Re: Divide by Three
/3 is little more than just using no framework, with a cost break. It doesn't fix the real problem of powers-miscosted-for-fantasy, and it introduces some serious granularity problems. The power difference between a 2cp spell and a 3cp spell would be huge.
So I would start by dispensing with /3 and implementing x5, which is to say that the character would spend character points for "power points" at a 5-to-1 ratio, that are used to buy powers. This way they can buy about the same powers, but without the granularity that occurs when you apply -2.25 in limitations to your 30 AP spell, and then divide the result by 3 again.
Then I would alter the magic skill to take advantage of this new granularity. The current system has two problems. The standard penalty is -1 for 10 AP in the spell, which itself is granular; spell AP generally stays within the range of 10-50, so spell difficulty tends not to be an issue. The bigger problem is that basing the penalty on AP overlooks the limitations. Why is it easier to cast a 20AP spell with no lims than it is to cast a 40AP spell with OAF and Extra Time 1 day? Basing the penalty on the real cost is more in keeping with genre conventions, IMO. (There is some potential for abuse this way, I know.)
Then finally I would divide up magic skill. Instead of having one blanket skill roll that applies to everything in the mage's repertoire, the character might have to buy skills for each GM-defined magic "school", each of which has access to 8-10 of the available powers. So a spellcaster might be very good at Binding spells but less so at Finding. Magic skills would have to be cheaper, probably 2/2 or 2/1 depending on how restrictive the schools are.
So this system would provide the cost break that magicians seem to need, yet improves on the original system by reducing granularity problems and providing incentives for casters to buy related powers for flavor.