Hello.
There are a few complaints that can be read again and again on the forums :
1. In the end game, arcane casters' damage become irrelevant, because their damage output do not follow the monsters' HP inflation. They can end up spending their entire mana bar on a monster without killing it.
2. Firewall is the only damage spell wizards and sorcerers are going to cast once they get it, as other spells are mana inefficient compared to it. Result : developpers had to work on many spells that are seldom used.
3. In the end game, casters' damage pales in front of that of melee classes.
4. In mid game, firewall is overpowered.
A few fixes have been suggested, but most make one or more of those issues worse while fixing another.
Here is my suggestion.
Give wizards and sorcerers an ability (I like the name "efficient destruction") that would make direct damage spells cost gradually less mana as they level. Something like 5% less mana per level after the fifth, so a 20th level caster would spend only 25% of the current cost on an attack spell.
Only direct damage spells would be affected. That would include metamagic feats used on them, but exclude damage over time spells such as firewall.
The result would be to make casters'dps actually an option, as they could sustain it long enough to actually use it, but not so much that squeezing the trigger whenever something move would be a good idea.
Since firewall would not be affected, it would not make it any more overpowered. Actually, I think it could use a little nerf, though as the only reliable source of dps for arcane casters now, it should not be nerfed before the proposed change has been tested as viable.
I think it fixes all of the aforementioned issues :
1. They can now cast many more nukes before running out of mana.
2. Other spells become more and more mana efficient, so they can compete with firewall. Arcane casters can now enjoy a little more diversity.
3. Casters can now afford casting nukes in quick succession, increasing their dps.
4. Arcane casters not depending on a single spell anymore, it can safely be nerfed.
Additionnally, it has no effect on crowd control and buffing. There are also many things that can change the cost of a spell, so existing code could be reused.
However, my numbers may be way off, I'm not sure how many more spells an arcane would need to cast to be viable at end game. Can the veterans help with those?