Linneman's Alternate Magic

Jonathan Linneman, over on the Monstrous Matters blog, posted a really intriguing idea for a simplified magic system.

He proposes a sister mechanic to Hit Dice: Magic Dice. If you want to level up in spell casting instead of fighting, you get a Magic Die instead of a Hit Die. When casting a spell, roll a pool of 6-sided dice equal to your MD.

MD pool ≥ Spell Level x 5.

Success: Spell is cast. Remove one die from MD pool (minimum of 1).

Failure: Spell is not cast. Add one die to MD pool.

This looks great. I love the idea of taking the bookkeeping out of spell casting. I love mechanics that make failure functional. And this looks elegant. But will it actually work?

I went over to anydice and looked at the numbers. I ran the Spell Levels up to 6 and the MD Pools up to 12, because my brain defaults to B/X magic.

Spell Level	Target		MD Pool	Minimum Roll	Maximum Roll
1	5		1	1	6
2	10		2	2	12
3	15		3	3	18
4	20		4	4	24
5	25		5	5	30
6	30		6	6	36
			7	7	42
			8	8	48
			9	9	54
			10	10	60
			11	11	66
			12	12	72

MD Pool	Percent chance of succeeding per Spell Level
	1	2	3	4	5	6
1	33
2	83	17
3	98	62	9
4	99	90	44	5
5	100	98	78	30	3
6	100	99	93	63	21	2
7	100	99+	99	86	50	14
8	100	99+	99+	96	76	38
9	100	99+	99+	99	91	65
10	100	100	99+	99+	97	84
11	100	100	99+	99+	99+	94
12	100	100	99+	99+	99+	98

Alright, so problems begin to creep up pretty early on. If you're replacing Vancian memorization with a chance of failure, you need a pretty good chance of failure. But Spell Level x 5 simply isn't steep enough a curve. Magicians quickly become unstoppable firehoses of magic.

But if you change the Target numbers like so:

Spell Level	Target
1	5
2	15
3	25
4	35
5	45
6	55

You get a success rate that is much more in keeping with the Old School progression:

MD Pool	Percent chance of succeeding per Spell Level
	1	2	3	4	5	6
1	33
2	83
3	98	9
4	99+	44
5	100	78	3
6	100	94	21	<1
7	100	99	50	1
8	100	99+	76	9	<1
9	100	99+	91	28	<1
10	100	99+	97	53	4	<1
11	100	99+	99+	76	15	<1
12	100	99+	99+	90	34	2

You can fine tune the Target numbers to more closely follow the Oe/BX progression (something along the lines of 5, 12, 16, 20...), but I was drawn to the idea of a numerical sequence that could be easily memorized.

This mostly looks good to me. A 1MD wizard is going to miss a lot, but will be pretty decent with 1st Level spells once they advance to 2MD. 

A 3MD wizard is going to be a dab hand at 1st Level, but will need to fail at least once to have a decent shot at a 2nd Level spell.

It's not perfect, by any means. You'd have to be level 15 before you could reliably cast a sixth Level spell. Maybe drop the Target for 6th Level spells to 50? Breaks the pattern, but gives a 12th-Level caster a 1-in-10 chance of succeeding.

Also, the dynamics of adding and dropping dice based on failure and success probably has all sorts of implications that need to be play-tested. F'rinstance, what do you do when a player purposely fires off higher level spells in order to bank enough failures and build up their MD pool?

Possible solutions include limiting the number of dice you can carry forward, limit the duration, or instill a calamity/corruption risk to casting above your level.

Howsabout this? Succeed or fail, if your roll results in any 1s, you must roll below:

1d8 + # of 1s	Calamity
1-4	No effect.
5	Inconvenient side-effect. GM picks something about the spell effect to go wrong.
6	Addled. Lose 1d4 INT or WIS. for 1d6 Turns.
7	Exhausted. Lose 1d4 STR, DEX, or CON for 1d6 Hours.
8	Spell Forgotten. Will need to be relearned.
9	Temporary mutation. Lose 1d8 CHA for 1d6 Days. GM makes up your crazy new appearance.
10	Permanent mutation. Roll on some chart or another.
11+	Extra-planar intrusion. You summoned something troublesome.

There's a lot to work out, here, and I present this information in the hopes of continuing the conversation. This could be quite a pretty little system.