Having now performed some playtesting, I'll say that this chart worked really well. It's hard to say that and not sound biased, of course, but here's how it played out in my game.
- As recommended, we only used it for the PCs, so it barely slowed down play. (Only 5% more rolling than auto-max-damage crits, and the same amount of rolling as roll-to-confirm.)
- It stayed behind the DM screen, so it created fun surprises for the PCs and left interpretation in the DM's hands.
- I treated the flavor suggestions (which body part is struck) as exactly that, and adjusting the flavor on the fly was no problem.
- Our opponents were mostly humanoid and low-HP, so the flavor suggestions were mostly relevant. (*But see below.)
- To fit 5e better, a couple of the wizard crits would have to be tweaked. (**See below.)
*When the PCs encountered a water elemental, I was tempted to say crits were just impossible. Instead, I treated the one crit that happened as an auto max damage. Unless magical damage types are involved, it's hard for me to imagine what sort of attacks could produce effects like slowing or tripping.
**In our first session there was actually only one crit by the PCs: a wizard using shocking grasp. The player followed his 20 with another 20 roll. The result (not spending the spell slot) was irrelevant to an at-will spell, which I thought was way too anticlimactic. So I allowed him to regain the last spent spell he'd cast. My description of this was a bit colorful (the shocking grasp flaring like an electrical storm and the enemy's soul being drawn into the wizard's hand) and became a defining moment for the character.
I'll say this: playing two sessions with a crit chart was more fun than the previous session I spent playing without one. Maybe it will get old, certainly there are more simulationist systems out there (like allowing "called shots" from a pre-defined chart on a successful crit), but for fun factor I think crit charts can be great.
I do feel that some people are overthinking the whole crit chart thing (like any random table, it's best taken with a grain of salt) but clearly people's mileage varies significantly on this topic. Perhaps so much so that it's best left in the realm of third-party supplements rather than rules modules.
