This post continues the development of ideas presented previously, with specific numbers and the introduction of a few new ideas to the mix. As before, I'll tip my hat to Wombat Warlord, who has worked with these ideas a lot more than I have, and without whom I probably would not have pressed on with any of the rules hacks of SIFRP.
Because the system hangs on degrees of success, there are a few more game-balance levers to take into account than in, say, D&D. Attack spells have a base casting difficulty determined by some applicable defense on the part of the target. It wouldn't be at all unreasonable to treat Combat Defense as a defense against anything that can be dodged and Intrigue Defense as anything that can be resisted with Will, leaving only some variation of Fortitude to create. This unifies spell mechanics with the rest of the combat system, as well as cutting out one of the levers one might use to balance spellcasting.
Things that are not attack spells but still need a success/fail roll and need to pay attention to degrees of success, such as defensive/buffing spells or divinations with better results for degrees of success, are generally set to a difficulty of 9 (or 12, for the spells of the highest rank; it's also possible that there are cantrips at difficulty 4). It works this way so that characters with Wizardry/Piety 1 are novices, not yet quite able to cast their first spell. Characters with Wizardry 2 do not yet find their magic to be reliable, but can sometimes complete spells - ideally when there's no pressure and failing a spellcasting roll doesn't carry any real drawbacks. At three dice, they'll succeed more often than not, and can score up to one extra degree of success. Failure remains a realistic probability all the way up to six or seven dice, though, thanks to the possibility of penalties from Injuries, Wounds, Strain, or Curses.
So difficulty is the first lever. The second is the cost in Health (and Injuries and Wounds) or Composure (and Strain and Curses). The amount of Health or Composure damage that a player can endure before taking an Injury or Strain scales pretty seriously, but you're just not going to see wizards with a Will of 2 or priests with an Endurance of 2. Presumably such people wash out of spellcasting training, or experience character growth, or something. Anyway. I'd expect most PCs and NPCs to have a 4 in their "mana reserves" stat. I assume therefore that 12 Composure/Health is the power reserve of the average spellcaster, and that divine spells intended to be cast in battle need to have a slightly lower Health/Injuries/Wounds cost than an identical arcane spell, since the priest may be taking damage from actually being in combat as well.
Injuries/Strain represent a pretty remarkable amount of additional power for priests and wizards. (As a quick note, I want to say that I am well aware of how strongly these design structures resemble Mental Consequences for wizards in Dresden Files. This is not an accident, but at the same time I didn't feel like DF got the balance of a wizard's staying power quite right - if this system does nothing but re-create that one with more numerical granularity, I can go home happy.) I feel like the cost for basic spells should be about 4 Health/Composure, and if anything that's rather conservative (see Places of Power, below). This might scale up to 6, 8, and 10 for higher tiers of spells. Four tiers of spells sounds pretty good to me for this rules hack.
As a side note: Injuries and Wounds are particularly important for powering healing spells - it would cause serious problems to the game's balance if a priest could ever turn Injuries into a survivable amount of simple Health damage, or Wounds into Injuries. SIFRP rules are deadly, which is why I'm interested in using them in the first place. The long-term consequences of Injuries and Wounds are central to the system's grittiness.
Some arcane spells would undoubtedly carry a comparable cost, always inflicting Strain or Curses. This line of thought led me to some really interesting world-building: what if the setting has Laws of Magic of one form or other, and the violation of those Laws immediately results in a Curse? This creates a situation in which casting more magic than you can handle is just as bad for you as casting forbidden magic outright. Maybe these things are all part of the magical tradition that is available to the PCs, and an enemy tradition inflicts Curses for completely different things - one tradition bans communication with the dead, maybe, while another bans making compacts with elementals.
The third lever is the spell's effect, of course. Cunning is the ruling stat for an arcane spell's potency, but a spell can reasonably carry its own modifiers for its final effect, much as many weapons deal a few points more or less than the wielder's Athletics or Agility. Of course, those extra points of damage are also multiplied when the caster scores two or more degrees of success, so we'll tread carefully around such adjustments; spells don't have an obvious cognate to a weapon that takes up two hands. (Which immediately suggests to me the need for a series of negative spell qualities to go with the more desirable spell qualities that Wombat Warlord has already created.)
Defensive and utility magics rely much less on the spellcaster's relevant potency stat. A healing spell that restores Health might still use the caster's normal potency stat, but spells to transfer Injuries or Wounds... not so much. Anything that doesn't operate on the same scale as Composure or Health is generally assumed to either use the potency stat in a different way, or ignore it altogether.
Places of Power
Then I started thinking about the function of armor in combat. To stop all warriors from piling on the heaviest armor available, armor reduces the character's Combat Defense (making it more likely that attacks will hit and will score additional degrees of success), but also mitigates damage. (This tradeoff is tuned pretty well, in our experience; an agility-focused fighter is a very expensive character build, but viable.) Anyway, I want to give wizards and priests similar ability to mitigate spell costs, particularly because I think that attention to location is very strongly in keeping with Birthright's themes (and is a common theme throughout fantasy).
Wizards, therefore, have one point of armor against spell-cost damage while located in any province where they have at least a zero-level Source holding. (This is a Birthright thing; wizards should be building up a network of sources and ley lines as far and wide as they can, so as to increase the number of provinces they can affect with Realm spells.) Furthermore, they have a base of two points of armor against spell-cost damage while within their personal Sanctum, and they can increase this further, up to (in theory) 10 armor. Spells always have a minimum of 1 Composure damage. A wizard intruding on another wizard's turf has an uphill battle ahead of him.
Priests have one point of armor against spell-cost damage in any temple, or on any holy ground, consecrated to their deity, even if it is a conflicting sect of their faith. Priests gain additional armor against spell-cost damage (all of which is, in in-game terms, reduced requirement - the deity demands less of their blood to grant the spell) based on the relative holiness of the place, typically as influenced by the places where martyrs of the faith died. The power of a true martyr's blood can never fade, unless the land is defiled; the slightest portion of the priest's blood stirs the soul of the martyr (who rests with the deity in paradise) to intercede on the behalf of the priest.
I would also like to come up with a workable, flexible cooperative casting system, because that's something D&D has never done well, but that feels very appropriate to the way the Birthright setting thinks about magic. I seem to recall that SIFRP already has rules for cooperating in skill use, which covers the initial Wizardry check (and also opens some design space for spells with difficulty so high that no one character can ever reasonably reach it). The casting cost is the only other thing that I'd need to design, since I think it makes sense for priests to share around the cost of a spell, and not unreasonable for wizards to do the same. The Ritual Fatigue chart used in Dust to Dust for cooperative casting would work pretty well here, though this rules hack mostly uses spells with a higher cost in Health or Composure than have been seen in the rituals of DtD.
(Also, this is the first step of working out SIFRP rules that I could keep twisting until they fit DtD, since Kainenchen and Ben both requested that I work toward that end.)