Playtesting D&D monsters

I've been running quite a lot of playtests recently, so thought I'd talk about a few of the ins and outs of the process.

This is through a D&D 5e lens, but I suspect it generalises.

If you're planning to publish a monster, thorough playtesting is a must. It's not, sadly, a case of "run it once for your regular group's entertainment". You must run it until you have complete confidence in your work.

There are lots of design elements that you're going to need to juggle your focus on in playtesting, but we can group them into three main things:

1. Double-checking the maths
2. Ensuring it's fun
3. Watching for corner cases

Double-checking the maths

If you meticulously follow the D&D Challenge Rating rules in the DMG, you get a perfectly serviceable CR... As long as you can draw strong parallels between the monster's traits and those for which the book provides appropriate adjustments. For a trait that's really off the wall and you're not sure how large an effect it might have, playtesting becomes that much more crucial.

Particular things to keep an eye on:

1. Monsters designed for very low or high CR. Low-level D&D characters are more vulnerable to individual dice outcomes; high-level characters can take advantage of synergies and have more options.
2. Huge damage swings. A monster with a magic item that rolls 1d20 for 19 different cantrip effects and a ninth-level spell on a 20 might be balanced at a low-middle CR according to the mathematical averages. In practise, though, you're going to have one of two different unsatisfying play experiences.
3. Offensive/defensive balance. As well as a monster being 'too easy' or 'too hard', being 'too arduous' to fight is also a design failure. D&D 5e (a) makes combat front-and-centre, (b) has a lot of fiddly rules for it, and (c) disincentivises players planning their turn before taking it. This adds up to a big risk of combat encounters dragging on. It's generally (although not always) safer to tweak the OCR higher and the DCR lower.
4. Character optimisation. Creatures with book values are 'safe'. Starting players will on average have a good time fighting them, won't typically struggle at all with a level-appropriate Medium encounter, etc. Optimisers will take the same monsters and run them through the blender in a single round. There's no silver bullet here. In fact, there's no bullet. You simply have to balance for the 'default case' and then expect DMs to make changes to encounter design (or individual monsters) as necessary.

Ensuring it's fun

Monsters can be mathematically sound but boring to play. It's easy to come up with examples. Imagine a beast that negates a huge slew of character powers but can be readily whittled down with a steady barrage of normal attacks, while attacking once per round itself. Or one that hedges a specific character archetype out of the fight, while being unable to harm such characters. Or one that can only be engaged at range.

Judging the fun is the subjective bit of playtesting, and as hours and hours of combat turns pass, it gets harder and harder. Break up the work. Roll the dice instead of using average damage, just to see how it feels. Work out in advance how you're going to track player feedback on what feels satisfying and what doesn't.

Hopefully, you'll find sudden joy in stuff you didn't expect, and that will inform your work as a designer. The other day I tested a weak undead horde creature that I'd given an 'overkill' trait: a good enough roll lets the attacker kill it and cleave through into an adjacent creature of the same type. It turns out that kind of outcome, mowing down chaff with a single shot due to it being such a visibly fragile enemy, feels great for low-level characters who lack many special abilities for area damage!

Watching for corner cases

There'll be all sorts of things you didn't think of while designing. This creature is inorganic; it should be immune to poison. This creature appears in groups; it would be better off without the multiattack.

This creature's special power doesn't actually work as described... or lets it do this cheesy thing... or perhaps more commonly, it works fine, but you didn't describe it clearly enough for a DM to use. Get other eyes on your work as soon as possible (but that's a whole other article).

You won't find the problems if you don't actually examine the monster interacting with the existing rules, in a 'real' situation.

Playtesting is a big task...

...but it doesn't have to be a chore. Be organised, have a plan going in, intersperse it with other kinds of work, take joy from the emergent consequences as you would from a real game, and then the repetitiveness of it won't kill off all your fun.