D&D 5.5e – they’re trying to call it “One D&D” but they tried to call 5e “D&D Next Edition” and I’m not playing along with the new dumb name anymore than I am with the old one – is mostly the same as 5e, but the beta test does show a couple of significant rules updates. One of them is adding in critical success and failure for skill checks: On a natural 20, you always succeed, on a natural 1, you always fail. This is a bad idea.
The reason why this idea gets support mostly seems to come from greentext stories about rolling a natural 20 and causing something insane to happen – the Bard seducing the BBEG mid-fight or a dwarf being too stubborn to fall in a pit or whatever. These greentext stories make great memes because they’re funny enough to be worth the two minute time investment it takes to read them. For some fraction of the playerbase, the point of D&D is basically to create an environment that generates these greentext stories. But that’s not a very big fraction. Most people balk at the idea of spending four hours playing a game in the hopes that a two-minute joke will fall out that they can post to Reddit.
And for everyone who plays D&D to tell a full-length story at the table, rather than to boil it all down to one funny moment posted to Reddit, critical success and failure both have negative effects on the flexibility of the skill system. The possibility of critical failure means that no one, no matter how high their bonus is, will ever reach a point of being able to do extraordinary things as a matter of routine. A character with a +5 ability bonus, +6 proficiency, and another +6 from expertise cannot roll lower than an 18, which means DC 15 checks succeed automatically. Removing the tension from die rolls is usually bad (abilities that do things like allow you to reroll any roll below a 10 are usually bad), but the massive investment of character resources required to reach the point where a DC 15 check is automatic means that it’s not a common thing. It’s a high-level superpower, which arises emergently from the game’s systems. You can kludge that back in with a special 17th-level ability for Bards that grants them autosuccess on DC 15 skill checks for any skill they have expertise in (or whatever), but it’s much easier to just let it happen as an emergent property of the skill system, that way you don’t have to write another rule into the books to cover it.
And critical success is, counterintuitively, very disempowering for players. Part of the GM’s job is determining what’s even possible to accomplish with a skill check. Under the current system, skill checks can be capable of achieving some pretty spectacular results on a DC 30, 35, or 40 (the latter in particular achievable only by stacking buffs on someone who’s also got a very high passive bonus). You can make it possible to do things like climb rain up into the clouds or convince a guard that he’s been polymorphed into a wombat, because you can assign these things an extremely high DC, so they’re available only to very powerful characters.
But some things might just be flat-out impossible. No amount of Arcana lets you cast spells without a slot, no amount of Perception lets you observe things from ten miles away, no amount of Athletics allows you to strike the ground and cause an earthquake that levels a city block. Which of these things needs to be impossible no matter what you roll and which are possible but only at extremely high DCs is left up to GMs – and if a character always succeeds on a natural 20, a lot more stuff is going to be moved over into the “impossible no matter what your bonus is” category. Most people’s D&D stories have room for legendary Barbarians climbing rain into the clouds, but not for a random farmer who tried it every time it rained and got lucky after a month or two.
Critical successes and failures push D&D towards slapstick, and while that’s definitely a style of play, it’s also definitely not the primary style of play, and the default rules shouldn’t be supporting it at the expense of the standard high fantasy pulp-not-parody tone that most people play in. And if you’re creating optional rules to encourage a wakcy playstyle, critical successes and failures are a good place to start, but don’t go nearly far enough on their own.