Brainstorming some fairly radical/major changes to D&D 5e

[jjordan]

Just throwing some ideas out here. It pretty much amounts to a rewrite of D&D.

-Point-buy characters rather than rolling.
-Classless system. Skill packages are suggested to represent the existing classes.
-Everything is skill-based. Characters can be built by acquiring levels in skills: proficiency (+Proficiency), expertise (+2xProficiency), feat (granting special abilities).
-Skills will include weapons, armor, languages and etc.
-3d10 rather than 1d20. Advantage gives you an extra die, keep the highest three. Disadvantage gives you an extra die, keep the lowest three.
-Existing feats and racial/species abilities and class abilities and magic are rolled into the skill system.
-Hit dice, ability improvements, spell slots, and etc... can be purchased using experience points. Levels are gone.

[JNAProductions]

Just throwing some ideas out here. It pretty much amounts to a rewrite of D&D.

-Point-buy characters rather than rolling.
-Classless system. Skill packages are suggested to represent the existing classes.
-Everything is skill-based. Characters can be built by acquiring levels in skills: proficiency (+Proficiency), expertise (+2xProficiency), feat (granting special abilities).
-Skills will include weapons, armor, languages and etc.
-3d10 rather than 1d20. Advantage gives you an extra die, keep the highest three. Disadvantage gives you an extra die, keep the lowest three.
-Existing feats and racial/species abilities and class abilities and magic are rolled into the skill system.
-Hit dice, ability improvements, spell slots, and etc... can be purchased using experience points. Levels are gone.

This doesn't sound like 5E at all. Why even use it as a base?

[PhoenixPhyre]

This doesn't sound like 5E at all. Why even use it as a base?

Agreed. There's a Ship of Theseus issue here--how much can you remove and still reasonably call it the same thing?

[jjordan]

This doesn't sound like 5E at all. Why even use it as a base?

There's a Ship of Theseus issue here--how much can you remove and still reasonably call it the same thing?
I will respond that material isn't really being removed. For the most part it's being reorganized and, in many ways, simplified. The key mechanic of rolling to exceed a target number is still there. The magic spells are unchanged. The spells slots are still there. The lore is unchanged. Players are already building custom backgrounds and Wizards has been moving to legitimize that practice. In fact, a lot of this is arguably the direction Wizards is already moving. I'll grant they are unlikely to go this far, however.

-Using point buys to create characters is already a popular option in D&D that allows players greater control over their character, being the embodiment of the take-20 principle.

-Moving to a classless system is the elephant in the room that will make everyone uncomfortable. But it's really the best solution to both simplify the game and give players more control over their character builds. Players are still going to make Wizards and Fighters and acquire the skills and tools specific to those design builds, but they are going to be less stuck with the WotC conception of what those should be. The optimizers will very quickly create models for these builds that will become de facto standards. And, honestly, getting the community more involved in the structure of the game would pay off in the long run.

-Making everything skill-based is the key to moving to a classless system. Keeping the skill system simple with only four levels (no skill, proficiency, expertise, feat level) is going to be important to preserve playability. Again, Wizards appears to be trying to figure out how to move to this without moving to this. Lumping weapon and armor and language and tool proficiencies into this is more of a simplification than anything. I mean, it's already in the system in some cases; the weapon and armor proficiencies already have no skill, proficiency, and feat levels.

-3d10 is the baby elephant in the room. It's still rolling to exceed a target number so the mechanics of the game are largely unchanged. It only extends the possible range of results by seven digits. If DCs are left the same then it skews the results toward player success. If DCs are modified then it opens up some lore options (we could extend some of the armor options, for example) and the bell curve still skews towards general success while being less prone to extremes of success. I'll admit this last is not going to read as a feature for a lot of players.

-Getting rid of levels and allowing players to acquire improvements a la carte takes away the Christmas morning level-up but, again, allows players greater control over their character progression and allows participants greater integration of their progression into the story. I know this will appeal to some players and repulse others. Moving to this mechanic still allows for batch upgrading of characters (aka leveling up) but with more flexibility. Players could acquire improvements after a single session if they have acquired enough XP, they could do milestone leveling when they hit certain levels of XP or certain places in the progression of the game, or even do it anytime their characters hit a location and take some time.

