The difference between those two items is going to be very minimal. The set bonus you get from spellfire is pretty weak, but it can help make up some of the deficiencies on the items in that set.
In a 240 second fight, you are going to have to cast a lot of frostbolts. Straight up arcane spell power is going to be a much weaker stat than general spell power as a result.
The model for arcane mage figures out how many max stack arcane blasts you can cast, and then it looks at your mana pool and removes as many of those casts as necessary to replace with a filler rotation. It perfectly figures out how to do maximum damage with your available mana in the given fight length. I realize that it will be hard for players to predict the fight length ahead of time to know how much arcane blast they can use vs filler. This model gives you the correct gear for ending a fight at zero mana, which is the goal with an arcane mage.
The filler rotation for mages using frostbolt comprises 1x arcane blast and however many frostbolts fit before you can cast arcane blast again without stacking the arcane blast buff. Unstacked arcane blast is the most efficient filler, followed by frostbolt.
If you reduce the fight length to 120 seconds, you will see the optimizer pick the spellfire gloves. That is the fight length at which you can use enough arcane blasts to make that item better than the gloves of the aldor.
Also, I am giving no value to the 2pc bonus of the aldor set. It will have some value in the game, but will vary greatly. So I’d say that this is even a very conservative estimate of how good that item is in your case.