D&D General Requesting permission to have something cool

[EzekielRaiden]

Repeated computer sim would be great for fine-tuning the math, no question there.

There's still important questions, though:

1. Is getting the math that finely-tuned a good thing or a bad thing?
2. Whose definition of "performing as desired" will be used to determine functionality?

How is it a bad thing to have a CR system that actually tells you the general tendency of how difficult a monster is? How is it a bad thing to know that players could choose the Noodlergy or Saucery subclasses of Pastamancer and overall be statistically similar?

Your first question is bizarre; it is like asking, "Is it good to know if a machine works or not?" I would argue that, barring purposes which genuinely should not be (e.g. "exterminate all life" or "enslave the minds of others" or other morally objectionable things), it is always better to know that something successfully achieves the purpose for which it was designed.

As for the second, the designers themselves. Hence why I advocate for designers having clear design goals. They made the game; they decide what the stuff in it is supposed to do. (5e's pillars are not clear design goals, but they are important principles from which design goals can be built, for example, if "socialization" is a critical component of the game, design goals related to that could include "every class has at least one tool useful for contributing to social encounters." The fact that social encounters exist and are important is not, in and of itself, a clear design goal, but it gives the foundation for building clear design goals.)

As I have said repeatedly, there are many things in D&D (or any game) that cannot, even in principle, be tested with this kind of modeling. Those things will always require real humans, with thought and judgment, doing testing. But a sword is designed to do a certain amount of damage. A spell of level N is meant to do less damage than a comparable spell of level N+1 and more damage than a spell of level N-1. Two subclasses of the same class should, in general, be comparable in their contributions to the party. Etc.

These things can all be subjected to simulation in addition to human testing for things that need the human touch. You don't need the human touch to test whether Champion's crit bonus is able to keep up with Battle Master's maneuver damage. You don't need the human touch to run a hundred thousand encounters of 4th-level parties against six ghouls to see if ghouls are killing player characters at the expected rates.

Human testing cannot be eliminated. Period. It will always be essential. Period. But simulation and modeling can be a powerful tool, both for checking to make sure that the things you make actually achieve the goal for which they were designed, and for helping us dedicate that essential human labor to the most useful things it can be, because it progresses slowly and people-time is valuable.

 

[Pedantic]

Sure, my simple cage match isn't enough but I could fairly easily expand it. I already track resource usage for battle master maneuvers for example so tracking spells used would be simple enough. You can randomize some things, I randomized how quickly the fighter used their second wind. I have no desire to do so but I could add multiple entities, initiative order, etc.

But it's still all going to depend on input variables, assumptions and what is being measured. What spells does the wizard have? How many targets are going to be in an AOE? Does the party focus fire or spread out damage? Do the monsters? How often is cover going to come into play, how often will the rogue be able to use stealth? Do we throw these groups into an AI that determines optimal tactics or do we just base it on the experience of the testers? Do we modify behavior at all to mimic how different groups are likely to play?

There are simply too many variables and I didn't even get to magic items, rolling for stats, does the DM use morale. The list goes on.

You now seem to be arguing that game design is hard, which is true but not particularly interesting. If you wanted a workflow, I'd probably start by doing what we've done here and listing out a variety of likely scenarios and variables, then follow-up with a same game test of several different party compositions to see if I can find confounding strategies or factors I haven't considered, ideally with different play groups, and then I'd iterate on the model a few times until I'm relatively happy with the current balance point, and do it all again with the next set of classes on my docket.

Or again, do you think there's some other job game designers ought to be doing? What is their role exactly, if not doing this?

 

[EzekielRaiden]

But do you have any reason to think they're likely to?

I mean, if BG3 is as wildly successful as it seems to be, and WotC is working on an in-house VTT...yeah, I think there's reason to believe that they expect to be modeling stuff in computer-land.

There's been quite a bit of commentary (IMO, more pearl-clutching) that 5.5e is being built to be more compatible with a computer implementation of the rules. If there's even a lick of truth to that, it seems to me reasonable to say that WotC has at least thought about such simulations.

Edit: And if WotC thinks they can reduce their (slow, expensive, time-consuming) in-house playtesting by having a computer run a bazillion combats overnight, they'd do it. Cutting costs has been one of the big watchwords of 5e's lifespan; for the first several years, they were operating with a skeleton crew. Given the recent high-profile outrage over attempting to invalidate the OGL, a quiet, completely internal method to speed up playtesting and reduce manpower expenditure sounds like it would be very, very tempting to the current management of WotC.

 

[haakon1]

The solution to the problem is for WotC to actually goddamn analyze their own game. A bit of statistical analysis
…
I will continue to criticize WotC's piss-poor survey design, lackluster to nonexistent mathematical testing, and frankly bizarre logic
…
So...yeah. I do have a solution. The solution is to actually bloody test the rules--and when rules are found to be busted, FIX THEM, don't just let them sit there broken for a bloody decade.
…

I have a solution.
(1) You go prove to Elon Musk, venture capitalists, or other investors that your design for D&D is better than those bloody boneheads in Renton and get them to bankroll you.
(2) You get Hasbro to sell you D&D.
(3) You them how it’s done right. Fire anyone who disagrees with you. Write the next edition yourself if necessary.
(4) Sit back and bask in the success of finally getting it right.

As Teal’c said to O’Neill in Stargate SG-1, and I paraphrase, many have said they could make D&D better, but you might the first to succeed.

 

[EzekielRaiden]

I have a solution.
(1) You go prove to Elon Musk, venture capitalists, or other investors that your design for D&D is better than those bloody boneheads in Renton and get them to bankroll you.
(2) You get Hasbro to sell you D&D.
(3) You them how it’s done right. Fire anyone who disagrees with you. Write the next edition yourself if necessary.
(4) Sit back and bask in the success of finally getting it right.

