Let's Talk Crits!

 One of- scratch that, the most controversial element of the Character Origins UA (I think we're just going to keep talking about this until the next comes out, sorry) is contained within the D20 Test segment of the rules glossary.

Let's talk about the simplest change:

In 5e, natural 20s and natural 1s only mean anything special when making an attack. If you're some high-level paladin with a +14 to hit with your sword, and you're attacking a lowly zombie - some minion of the lich boss you're trying to defeat - who has an AC of 8, getting a natural 1 means that even though your total to-hit roll is 15, it still misses.

The benefit here is that it gives the roll a little unpredictability. Yes, with that hit bonus and the monster's very low AC, the chance of you missing is tiny, but this gives you the opportunity for a moment to screw up and miss despite how grand and powerful you are.

I think the opposite scenario is far less likely - the highest AC of any monster in the game is 25 (though Sul Khatesh can cast shield on top of a base AC of 22, putting hers at potentially 27). Most characters have a +5 to hit with their main weapons or spells at level 1, so a natural 20 even at that low level would still hit a Tarrasque anyway without the "crits always hit" rule. But in the unlikely event that you wouldn't (perhaps you're a wizard caught in an anti-magic zone and need to use a light crossbow or something with a fairly low Dexterity score,) this allows you that one-in-a-million (ok, one-in-twenty) shot, which can be a fun, dramatic moment.

Anyway, I think most players are pretty happy with the crit mechanics for attacks.

What the UA changes is that this effect now applies to the other d20 tests - ability checks and saving throws.

I want to tackle saving throws just to get this out of the way:

I don't think I've seen anyone complaining about the potential to crit success or crit fail on a saving throw. The saving throw is something that, in-game, has very clear mechanical consequences for success or failure. Much like an attack, a saving throw's binary outcomes are usually to be expected one way or the other, and it's not disruptive to have a success or failure.

The difference that this change would make is that it would allow, rarely, for people to succeed on saves against very powerful foes. Some monsters have saving throw DCs that go above 20. For example, a Black Greatwyrm's save DC against their breath weapon is 26. Without stacking things like Aura of Protection, Bless, and other bonuses, I'd say that even most level 20 characters would automatically fail against that. You actually need proficiency in Dexterity saves or some way to artificially boost your Dexterity above 20 in order to hit that even on a natural 20.

As the stat block reads right now, the DM simply gets to do an average of 78 damage to players in the cone's area. This change would give players a chance - just a tiny, 5% chance - to reduce that damage to 39.

The one scenario I can imagine this being a problem is if a DM wants to make sure players fail against, say, a poison that knocks them unconscious to set up a scenario where, for example, they have to escape from some prison. I think this is a situation where DMs need to simply make use of "cutscene magic" to say that something happens, despite the efforts of the players. DMs should also talk with players ahead of time to make sure that the players are ok with momentary loss of control like that.

The controversy arises with ability checks.

Ability checks' are the broader, catch-all d20 tests that handle things usually outside of combat and regimented, initiative-based play. This is usually where players get creative, using their skills in unexpected ways.

The classic scenario here is "can I seduce the big bad?" Your party's bard might have a +5 to Charisma, expertise in Persuasion, and thus expects that they can get a super-high number on a roll. And then they nat 20, so that's an automatic success, right? So that means the bad guy is now going to do whatever they say, right?

Unlike attack rolls and saving throws, players have a tendency to roll these unprompted, and they often like to set their own parameters for success or failure. While they'll still respect the DM to set a DC, they expect that a success means they get to do what they intend.

And that's the problem.

And I think it's as much a cultural problem as a rules problem. Ideally, in-game, players would say something like "I want to try to seduce this evil necromancer," and it would be up for the DM to ask for an ability check - or to not ask for one.

The DM could respond in a couple ways. "No, the necromancer does not seem responsive to your attempts to arouse their interest." "Make an insight check," and then if they succeed, they'll realize that the necromancer is too committed to their plan to entertain such thoughts, or maybe they're not interested in the player character's gender, or perhaps they're already in a really fulfilling, committed relationship and don't want to screw that up.

The problem here arises when the player just rolls the check without prompting and then expects the DM to respect the result.

On the DM side of things, you also need to know when not to allow a roll - the guidance suggests that if a roll would require a DC of less than 5 or more than 30, you shouldn't bother - less than 5 should succeed without a roll and over 30 should be considered impossible. Don't make someone roll for something if you don't want to proceed with either outcome.

Now, all that being said, one of the philosophies behind One D&D changes - and I think very much including this one - is to try to meet players where they are, and reflect the way people actually play. And that practice of players calling out their ability checks is one of those ways people just play.

