D&D 5E do CRs seem a bit arbitrary?

[SirAntoine]

Challenge ratings don't work very well in D&D, where anything can happen. At best, they offer some guidance, but if it takes a long time to work with them that benefit is overshadowed in my opinion. I prefer to carefully read the monsters' descriptions and pick monsters which would really give the PC's the level of challenge I want for any given encounter. This can take a little time, too, but much less I think.

 

[FormerlyHemlock]

My next concern is that I will will build what feels like an appropriate encounter for the party, and then look at the total XP and realize it is way past lethal and have to start toning it down.

Why tone it down? All the best encounters are Deadly x2 or better. I use lots of encounters that turn our to be Deadly if you count up the XP. My guys aren't even powergamers and it still works.

 

[I'm A Banana]

And yet 4e's encounter budget system worked perfectly fine, and provided a highly robust system. You couldn't just throw special abilities on willy-nilly, but as far as stats and level were concerned, the budget worked more or less exactly as advertised.

The price we paid for that tight budget was a decrease in the meaningful variation of options and a system that resisted dramatic swings in fate. Given the amount of complaints about 4e's "sameness" and its "draggy combat," it's pretty safe to say that not everyone was on board with paying that price.

Personally, I think CR does a fine job, because I expect my D&D combats to sometimes be wildly swingy. Sometimes the party cleric will roll good on turn undead and turn the zombie fight into a cakewalk, and that's fine, because it lets them be awesome for a moment. Sometimes, they'll roll low and save-or-suck effects will lead to the wizard's untimely demise. That's fine, because it makes the risk of failure very real. Greater precision in 4e didn't lead to a more fun game for me - in fact, I pursued measures fairly early on that restored the swinginess to the system for my own games.

 

[Roger]

A bit, but let's consider what we've got here.

Let's start with a Grick. 14 AC, 27 HP, potential to do 13 damage with two +4 to hit attacks. Probably get a surprise round, which grants it advantage to hit once, but otherwise something I'd call a very moderate challenge to a level 2 party.

Grick has damage resistance to nonmagical weapons, which isn't quite as good as twice as many hit points, but it's sort of in that ballpark. And it has a weaker version of multiattack -- instead of just making two attacks, it has to hit with the first one to get a chance at the second.

Move next to a Gnoll Pack Lord, fighting alone. Now we've got 15 AC and 49 HP, with the chance to do 16 damage in two attacks. Tough, and definitely more dangerous, but still the chance of PC death is quite low.

As you say.

Now the Polar/Cave Bear: only 12 AC and 49 HP, but +7 to hit on two attacks for a total of 21 damage each round. It's got a good chance to hit and a good chance to drop a character each round, and 49 HP is easily 2+ rounds for an average party, even focus-fired. Still not a likely TPK but the damage is way up and it's still a CR 2.

Hits harder than the gnoll or grick, but also has lower AC and lower HP than those monsters, so it looks to me to be in the right ballpark. It's also just a beast, so lots of druid/ranger stuff is a possibility, so yeah, maybe on the harder side of CR2, but maybe not.

Ghast - not too tough with 13 AC and 36 HP, but turn defense means clerics no likey and melee will have a bad time with it - save or disadvantage to hit. The main thing is its +5 hit, 10 damage and DC 10 Con save or paralyzed - which means next round you are down (coup-de-grace bite for 19 damage). Sure it takes a couple rounds and a failed save, but the chance of a PC dying goes up.

A lot of things need to go wrong for the hapless character here -- they have to get hit at +5, then fail the Con save, then get hit at +3 with advantage, and then after all that they take 19 (more) damage. Which isn't a good day, but not every day in the front line is.

But what about the Gibbering Mouther? It's a first-round danger: DC 13 Dex save or be blind in an AoE, which gives it advantage to hit you for a whopping 17 damage. That means someone is very likely to go down in one round - which combos in a nasty way with its "DC 10 Wis save or be useless" power that happens EVERY ROUND to EVERYONE. So there's a decent chance you'll helplessly watch your downed buddy bleed out while you hit your friend or stand in a corner, especially since it has 67 HP to chew through. Chance of death goes way up.

Well, yeah, but it's also 67 HP that has an AC of 9 and a move of 10'. So it might eat your blind gibbering buddy, or you might just all stand back and fill it full of arrows and have a good laugh about it later.

Now check the Gelatinous Cube. Very good chance for a surprise round, where it's a save with disadvantage or get sucked in and take 10 damage. If you fail another save, you're down, no question: 21 auto-damage. No one is going to stay up from 31 damage - most level 4 PCs won't. And if you stay in another round, you're dead completely from the 21 auto-damage again. And your buddies can try to help with the 2nd save but it's 10 auto-damage to anyone who does - that is ROUGH. But at least it's two saves, and they are DC 12 Dex then Str - most will at least have a good shot at one of those. And the thing has auto-hit AC. But 84 HP means it will definitely get many turns, and considering it is large there's no reason it can't do this to multiple PCs at a time, possibly as many as four. Quite a bit more scary - still CR 2. There is a good chance multiple PCs could get killed.

No doubt about it -- any day you walk into a gelatinous cube, that's going to be a bad day. For the poor sap walking point, anyway. Everyone else can point and laugh from a safe distance.

And then you have crazy combo CR 2s, like the Giant Elk.

