Son of a
Hi,
In classic message board tradition, I've just lost thirty minutes of composed text that I was pretty happy with. So now I'll have to be represented in this thread by this half-assed reply. Reason #2 why I usually stay out of these discussions.
A CR is a broad value because its unit of measure is larger than that of the subject being measured. A +1 BAB (or save or ability) doesn't equal a +1 CR. Several factors have to accumulate before a CR gets bumped to the next order of magnitude. What's more, CRs describe a condition which can only be met after random factors are applied. Water doesn't boil at 1d20 + 90 degrees, by example. So, any given CR must be wide enough to encompass all the possible factors. The system is even so broad that CR 1 is not an acceptable low, so we have CRs 1/2, 1/6, etc. A precise measurement could be crafted with fractional CRs (e.g. CR 3.75, but let's not do this). That's what I mean when I say "broad."
I'm not saying that I think a mathematical formula for CR isn't possible, I'm saying I don't think it's necessarily a good idea. I'm concerned that a stricter method would be used as a tool for exclusion. CRs can be used right now as a gauge of a designer's fluency in the system (Wulf, you implied you had publishers in mind), and I have a slight fear that a CR formula would become a systematic method for grading a designer's work; an artificial scale without concern for design choices (by designers) and personal preference (in consumers). I'm concerned that a CR formula would become a compeller, emphasizing statistics and metagame over flavor and the RPG as a whole.
Now, Mearls' method seems more like an explanation or study of the existing system. It seems designed to quantify the concepts considered when assigning a CR rather than creating a mathematical construct to calculate a CR. This speaks to DM education about CRs, which is something I supported in my first post above. I'd be very interested in a method which is empowering without being restricting. IMO, such a method should be accompanied by essays on theory and application, if its going to be presented to end users and not just other developers. The Monster's Handbook contains just such essays, so maybe I have nothing to fear.
To be a pedant, I'll mention that my distrust of a CR formula probably comes from the terminology. A formula is formative (of course), and I think the job of forming a monster should fall to the designer, who decides if a given monster should be built first as a statistical entity or a narrative one.
As an aside: undead certainly can be an uncommon menace to even an average PC group if that groups lacks certain "average" characteristics (most involving clerics). Undead creatures usually have traits to counter-balance their Constitution (AC, which has a wider scale of random factors than weapon damage, is a more powerful defense mechanism than a volume of hps since randomness favors monsters) and BAB (ability scores and touch attacks being common). Although undead are designed to be suspeptible to a variety of binary factors (I used to call them "toggles," but I'll borrow the term here), if those factors are not present then the undead become more formidable. Often, undead can initiate binary contests, too, which emphasizes dangerous randomess at low levels. It's CR does assume certain things about party make-up, after all, but gets by with averages. Since no one group is "average," though, and most encounters are met by a given player group just once with one set of variables (the PCs), different groups can experience very different results with the same CR.
This is a reason why high-level CRs can be troublesome, I think. High-level combats involve a great many binary conditions (related to saving throws), which is what keeps them moving at a pace roughly equivalent with low-level combats, despite hundreds of hit points and dozens of attacks. However, binary conditions in d20 can put more emphasis on the "sure thing" over the die roll (a +18 attack bonus dilutes the impact of the d20), which empowers PCs over NPCs. It's possible for a single loss in a binary contest to result in defeat conditions for a CR 20 monster. Was that 1 round of combat a suitable challenge, then? Randomness makes things a bit hinkey sometimes.
The math isn't my strong suit. I prefer to keep the pretty clothes on the game mechanics to preserve my escapism. Do you find that's a problem for your, Wulf or Mike?
word,
Will
