Having chewed through some of the math, I think I would make these changes (changes have been underlined):
Magic Missile
Cantrip evocation
Casting Time: 1 action
Range:
60 feet
Components: V S
Duration: Instantaneous
Classes: Sorcerer, Wizard
You create a glowing dart of magical energy. The dart hits a creature of your choice that you can see within range. A dart deals
1d4 force damage to its target.
However, a dart produced by this spell may not deal more damage than your spellcasting modifier (minimum 1). The darts all strike simultaneously, and you can direct them to hit one creature or several.
The spell creates more than one dart when you reach higher levels: two darts at 5th level, three darts at 11th level, and four darts at 17th level. You can direct the darts at the same target or at different ones.
The math problem I'm seeing relates to when the spell ends up being better than the alternatives. Because the spell now has no spellcasting modifier benefit, it's equally potent when you have a low modifier (+2 or less). That means if you're multiclassing, or a High Elf, or whatever, you could have Int 6 and you'd benefit just as much as a character with Int 20. That's not a good design. The fix to cap damage per dart to the caster's ability bonus is fiddley in a way a really hate, but it's about the easiest way to do it.
Additionally, I might consider changing mage armor to make it give you resistance to magic missile. The immunity clause on shield doesn't make much sense anymore since it's such a short term spell now. Another alternative would be to switch it from force to piercing damage, but I don't think damage type is that meaningful.
I didn't make this clear in my initial post, but I'm not planning on using it at all. It's just a design exercise. That's why my question was, "What would a magic missile cantrip look like?"
So, I would see this as violating my own "new spell rule 0" - no neexdpell should be the new "must have" that practically everyone takes.
I think you're overvaluing it. Would you take it if each dart dealt only 1 damage? I don't think anybody would. I would go so far as to say that it's
strictly incorrect to take it if each dart deals exactly 1 damage. So we're just arguing about the balance point. Where's the balance point between 1 and 1d4?
Math-wise, it's fine, but some people might find it kind of boring.
That's true of anything, though. It's a play style preference more than anything.
D&D was always stingy with force damage, especially free/cheap ones. Like, look at 5E's Eldritch Blast, it works very well for damaging baddies, but it's a sitting duck when you need some property demolition (it only damages creatures, not objects).
Yes, that's why this spell has the same implicit restriction.
I don't think there's anything wrong with the first draft you've posted,
@Bacon Bits, it seems balanced and playable. It's just not my style.
That's fair.
Another effect of giving Magic Missile as a cantrip is that it has a substantial impact against spellcasters maintaining Concentration, and enemies who are unconscious and making Death Sves (and so, typically, Big Bads). Auto-damage without any expenditure of resources, even if it's a single point of Force damage, represents a significant benefit.
That's still really narrow. It still costs an action to cast, and it's rarely going to trigger a save with a DC higher than 10. And this is 5e, not 3e, so force damage isn't remotely as good. It doesn't "penetrate the Ethereal," for example, or automatically have any of the properties that it used to. The designers clearly remember those rules existing because lots of ethereal things are subject to force damage, but they're not actually present as general rules for force damge. In the de facto sense there's really only 4 damage types in 5e: non-magical, fire, poison, and everything else. Yes, you might run into shadow dragons, ghosts, and helmed horrors all day, but that's not a meaningful consideration to the design. If it were, eldritch blast wouldn't exist at all.
My only concern is potential abuse, like a Wizard School with a 100 apprentices all focus-firing to eliminate one target after an other.
I'm not remotely concerned about that. Any situation where that kind of thing would happen is not one which would happen in the game normally. Further, 100 apprentices could also cast magic missile the first level spell. And for the same cost, you could raise an army and obliterate anything already. Scaling out is not a concern. I really don't care if the game breaks when you turn it into a tower defense simulator.