As Teal’c said to O’Neill in Stargate SG-1, and I paraphrase, many have said they could make D&D better, but you might the first to succeed.

I think your facetiousness is showing.

 

[Oofta]

You now seem to be arguing that game design is hard, which is true but not particularly interesting. If you wanted a workflow, I'd probably start by doing what we've done here and listing out a variety of likely scenarios and variables, then follow-up with a same game test of several different party compositions to see if I can find confounding strategies or factors I haven't considered, ideally with different play groups, and then I'd iterate on the model a few times until I'm relatively happy with the current balance point, and do it all again with the next set of classes on my docket.

Or again, do you think there's some other job game designers ought to be doing? What is their role exactly, if not doing this?

How many complaints do we see about how the CR system "doesn't work" and that the game is too easy? Because that could be easily caused by 2 factors. First is that people simply don't understand what CR means or how to use it because they don't read the DMG. The other is that they set the bar low because they're assuming a 4 person group of newbies using point buy and no magical items. Were the latter the correct assumption? I happen to think so, you don't want new players to get discouraged and the DM will always have to adjust difficulty based on the party and the encounter scenario anyway.

Let's take 2 groups I ran for group A and group B over the same timeframe, using same basic assumptions. Group A was simply far more effective than group B between tactical acumen of the players and perhaps the mix of classes chosen. So it was obvious pretty early on that group A could handle 1,000 XP in encounters per long rest, group B could only handle 750 XP in encounters per long rest. This was with minimizing build options, similar magic items, etc.. Group A was simply better at combat.

So how many groups fall into category A and how many fall into category B? Heck if I know. That's just one sliver of all possible variables. The outcome is going to vary by what variables you put in and what you measure and a simulator isn't going to change that. If you work out the math correctly, a simulator isn't necessary but it will still always be an educated guess on all the variables and all we can hope for that the results are roughly in the correct ballpark.

 

[haakon1]

There are people that are really into creating a "complete character" at chargen, including a multi-page backstory, fully formed personality, and even a plan for their future levels. And that's totally cool!
…
On the other hand, some people find it interesting to see how a character becomes revealed to them during play- that they start with the sketch of an idea
…
Neither approach is better, and the former is certainly more prevalent today than the latter. But both can be fun and interesting.

What makes you certain the “CharOp 20 level build plan” approach is more common than the “make a 1st level character and see what happens”?

For the 60+ people I’ve DM’d, admittedly all using 1e/3e/3.5e rules and 20 or so more with non-D&D RPG’s, I think a player has told me their “build” plans for higher levels from the start maybe twice?

That said, either Char Gen approach can be linked with “here’s an elaborate back story”, “here’s about a half page of backstory”, or “he’s like an orphan or something” refusal to backstory.

 

[Minigiant]

No the problem with 3.0 and 3.5 was it started out with a design philosophy stated by Wizards that they were throwing everything out there they could think of and it was up to the DM to decide what was appropriate for their games. But eventually every single game turned into a fight because someone wanted something that was either ridiculous, (say the 10hd battle cat out of the barbarian splat book that was a better warrior than the barbarian, and was so poorly written that many thought thier 1st level barbarian could have one.), or simply didn't match the game being run like say a Neogi cleric in a non spell jammer game. That's the first time I remember "core rules only" becoming a normal thing. It wasn't that the players became magic user's it was that they became anime action hero's. Nothing wrong with that kind of game but If I'm going to run that game TMNT or Hero's unlimited has better rules for it.

Quality control went to the toilet and 75%+ went to casters to abuse it.

Nature human instincts.

It's why the aliens won't make contact.

They know if we learned their tech, we'd wreck them using loopholes

 

[haakon1]

Quality control went to the toilet and 75%+ went to casters to abuse it.

Didn’t you say 3x was “all casters”? Can we settle on “50-60% casters if you include Divine Casters and Bards”?

I really really like (present tense) 3x, but obviously you didn’t.

People who like complexity can try and eke out some small advantage, but only by showing how super awesome they are (which is the point, right? ).

Maybe. I’m not even sure what the point of this conversation is,

Feels like:
“D&D is broken! It’s not balanced. Someone needs to fix it.”
Meaning perhaps: I want a more powerful build but my mean DM said stick to RAW. Or: I don’t like classes I don’t play, and I think people who play Fighters are chumps having wrongbadfun.
If only everyone could be as awesome as me and understand I’m right!

Versus
“Meh, I like D&D. If you don’t like every bit of it - and who does - you can find some alternate rule, or homebrew, or a different system.”
Meaning: Stop whining about the same stuff and telling me I’m stupid for liking the World’s Most Popular Role Playing Game (tm).
If only everyone could be as nonchalant as me and just play!

 

[Lanefan]

Sure, my simple cage match isn't enough but I could fairly easily expand it. I already track resource usage for battle master maneuvers for example so tracking spells used would be simple enough. You can randomize some things, I randomized how quickly the fighter used their second wind. I have no desire to do so but I could add multiple entities, initiative order, etc.

But it's still all going to depend on input variables, assumptions and what is being measured. What spells does the wizard have? How many targets are going to be in an AOE? Does the party focus fire or spread out damage? Do the monsters? How often is cover going to come into play, how often will the rogue be able to use stealth? Do we throw these groups into an AI that determines optimal tactics or do we just base it on the experience of the testers? Do we modify behavior at all to mimic how different groups are likely to play?

There are simply too many variables and I didn't even get to magic items, rolling for stats, does the DM use morale. The list goes on.

Indeed; but I think @EzekielRaiden has the right of it here: turning BG3 loose on this to run simulations, using both pre-set and randomized variables, over and over again as a sort of AI-adjacent data mine seems far easier (not to mention faster) than doing any of it by hand.