[JNAProductions]

I'm not saying you should completely and totally ignore 5E-by all means, lift ideas from it.

But I think you'd be much better served starting with something like GURPS or M&M as a base, rather than 5E.

[jjordan]

I'm not saying you should completely and totally ignore 5E-by all means, lift ideas from it.

But I think you'd be much better served starting with something like GURPS or M&M as a base, rather than 5E.

Where GURPS is concerned I disagree. The 3d10 dice mechanic would, with simple substitution, skew towards failure as well as nearly doubling the possible results range (which, granted, GURPS would benefit from). The skill trees in GURPS are a nightmare to deal with and would require substantial modification, particularly if you want to retain the simplicity of play WotC has achieved. Which brings me to the subject of action resolution mechanics that would need to be substantially changed (moving from a 1 second action phase to a 6 second action phase leaps to mind). And then all the lore would need to be transferred over; every spell would have to be rewritten. Building this from a GURPS base would be a lot of work. A lot more work.

I'm not at all familiar with M&M (the last time I played superhero RPGs we used an early edition of Champions or Villains and Vigilantes). But M&M is based on the d20 system so I *think* it wouldn't be substantially different from altering D&D 5e and since I want to preserve a lot of 5e I'd probably be better off starting from that base.

[Yakk]

Just throwing some ideas out here. It pretty much amounts to a rewrite of D&D.

-Point-buy characters rather than rolling.
This isn't a change, there is that in standard 5e.

-Classless system. Skill packages are suggested to represent the existing classes.
This is a significant change. However, note that MCing is similar to this.

-Everything is skill-based. Characters can be built by acquiring levels in skills: proficiency (+Proficiency), expertise (+2xProficiency), feat (granting special abilities).
Unless you are better at math than you appear to be, you'll break bounded accuracy here, and render most problems into auto success/failure.

-Skills will include weapons, armor, languages and etc.
Unless you are very good at game design, the result is that optimal players often know ridiculously narrow sets of things. "I know all weapons" becomes a suboptimal choice.

-3d10 rather than 1d20. Advantage gives you an extra die, keep the highest three. Disadvantage gives you an extra die, keep the lowest three.
Ignoring advantage/disadvantage for now, 3d10 moves the mean up to 16.5 (so +6) and changes the standard deviation. It also requires a different critical hit/miss mechanism (which in core 5e is only on attacks remember).

The Variance of 1d20 is 399/12, the Variance of 3d10 is 297/12; 4/3 isn't a big ratio, and its square root is under 1.2.

Boosting every DC by 6 and adding +6 to all d20 rolls will be nearly indistinguishable from this one. A SD of 0.85x ish of d20 has tiny effects on actual gameplay. It does raise the brainpower required to resolve a roll, which slows play in certain ways, which may be of use if you want each roll to have more consequence.

Now, your change does make advantage/disadvantage weaker; 4d10dl has less of an impact than 2d20dl compared to 3d10/1d20.

I suspect that you think this will make a bigger difference than "reduce size of advantage/disadvantage, make numbers 6 bigger, change crit mechanics". You probably have fallen for the bell-curve illusion, like most people who add up multiple dice for the core resolution mechanic.


-Existing feats and racial/species abilities and class abilities and magic are rolled into the skill system.
So that is a lot of work.

-Hit dice, ability improvements, spell slots, and etc... can be purchased using experience points. Levels are gone.
The usual side effect of this is even less ability for the DM to have an idea of how competent PCs are given a certain point of build points/xp/etc.

Inside party balance wise, the advantage is that a gonzo system doesn't let players lookup charop results. The disadvantage is that you are probably 2 orders of magnitude worse at balancing than the writers of D&D are, so it won't take much brainpower to break things horribly.

---

In short, have fun.

[jjordan]

This isn't a change, there is that in standard 5e.Yup.


