 highest stats & the lowest allocated stat will drive most concepts, placing the lowest stat is almost as important as placing the highest. I say this from the pov that whenever I build (a significant npc/) PC the placement of the 8 is the source of greatest thought.
But to your question:
The 36pt array is only worse than 36pt buy in that it forces the allocation of points to any given score. I would not draw a comparison with the default array and 25pt buy because imo, default array is almost always the best use of 25pt buy. You would need to look at the actual 36pt array (your's presented, 16,15,14,13,12,11) and look at the 'dead' points, i.e. the 15,13,11 which results in 4 'dead' points.
Now we need to careful with using the 'dead' point term because although there is a tendancy to boost the highest stat every 4 levels there is also a desire for immediate gratification competing with this. Another point is that 13 scores can be considered well spent if the otherwise non-essential bonus opens up feat chains.
I come from the pov that every 'bonus' lost by the player from a 25 point baseline is worth .5 point per ability attribute, i.e. 4d6dl loses minimum security and so is worth the average of 28pts. I also believe that there should be a diminished return as every incremental loss of control by the player should receive less compensation, otherwise the linear boosts will drown the 25pt buy in sheer numerical superiority.
So if I reverse engineer your 36pt array along these principles I would consider it equal to
32pts; -3 for loss of customization, -2 for 'dead' points (halving the actual loss because the 15 & 13 open feat chains/immediate gratification), +1 for diminished returns.
However we can all respect the fact that nothing like actual playtesting will confirm or deny my theory.