D&D 5E Fixing Challenge Rating

[Quickleaf]

I really feel that CR as it is also strips a lot of magic from monsters. Trying to so rigidly design them IMO makes the designs off feel uninspired, and makes the danger of something feel a lot more transparent. Since the point of bounded accuracy is to make it so that you can challenge higher level players with lower CR monsters, I feel like the current implementation makes it so those lower CR monsters really just don't feel like a true threat -- more an annoyance as you're trying to deal with the cool real danger in the fight.

I actually think that more robust, holistic, and transparent mathematics facilitate making more creative monster designs.

I agree with what you're observing – that 5e's monster design presents a "sameness" trap that it is easy for a designer to fall into, and even more so if the designer prioritizes creating a monster whose CR can be clearly modeled under the existing guidelines.

One of my projects involves lots of really wild / creative monster design within the 5e umbrella – some is stuff that has no parallels in the game. Evaluating those divergent/creative ideas from a CR standpoint with the current system? Hah. It's educated guesswork at best, nightmare at worst.

I understand wanting to push back against an unfortunate trend in online discussions to hyperfocus on this one number (CR) at the expense of pushing the creative frontiers of the design – heck, I hate that tendency more than most folks, and push against it with my own stuff – but I think it's inaccurate to assume that Maths & Creativity are inherently opposed forces such that developing Maths must come at the expense of Creativity. Nothing about that is inherent.

 

[Maxperson]

Based on the conversation in this thread, Scott, Teos, and I agreed to add the Lazy Encounter Benchmark chapter from Forge of Foes into the Creative Commons as well. You can find our full CC BY 4.0 Lazy GM's 5e Monster Builder Resource Document here:

The Lazy GM's 5e Monster Builder Resource Document

Lazy GM guidelines for building and running 5th edition compatible monsters.

slyflourish.com

Fantastic! Things like this are a large part of why I love the community so much.

 

[Krakenbreath]

Hey everyone! Like a lot of DMs, I've struggled to get CR to work reliably in my games. Unlike a lot of DMs, I can honestly claim that it's my fault.

5e drafts heavily off of 3e's core mechanics, so it made sense to recruit its encounter building tool. Rodney Thompson and Peter Lee both pushed to do something else, but we already had a small budget, a tiny team, and lots of work. I locked us into CR because it fit with our timeline and was a tool that our existing DM base already understood. Looking back, I think I made the right call as a producer, but it wasn't a great call from a design point of view.

Over the past two weeks I've been tinkering with an alternate approach to encounter building, one inspired by games like Warhammer 40k. It assigns a point value to characters and creatures. A balanced encounter has equal points on both sides. If the characters' point value is below the monsters, it's a tough fight. If the reverse is true, it's an easy fight.

EDIT: System tested and found wanting. Replacement up:

https://github.com/mikemearls/5e_point_encounters/blob/main/challenge_points.md

Here's the old version for memory's sake:
------------------------------------------------
I've put the bones of the system up on GitHub:

GitHub - mikemearls/5e_point_encounters: A system for balancing encounters in 5e using point values

A system for balancing encounters in 5e using point values - mikemearls/5e_point_encounters

github.com

The math is still early, so expect changes as I spin up some code to run a deeper analysis of the monsters and characters in the 5e SRD. Hit me up here with any questions or comments.

Mike, this seems crazy high to me. At 10 points I would expect the monster to be around the equivalent of a PC. By the tables proposed, it shows pretty much Monster CR = Character Level. By my experience, and some checks of Monster Book ‘Character’ CRs, the highest number I’ve got was Character Level = 3/4 Monster CR. This system also is WAY harder than the MCDM Flee Mortals system and the Lazy DM Benchmark. Is something up with the tables, or am I reading something wrong (or am I a pansy DM)?

 

[clearstream]

Ideally, I'd think the CR system should be able to provide a pretty standard "if they fight this, they will likely leave in this condition" outcome. The best case scenario is that you can then modify the PC input side (say, having some adjustment to the point totals, in the point based model being discussed) to account for the party in different resource states.

I don't think CR needs to account for every possible PC plan or encounter outcome, but that tool would have value even for you in the style suggested above. Your later encounters will have a PC party at a higher resource level than expected, and you'd be able to dynamically adjust for the new expected outcome and have some guidance in how to adjust later encounters if you wanted to preserve resource expenditure/challenge at some specific level.

I additionally don't see the value of four categories. Two - Deadly and Moderate - would do. A deadly encounter is one in which observed outcomes include a noticeable character death rate. Whilst the observed outcomes of attritional are just that (death hardly ever observed, but with noticeable costs.)