You had your chance to learn Giant Elk so you could explain to them why they shouldn't be stepping on your face. It's hardly the elks' fault you kept falling asleep during Introduction to Giant Elk Love Poetry class.

So do the CRs seem a bit arbitrary?

Yeah, sure, a bit. It might be useful to have another number to reflect the variability of the monster -- if CR is the average challenge, some way of expressing the standard deviation might be useful to some DMs. But it might be more hassle than its worth.



Cheers,
Roger

 

[evilbob]

Grick has damage resistance to nonmagical weapons

Wow, I completely missed that. You're right: this bumps the Grick up to be much closer to the other bruisers. Thanks!

Well, yeah, but it's also 67 HP that has an AC of 9 and a move of 10'. So it might eat your blind gibbering buddy, or you might just all stand back and fill it full of arrows and have a good laugh about it later.

Definitely ranged is the way to go with this guy - it's just amazing how bad it can be if it actually gets close to you, especially since you may not even be able to run away. I guess its real saving grace is that it is impossible for the thing to surprise you - it always starts babbling when it can see you - so other than the poor fool who charges in before getting those knowledge checks, I guess this does balance it out quite a bit.

Oozes are also much safer from a distance, lacking any ranged abilities and being particularly slow. Maybe that's why the Gelatinous Cube is rated so low, relative to its power - if you have a 200' hallway to backtrack into, it's just a pin cushion (it can't even detect you over 60' away). I'd still call that thing a CR 3, though; it's still stronger than an owlbear and an owlbear also isn't known for its archery skills.

So in the end, it seems like I was probably mistaken; the power level is closer than I thought, other than the one outlier who may need to be CR 3 - but that also has no ranged capabilities and therefore is even more dependent on its environment to pose a threat (although the chance of PC death is still relatively high). I guess "CR 2" sort of is shorthand for the normal bruiser-level you'd expect, plus the start of several "mess up and die" creatures where several bad rolls or a particularly nasty battle location can really get you.

 

[CapnZapp]

I think CR is a terrible system. D&D has too many moving parts for any sort of consistency. They should just divide up the enemies into 3 or 4 tiers or something. All the steps taken to determine CR is just more trouble than it's worth IMO. I like how numenera / cypher system enemies are made. You get one number, and bam. Done.

+1!

 

[Tormyr]

[MENTION=15700]Sacrosanct[/MENTION]: Yeah, that's weird - Kamikaze Midget said that. I probably just messed something up when trying to reply too quickly - sorry!

There is a bug on the boards when you use the multiquote button in a thread. Any regular quotes you do later in the thread include the original multiquote as well. When you cut out the parts you don't want to quote, it is easy to match up a user name with a quote they did not make.

 

[Tormyr]

I have reverse engineered several of the MM monsters to get a better understanding of how certain elements of the guidelines are calculated. This was especially helpful for figuring out DPR on spellcasters with AoE attacks. I am able to get almost all of them to the CR in the MM. The rest are off by a step (1 vs. 2, 1/4 vs. 1/2, etc.).

My experience has been that not only do the CR calculation guidelines pan out, but the encounter building guidelines do as well. As I have been converting Age of Worms, I have converted dozens of monsters that are not in the 5e MM. After that, I built the encounters to mimic the Encounter Level of the 3.5 encounters. If I had a level 10 party in what would be an EL 13 encounter, I make what would be a medium encounter for a level 13 party. This then turns into a Hard encounter for the level 10 party. Since the DMG came out, the encounters have gone very predictably in terms of difficulty per encounter.

The process of determining monster CR and encounter difficulty is very easy for me as I have a couple of spreadsheets for the purpose. They allow me to quickly tweak a monster or encounter to get the desired outcome. You can find them by clicking my user name and then clicking the link for the downloads I have posted.

 

[FormerlyHemlock]

Oozes are also much safer from a distance, lacking any ranged abilities and being particularly slow. Maybe that's why the Gelatinous Cube is rated so low, relative to its power - if you have a 200' hallway to backtrack into, it's just a pin cushion (it can't even detect you over 60' away). I'd still call that thing a CR 3, though; it's still stronger than an owlbear and an owlbear also isn't known for its archery skills.

The problem with the CR system as a whole is that it's basically just a way of measuring the 3/2 power of HP and damage scaling. You could bump a Gibbering Mouther up to CR 10 (and triple its HP and DPR in the process) and it wouldn't change anything substantial about its threat profile, it would just give more XP to whoever kills it to death with arrows.

In other words, the problem with CR is that it's a scalar and threat profiles are vectors. Since I don't care to do the work to convert CR to a vector, it's simpler just to use it for what it is: a fairly arbitrary way of calculating XP rewards, and otherwise best ignored.

 

[FormerlyHemlock]

My experience has been that not only do the CR calculation guidelines pan out, but the encounter building guidelines do as well. As I have been converting Age of Worms, I have converted dozens of monsters that are not in the 5e MM. After that, I built the encounters to mimic the Encounter Level of the 3.5 encounters. If I had a level 10 party in what would be an EL 13 encounter, I make what would be a medium encounter for a level 13 party. This then turns into a Hard encounter for the level 10 party. Since the DMG came out, the encounters have gone very predictably in terms of difficulty per encounter.

That's great! It means your players are right in the sweet spot for CR: they play encounters in the way that the designers expected most players to play. In your case, CR is working exactly as it was designed, and that really is great. It won't apply to all groups, but it's nice that it does for you.