This is a significant change. However, note that MCing is similar to this. Yup, again. And people are already modifying backgrounds and classes by swapping things in and out.


Unless you are better at math than you appear to be, you'll break bounded accuracy here, and render most problems into auto success/failure.That's a risk. :)


Unless you are very good at game design, the result is that optimal players often know ridiculously narrow sets of things. "I know all weapons" becomes a suboptimal choice.Not very good at game design but I'm not going to shoulder the burden of hyper-specialized characters. They've always existed and they will always exist and some players will enjoy the heck out of them and other will not.


Ignoring advantage/disadvantage for now, 3d10 moves the mean up to 16.5 (so +6) and changes the standard deviation. It also requires a different critical hit/miss mechanism (which in core 5e is only on attacks remember).Well, only if I want to ensure criticals happen with the same frequency. I can keep the mechanism (probably a 3-4 for a critical fail and a 29-30 for a critical success) while accepting that criticals will happen much less frequently. It's only if I want to get up to that 5% chance that I have to alter the mechanism. Admittedly it might be something I want to do because players find it fun.


The Variance of 1d20 is 399/12, the Variance of 3d10 is 297/12; 4/3 isn't a big ratio, and its square root is under 1.2.

Boosting every DC by 6 and adding +6 to all d20 rolls will be nearly indistinguishable from this one. A SD of 0.85x ish of d20 has tiny effects on actual gameplay. It does raise the brainpower required to resolve a roll, which slows play in certain ways, which may be of use if you want each roll to have more consequence. Wait.... Shouldn't I actually be multiplying each DC by 1.35 to account for the 27 possible totals of the 3d10? That means, rounding up, a DC2 would be a DC3 and a DC20 would be a DC27. That's considerably different from just adding six. And I'm not sure I want to change any of the DCs anyway because if I scale them the effect of the bell curve is to make success less likely.


Now, your change does make advantage/disadvantage weaker; 4d10dl has less of an impact than 2d20dl compared to 3d10/1d20.Yes, it does.


I suspect that you think this will make a bigger difference than "reduce size of advantage/disadvantage, make numbers 6 bigger, change crit mechanics". You probably have fallen for the bell-curve illusion, like most people who add up multiple dice for the core resolution mechanic.Yep, totally bought it. I mean, the bell curve does indicate that you will tend to get results which are closer to mean and that does reduce your chances of rolling at either end of the spectrum. A player has a roughly 30% chance of accomplishing a DC15 task in a 1d20 system, absent any modifiers. In a 3d10 system the scaled DC would be a 21 and a player would have a roughly 20% chance of accomplishing that, absent any modifiers.


The usual side effect of this is even less ability for the DM to have an idea of how competent PCs are given a certain point of build points/xp/etc.Yeah, that's a really good point.


Inside party balance wise, the advantage is that a gonzo system doesn't let players lookup charop results. The disadvantage is that you are probably 2 orders of magnitude worse at balancing than the writers of D&D are, so it won't take much brainpower to break things horribly.Also good points, and probably an underestimate.

[Man_Over_Game]

Important note related to the dice change: The more dice you add, the more predictable the results.

Consider the comparison of two roughly even dice calculations: 3d6 vs. 1d20

Say that a 5 caused you to miss with a 1d20, so you have a 25% chance of missing. That same 5 on a 3d6 is a 4.63%, for a ~95% chance of hitting. Comparing a simple +/- 2 points from stats and AC that modify your chances to succeed from the midpoint (that is, going from 10 to 8 or 10 to 12 as your AC), that is a whopping 50% difference in your odds (changing a 50:50 chance to a 25:75 or a 75:25 chance).

Or, in essence, the more dice you add, the more predictable the outcome. The more predictable the outcome, the bigger deal stats become. And to add to that, the more important stats are, the less the underdog has a chance to succeed.

And for some games, like Chess, you don’t want any unpredictable outcomes. But that might not be a good solution for a TTRPG.

