(Well, now you've made up for it. )
And I think the elegant solution is both obvious and not against the rules - that "going first" doesn't preclude fast-casting later in the same round.
Well, you can "degage"(sp?) (evade/elude/disengage) a block.
A fast-cast is not necessarily analogous to a block or parry - it's analogous to a magical flinch. What that flinch does, whether defensive (perhaps a block/parry) or offensive (a quick stop-thrust) or something else is up to the caster.
First, let's disabuse some folk of the wrong model of how fast-casting works.
In Ars, "1 combat round" is an undefined amount of time. There is an approximation ("about 6 seconds"), but that's not a rigid Olympic-timekeeping-quality amount. And each round is then abstractly broken into initiative phases - these are not even "fractions" of the round, but undefined "your turn" and "her turn" and "their turn". Initiative determines "who goes first" - and fastcasting doesn't change that. The fast-caster is always reacting, never "jumping back in front" of the action.
Let's say Mage A rolls a "7" and Mage B rolls a "5". Mage A acts on 7 - Mage B can try to fastcast, but that will never occur faster than the initiative of the "surprising event" being reacted to.
In our example, Mage A goes first, and he casts a spell, or swings his mighty quarterstaff down on Mage B's head - or maybe he's unconscious and the "surprising event" consists of him falling out the 3rd story window down toward Mage B - all same same. Mage A's action has occurred on Initiative 7.
Now, Mage B can fast-cast as a reaction to that event. The event is already happening - the PoF has been cast, the quarterstaff swung, the body falling - can't retro it back to not-happening. What can be done is to react to stop that effect from landing (or otherwise avoid it), or (as suggested above) get in a quick retributive strike before Mage B suffers the inevitable effects. (Maybe Mage B has a ward vs. flame, or a Ward vs wood, or a Ward vs corpus - nothing says the fast-cast has to be defensive (even if the rules "imply" that).
So now we assume Mage B makes their Fast-cast roll*, beating a difficulty of Mage A's initiative. Mage B's spell is spontaneously cast* as the first spell/attack/body takes effect (unless multiple spells are being fastcast/etc.) This can be construed as a new "surprising event" - nothing that I can find in the RAW says that once you take a turn or cast a spell you can't then Fast-Cast in response to a new "surprising event".
(* more on all this dice-rolling later, so hold that thought...*)
The advantage Mage A has is that his fast-cast spell is against an easier difficulty factor, since Mage B rolled lower initiative than Mage B. Keeping in mind the last paragraph, if Mage B blocks the spell - that's not all that surprising, no fast-cast for Mage A. If Mage B sends a spell toward Mage A, that is surprising - and Mage A can try to fast-cast to avoid that attack, same as Mage B did Mage A's attack "before" Mage B could even act. Fair is fair.
More, if Mage B misses that fast-cast roll (quite possible!), they are done casting spells for that round! They can do something else on their initiative, but not cast a spell. (Some SG's require that all combatants declare their intended actions before determining initiative for the round - makes things feel more like "combat" and less like a chess game.) And if they do something "surprising", then Mage A has the option to... fastcast, as a response to that. (And unlike Mage B, Mage A doesn't care if they lose their ability to cast spells for the rest of that round - they've taken their action!)
And all this fastcasting is "with SG approval" - I hope few SG's would allow a mage to claim "surprise" that his target did/didn't die from the first spell, or "I'm surprised that I'm still standing, so I'll now fastcast...". That's not in the nature of FC'ing. (I would even suggest that it has to be SG-invited if not SG-approved, never assumed by the Player as an option they always have.) This gets back to what "an initiative round" is - a fast-cast is not a "half-action" or a "surprise attack" or a "quick action", it's a way to allow a mage to be magical when they really need to be. If they don't have anything legit to react to, initiative and actions - and spellcasting - proceed as normal.
(If magi in your Saga are always trying to FC every combat round until they fail their roll, sending off random spells at whim, think about what I've said above!)
I'm not even sure there is a "technically" needed here - that's how the rules read, and it makes sense. (Or, at least, the opposite interpretation makes far less sense.)
- Has a potential to - this is far from guaranteed! (And here we are getting back to the notion of "assuming Mage B makes all the rolls" to fast-cast...)
First, the mage has to beat Mage A's initiative, and that roll has already proven to be higher once (so odds are it's a decent roll - not always, but more often than not.) For these tenuous additional spells, each takes a cumulative -6 to that difficulty roll - that's more than an average d-roll, so that's a serious knock. Take a LOT of Finesse to make up for that.
And then the spell has to be Spontaneously Cast at -10 to the Casting Total - and only the rare archmage could cast such as non-fatiguing and thus guarantee that roll. (And with 2 extra botch dice, it's not something done lightly!)
And Mage A can always try to counter any "surprising event" these new spells may create with a FC roll of their own - and against an easier ease factor.
To me, this creates more of a feel of intensely fast parry-riposte spell dueling than any "paradox of time" - each mage is seeing what is happening and countering it (or praying/hoping to), until one fails. And then, if the other can make yet another roll with the added-6('s) of penalty, they keep going.