When it comes to fatality rates, a rough equation to consider is​

​

1/(%^E)​

​

Where % is survival rate, and E is encounters over adventuring career. You can derive encounters over career by picking your sweet spot (say level 9) and dividing total XP needed by XP per encounter (in any event, just sum expected encounters per level up to 9th).​

​

Say 95%, aiming for 9 levels expecting 2 deadly encounters per level.​

​

0.95^18 = around 40%, 1/.4 = 2.5​

​

Players would roll 2.5 characters each to have one reach level 9. 10 over a party of 4 to have 4 reach 9th. That's a rough estimate: you can set up a spreadsheet to improve it based on your reroll assumptions (e.g. are chars rerolled at 1st?), and obviously the math could be more sophisticated​

​

The observed fatality rate in one encounter is then 1-(%^P), where P is party size. If P is 4, it's near enough to 20%. Predicting a death every 2 or 3 levels gained.​

Deadly encounters are then tuned to yield the desired character reroll rate over party careers (levels 1-9 in my example.) To get fancy you can incorporate a revival magic assumption. For example that half of deaths are rezzed. That lets you bump up the rate chars go down without interfering in their careers.

I believe that narrowing to two categories will meaningfully reduce the playtest/iterate design effort with little to no negative impact on play. @mearls for vis.

 

[mearls]

To my reading, the system attempts to fix CP-per-character, and vary monster-point-cost by character level.

That doesn't take full advantage of your premise. Rather, fix monster-point cost, and vary CP-per-character by level.

The pictured future-MM then prints a point-cost for each monster (just like a WH codex). In the DMG would be a table of character CPs-per-level-per-difficulty-band. CRs would no longer be needed.

That would be easy to use. (With apologies if this critique is based on grasping your design incorrectly.)

I'd like to vary CP for both monsters and characters. I'm tackling monsters first, characters will come later.

In terms of still using CR, I see it as a bridge to what everyone is already using. If you have a book that uses CR, you can get a rough estimate of points. The creatures I'm designing right now just use points, with CR added as a reference for people who use that system.

 

[mearls]

Mike, this seems crazy high to me. At 10 points I would expect the monster to be around the equivalent of a PC. By the tables proposed, it shows pretty much Monster CR = Character Level. By my experience, and some checks of Monster Book ‘Character’ CRs, the highest number I’ve got was Character Level = 3/4 Monster CR. This system also is WAY harder than the MCDM Flee Mortals system and the Lazy DM Benchmark. Is something up with the tables, or am I reading something wrong (or am I a pansy DM)?

Keep in mind that the total points for an encounter is based on PCs.

So, given that CR is meant to represent a single creature that can fight four PCs and reduce their resources by around 25%:

10 points per character * 4 for four characters = 40 points total
A CR X creature should take out about 25% of an X level party = 40/4 = 10 points

If you're running for one PC, you'd have 10 points. A creature with CR equal to that lone PC's level would likely end in a dead character.

 

[mearls]

I believe that narrowing to two categories will meaningfully reduce the playtest/iterate design effort with little to no negative impact on play. @mearls for vis.

Wow, that is a great point. The nice thing about that approach is we can treat it as a threshold: For party X, keep below totals point Y to avoid character deaths, go above Y for potentially lethal encounters. That feels simpler and elegant. Thanks for breaking down the character death rate like that - super insightful!

 

[mearls]

I understand wanting to push back against an unfortunate trend in online discussions to hyperfocus on this one number (CR) at the expense of pushing the creative frontiers of the design – heck, I hate that tendency more than most folks, and push against it with my own stuff – but I think it's inaccurate to assume that Maths & Creativity are inherently opposed forces such that developing Maths must come at the expense of Creativity. Nothing about that is inherent.

One of the things that I loved about designing Magic cards was how much pressure it put on creating interesting effects beyond creature numbers. For TTRPGs, I think we sometimes focus too much on stats rather than on the effects a creature is using. TCG design is very much the opposite.

 

[mearls]

General thought based on observations from a few folks in this thread - I think this system could support multiple ways to build encounters. Warhammer literally has this. There are three ways you can build armies, one of which is traditional points, another has much fuzzier point values, and the third is to just show up with whatever minis you feel like using that day. I think multiple paths could be really powerful.

 

[mearls]

So, then, the powerful tool to grow the TTRPG community is to bring new DM's into new groups and public spaces and lean into the performance of D&D. We build skills like improvisation. We build our knowledge of various fantasy tropes and works (a little Appendix N homework). We include advice in the DMG for how to not freeze up, how to be a good host, how to attract players, how to help the party build characters with hooks and drama and backstory, how to present a setting without lore-dumps, etc. A lot more advice cribbed from screenwriting and acting and improv than we have now!

Fantastic point - we have seen TTRPGs grow tremendously due to streaming, I think because of exactly this. Once you see someone DM, you now have a model. The Yo Yo Ma analogy breaks down because there is a technical aspect to music (how to physically use an instrument, read music, etc) that you can't pick up by watching. Where with a streaming game, I can see the mannerisms and shifts in the group and pick up those cues.

Man, I love this community. This thread has been such a boon to wrapping my head around this.