[Morphic tide]

I'd suggest falling back a bit to Generic Classes, where there's a very small number of classes for extremely broad roles. This has generally taken the form of a front-line class, a skill-based class, and a magic user. Keep one Proficiency Modifier, keep discrete levels that mandate you gain HP with offenses, but then have the particulars as open-ended as you can manage. This is because 5e is a class-based system so massive swaths of the game are built around assuming you're gaining durability over time passively, whereas an open degrees of proficiency system where HP is something you have to choose to get is prone to hyper-focused madmen who are too binary to work into a campaign because anything that survives their first turn will be killing them in one hit.

So basically, roll back discrete purchasing of core character functions like health, proficiency scaling, and spell slots, so that regular game content operates mostly as normal because you're not freeing up the budget for a fully degenerate glass cannon because you preserve the anchor of bounded accuracy and don't let them ignore all survivability and endurance for even more damage.

[Anymage]

The devil of turning any D&D derived system into a point based builder is very much in the details, since more freedom by its nature will make it easier to minmax and home in on the strongest options. So I'm unsure about that from the getgo before even seeing the point costs you'll want to apply to everything. Classes do bring strong archetypes and package deals, things D&D has gone all in on, so I wonder why that's the base you want to start from instead of something initially built from those premises.

Doubly so because the binary proficient/not proficient (with expertise being a third case for skills) is a playability thing that very much sacrifices realism. Different people can have differing levels of skill even beyond the innate aptitudes reflected in stat mods and the slowly growing difference in proficiency mod. The logical extension of a point based system would most likely open up skill modifiers to be individually raisable. Either with a cap like 3.x's skill system, or just counting on increasing costs to make investing all your points in one thing have diminishing returns.

And with that sort of thinking in mind, what makes you want to start from 5e as a base and then heavily mod it instead of starting from something more point based in the first place? Other than the dice change your whole focus seems to be about enhancing build freedom. Knowing what goals and tastes you bring that makes you want to start from a 5e D&D base might help people point you at other stuff that you can crib from more effectively.

[Yakk]

https://anydice.com/program/21532

Click on "at most" and "graph". I really ask you to do that; it make the rest of this post make much more sense.

This is 1d20 vs 3d10, where I just translated DCs and scaled them by 86% towards the average roll. (I then multiplied everything by 100, because anydice doesn't like fractions).

The only significantly different region is the "outer 5%" which corresponds to a natural 1 or 20 in a d20; the crit hit/miss mechanics. All other changes from rolling 3d10 are so small, they are smaller than adding a +or- 1 to your d20 check, but only on some DCs.

NdX on roll-over systems are illusionarily different if you look at the "roll exactly", but you don't roll exactly in roll-over systems. You roll over.

And the integration of the probabilities means that correcting for average and standard deviation makes most of the other differences evaporate, other than the outer 5% "crit" hit/miss mechanics.

The Variance of 1dX is (X^2-1)/12; variance on NdX adds up linearly. Standard deviation is the square root of variance.

You take your DCs, scale for a change in average (mean), then scale towards the mean based on the differences in standard deviations.

So a 3d10 system is a 1d20 system with 6 higher DCs and each DC point away from 17 is 16% "bigger"; ie a DC of 23 in 3d10 corresponds to a DC of 23-6=17, (17-11)=6, 6*1.16 = 7, 11+7 = 18, or DC 18 in 1d20 system.

Graphing it we get https://anydice.com/program/21533 -- 15% for 1d20 on a DC 18, vs 12% on 3d10 on a DC 23. Difference of 3%, or 1 time in 30.

Similarly, modifiers to your 3d10 roll are 16% more important than modifiers to your 1d20 roll.

People usually use 3d10 or other dice combinations out of some belief that the bulge in the bell curve in the middle has some other noticeable effect; it really doesn't in roll-over systems. In effect, it just scales DCs and modifiers by relatively linear effects.

---

This isn't to say "don't use NdX". But I'm saying be aware of the actual effects of it.

I'd also suggest you make the multiple dice do something less boring than "add up". Something can happen on pairs, triples and on runs. Their probabilities might not be great.