Found a possible error in the equation. Before I was assuming A = 446.5 which was based on
http://db.vanillagaming.org/?spell=10181. However I looked in my spellbook the frostbolt damage ranges show that the average is actually A = 486.
Anyway I updated that so let's do an example to show that the ratio between crit and spell power holds true even if you consider damage multipliers.
So let's say your frostbolts hit for 1000 unbuffed (Sp = 631) and you have 25% chance to crit. We want to know how much spell power is equivalent to 1% crit for Sp =631. So I come up with 1% crit = 9.8 spell power. Now let's check to see if this ratio holds up if we take 100 frostbolts and COE (1.1 damage multiplier) the mob we're attacking. For simplicity let's say we have 100% chance to hit.
So we will have 74 hits and 26 crits = 1.1*1000*74 + 1.1*1000*2*26 = 138,600 damage.
Okay now let's instead increase our spell power by 9.8 and set crit back to 25% and see how much damage we do.
So we will have 75 hits and 25 crits and our new hits with the added spell damage will be 1008.
So our damage is 1.1*1008*.75 + 1.1*1008*2*.25 = 138,600 damage.
As you can see the damage the same as if we added 1% crit by adding 9.8 spell power. Now you can't get items with 9.8 spell power this is just an example but if an item had 10 spell power it would be clearly better at this condition. The point of this example is to show that damage multipliers don't make a crit ratio all the sudden become better in favor of crit it remains constant because damage multipliers are applied to everything.