With either the equation he used or Dervy's, the results would have come out the same. The issue is with his simulation; it's essentially a solo dummy parse, showing that 10% off cast time adds as much damage as doing 20% more damage per spell.
You'll get about 12% more damage from extra casts and that boost to thunder DoT tics, but the only thing he has giving that extra damage is Thundercloud and Firestarter procs. Something is wrong with his code.
