As far as I can tell, the revised version reads the same.
I agree that, at first glance, it seems like some kind of violation of something for someone to scribe a scroll of Fireball at minimum strength, slot-wise, and then for someone (possibly the same someone) to come along (maybe as little as 5 minutes later) and use the scroll of Fireball to cast Fireball as though with an 8th level spell slot. Where is the magical power coming from, right?
And yet...
Digging into it, let's consider the relevant statistics and salient details of the entire system encompassed by the premise:
The DC for the spellcasting-ability check, to cast a spell from a scroll "at a higher level, or even ... of a higher level than the character can normally cast", is DC 10 + the intended level of the spell. That is, to cast, using a slot of level X, any spell that would normally be cast using a slot of level < X, or any spell where X exceeds the level of the highest spell slot possessed by the caster attempting the action, the caster must make an ability check against a DC of 10 + X.
The cap on spellcasting attributes is, for most characters at most levels, 20, for a +5 modifier. In character builds where this constraint is overridden, access to high-level spell slots is more or less a given, and we can therefore confine ourselves to the consideration of a +5 modifier being the effective cap on the relevant check. ("But Dosk, what about the stronghold modifier?" -- Yeah, I'll get to that, bear with me. ;D )
Let's assume that, as a spellcaster, the character attempting this has at least a +2 modifier to their spellcasting ability score, and examine the bounded conditions arising therefrom.
Being able to cast a cantrip as though using a level 1 slot from a scroll requires a DC 11 [spellcasting ability] check, meaning that accomplishing this will be, according to the Law of Large Numbers, successful 60% of the time for a caster with a +2 modifier (as they need a 9 or higher), 65% of the time for someone with a +3, 70% for +4, and 75% of the time for someone with a +5 modifier.
As the level of the slot being used increases, those "easiest case" percentages lose 5% per cast-as-though-from-a-slot-of-Xth-level increase.
Trying to cast a spell from a scroll as though using a 9th level slot (no matter the level of the spell on the scroll) is technically possible 20% of the time for someone with a +2 modifier, and 35% of the time for someone with a +5 modifier. It's technically possible to do 10% of the time for someone with a casting stat of 10, and 5% of the time for someone with a -1 modifier to their casting stat (requiring a natural 20 on the roll).
So: absent the resources of a stronghold, upcasting from a scroll is between roughly 60% and 30% successful, depending on who is doing the casting and what level of slot they're trying to upcast with, with the more valuable upcasting being less than 50% likely to succeed across the full range of potentially-fielded values.
And that's where the power comes from (to address the fluff balance verisimilitude question). The result of failing the check is that the spell still fades from the scroll, just the magic doesn't go where the caster wanted it to. (This is an interpretation for narrative reasons -- if it works for you, great, if not, don't use it. I'm not saying this is what the rules do mean, I'm saying this is what the rules could mean.)
Ahh, but the Stronghold skews the percentages. It also imposes a 12-hour cast time and can't be any stronghold, it has to be "a magical stronghold appropriate to their class". This also makes sense to me (from the perspective of "where does the power come from?").
All in all, the likely costs in terms of failed castings would seem, in my view, to more than balance the likely benefits to any reckless attempts to exploit this feature, and the ability to allow a chance (that can be improved) of achieving something themselves is likely to spur my players (at least) to engage in more interesting narrative pursuits, be those securing permission to use a suitable stronghold before they have access to one, or trying to figure out some other way to get a bonus to the attempt by being creative.
I don't see the mechanic as being terribly disruptive the way that it seems to be printed. At worst, it means that players will sometimes get lucky and manage to conserve resources. At best, it's a whole new avenue through which interesting stories can be pursued.