In that case, I can understand why a lot of DMs want to restrict this rule. I think that the rule can stay if we just make sure that DMs are the ones who determine if an ability check is warranted.

But we're not done!

The other change, and one that I think has more people up in arms, is the change to critical hits.

There are three main aspects of this change - all which reduce the power and presence of crits.

The first is the limit on what crits: only weapon attacks and unarmed strikes would crit under this rule.

Let's unpack that: this means that attack-roll spells like Guiding Bolt or Chromatic Orb would no longer benefit from a crit upon a natural 20. This would also apply to a lot of cantrips like Fire Bolt or Chill Touch.

Now, I get that spell crits can be insane - Guiding Bolt is a 1st level spell, but on a crit, it does 8d6 damage - the equivalent of a fireball (though only to one target). The thing is, most leveled spells require not an attack roll, but a saving throw on the part of the enemies. Fireball, for example, requires a dexterity save, and even if the target succeeds, it winds up taking half damage. Thus, these saving-throw-based damage spells are often "safe." Even if you get unlucky and all the enemies beat your save DC, you still did something useful in combat.

The risk, then, with an attack-based spell is that it can be totally wasted. That risk should come with a reward, and I think that award is easily implemented with a critical hit.

The next aspect of the change is that the critical hit only affects weapon (or unarmed strike) damage. Note here that unarmed strikes only benefit from a crit if you have some feature like Tavern Brawler or Martial Arts to give yourself a damage die. But this means that any additional damage on top of your attack would no longer benefit from a critical hit.

The two classes I see getting hit the hardest by this change are Paladins and Rogues (though I think Rangers who use Hunter's Mark would also feel a bit cheated).

Flashing back to when I was a new DM and we had a paladin in my group. When he rolled a critical hit, I was flabbergasted that he got to see the roll before he decided to use Divine Smite. Thus, a conservative paladin could hold spell slots for those moments to do an absolutely devastating amount of damage. It was an eye-opening experience for me as a DM. Later, in Curse of Strahd, I'd play a paladin and I will tell you this: there is no rush in D&D that I've felt that is more exciting than smiting on a crit.

Rogues, of course, get most of their damage from Sneak Attack. Their actual weapon damage is fairly negligible - using a 1d4 dagger versus a 1d6 shortsword makes very little difference when you're doing 5d6 damage on top of it either way. Rogues want to crit even more than paladins, and indeed, basically the only reason to play an Assassin rogue is to get those guaranteed crits against Surprised targets - it's the subclass built for taking out an enemy before the fight even starts.

Now, I'll grant that the intention here might have been to try to rein in some of the higher-damage classes to be more in keeping with other martial classes (though the Great Weapon Battle Master Fighter in my Ravnica game needs no help at all doing damage - on a recent boss he did something like 89 damage in a single turn.) Still, this feels like the least fun way to do so.

The final proposed change is that only player characters can crit anymore. Monsters would be unable to do so. First off, this clearly makes adamantine armor and the Grave Cleric's Sentinel at Death's Door basically useless. On the other hand, I have had experiences at low levels when an errant crit auto-killed a player character. (In one scenario, a character introduced about 5 minutes earlier was torn apart by a swarm of Gremishkas before they had a single turn in combat - hilarious, but not exactly ideal in a scenario where it was a character that the player was invested in). However, the most consistent mark of character growth as they level up is higher HP, and pretty soon it becomes pretty unlikely that a monster can one-shot a character with an attack, so I think the threat of crits comes down to a reasonable level.

The proposal that Jeremy Crawford mentioned in the hour-long interview introducing this UA was that monsters will have more recharge abilities, allowing that, rather than crits, to be the "big scary ability" that they can use. Because a monster doesn't have to use that ability, this gives the DM a little more flexibility to tune the fight on the go (something any veteran DM learns to do).

For the most part I only really think this solves a problem at the lowest levels of play - even by level 3, player characters are resilient enough to take the odd crit, as long as it's not from some way-too-high-CR monster.

So, I'd say the opinions I've landed on are the following:

Crit success and failure on saving throws: no problem, totally fine.

Crit success and failure on ability checks: fine as long as players understand that the DM gets to decide when an ability check happens.

Inspiration on a Nat 20: We didn't talk about it here, but yeah, it's great and I love it. Already made it a thing in my game.

Only weapons critically hit: Not really into it. Seems like an unfair nerf.

Only the weapon damage crits: This is a big fun-killer. Big no.

Monsters don't crit: Mostly no. I might stealthily apply this for a level 1 or 2 party, but even then, probably not.