What attributes used to exist?
- Base Attack Bonus (hit rating, weapons only)
- Individual save bonuses
- AC, but not really an attribute
- Caster level, for purposes of damage, SR, occasional level contests
- Spell save DC
A strong case could be made for an Attack and a Magic bonus, where 1/level of each is the "prime" amount. Generally, nobody should get both at 1/level.
What about Defense? I mean, the alternative is individual save bonuses (Fort/Ref/Will, or even all 6 stats, plus perhaps AC, or even multiple ACs). Could get weird either way.
Meanwhile, there are a lot of other things that attributes might improve, such as:
- crit chance / crit power
- initiative
- mobility
- haste
This of course gets back to the classic debate about what really is the point of proficiency bonuses. They separate one PC from another, saying "you can do X, but you're bad at Y". This idea has pros and cons. It seems exclusionary for a rule's only purpose being to limit what each PC can do. However, it also seems weird to play in a system where everyone is equally good at everything. The ideal situation seems to be that everyone has different stats, and the math works out that everyone is roughly equally powerful, but with strengths in different areas.
To achieve this, it is vital that prime rolls--attack rolls for fighters, spellcasting for casters, etc, must always be at a prime, level-based proficiency bonus. Being below this line has absolutely no benefit to the PC, and being above it trivializes the game. Smarter games like WoW realized this long ago and shifted focus away from prime stats to secondary stats.
But as long as we're using a system ultimately derived from d20, we should probably stick with the basic formula: 1d20 + skill mod (i.e. proficiency) + stat bonus.
Of course, proficiency is secretly two things: level and build. Well, it is in 3e and PF2--not so much in 5e. How can we compare these options?
- 5e is simple. That's great. It's neutral for multiclassing, as it just factors out.
- PF2 (and 8.2, and various similar systems that offer a few simple ranks of proficiency) are bad for multiclassing, and not as simple as 5e.
- 3e style is best for multiclassing. Each individual level gives you a specific bonus to each attribute--even if it's a fraction. It's not simple, per se, although the math is quite trivial, and exceptionally so for a computer. Also, it doesn't require choices over and above "what class do you want?"
- Similarly, a fully skill-based system would be fine, if classes didn't exist, although the attributes would have to be well-chosen to have enough flavor for anyone to care about which one to choose.
Okay, but which attributes should exist? Recent experiments suggest the following:
- Not too many! How many is too many? Not sure. 10 is too many. 3 may be too few, but maybe it isn't.
- Some should be okay to "dump" for certain builds. None should be universal dumps.
How about this?
- Attack
- Speed
- Toughness
- Skill
- Magic
- Willpower
Rougly analagous to the 6 core stats, but not exactly. Here's what they do:
| Attribute | What it does |
|---|---|
| Attack | Adds to hit and hit power with weapons |
| Speed | Adds to movement speed, initiative; improves Reflex saves |
| Toughness | Adds to HP; improves Fort saves |
| Skill | Adds to crit and crit power; adds to skill checks |
| Magic | Adds to spellcasting checks, DC setting for spells |
| Willpower | Adds to spell damage; improves Will saves |
All have the same schedule, where:
- 1/level is the best
- 0.75/level is mid
- 0.5/level is the worst
The math gets more complex for some derived stats, such as:
- Weapon damage goes up one die per 5 Attack
- Magic damage goes up one die per 5 Willpower
- Max power rank is 1 per 5 Magic
- Movement speed increases by 1 square per 3 Speed
- HP: tbd
- Crit chance goes up by 1 per 2 Skill
- Crit power goes up by 50% per 5 Skill
Also, I think what is currently called "Magic" should refer more to magical endurance. Willpower might control damage dice in addition to da