A few questions, since this is spitball, you might not have gotten to these yet, but if you have, I don't notice these details above.
1) Is this an open ended, or closed roll?
2) Are there UM results?
Comparing this to the previous systems, I disliked the RM2 method, since it required you to roll twice, but on the upside it did make tampering with spells somewhat risky. I had issues with the RMSS Spell Casting Static Maneuver method, in that it tended to act like "Magical Ritual by proxy" by making tampering too safe and easy. . . .RMC made the RM2 method into 1 roll, but kept the risk levels fairly high.
This system seems to resemble the SCSM method most, so it runs the risk of having similar issues, I suspect this is familiar to anyone who used SCSM or one of it's variants:
"I use flamboyant gestures, loud incantations, take 10 rounds, plus my skill, level and stat bonus, plus etc, etc". . .which resulted in a total bonus that allowed things like overcasting 10 levels casually.
Since all the bonuses and penalties come out of the same pool, if you can accumulate a lot of bonus, you can ignore or circumvent most or all of the restrictions in place on casting. This mostly was a serious problem with out of combat casting, where casters seemed to routinely overcast. (As opposed to the ESF methods that raised the UM failure rate, with few angles to reduce it.)
This method does address some of the problems in SCSM by toning down the input bonuses, but some care needs to be addressed to the scale of all the inputs. . .like a fair elf mentalist with a 101 presence is going to start off 0 level with a +45 spell casting bonus just due to stat and race. With the "per rank" bonus reduced to +1/rank that's 45 rank equivalents which seems like a really out of control bonus input, with that sort of headstart, that character could casually overcast 6 levels with 10 ranks in a list. There are loads of inputs that all need to be put on the same scale as a method like this is refined and detailed to bring all into balance. That's all just detailing, but it's important detailing.
I'd suggest considering the RMC scaling options (options 3.1 and 3.2) for incorporation into the logic, where a 5th level caster wanting to cast a 1st level spell with 5th level effects (RR vs 5th level, X/lv effects at 5th level) needs to spend 5 PP and it casts as a 5th level effect.
Overcasting seems to be the boogum in any SCSM type method of summed bonuses/penalties. . .game balance is only slightly effected if you balance say "no hands" and "No incantations" against "Level bonus" and "Take extra time", while instances where a caster piles up every possible bonus to offset an overcasting penalty tends to be where summed modifier methods begin to resemble mini magical rituals and bend or break game balance. . .perhaps just use the ESF logic of raised UM failure rate only for overcasting? (Or have an option to make overcasting use the ESF method for GMs who want to allow overcasting, but maintain it as risky, rather than casual.)
As a complete aside, a lot matters as to what feel you want as GM, and I suspect that if HARP can have 3 combat systems, it might be worthwhile to consider multiple magic mechanics or one core mechanic with options to signifigantly shift it's results. . .for a no magic game, it's easy, for a high magic game, with lots of caster stretch room, a SCSM style method seems ideal, while in a more limited magic game, the RMC style method seems to work better.
