Revive 2d10

[mlenser]

I removed it for the reasons specified in https://bitbucket.org/mlenser/tabletop-homebrew/issues/314/solve-2d10s-problems-or-abandon-it, but I believe those can be overcome.

I've long thought that 2d10 is overall a better system. https://faustusnotes.com/2018/05/23/what-is-the-best-basic-die-roll-for-dd-5th-edition-a-probabilistic-analysis/ covers some of the math

If I go through with this I'll make 2d10 the default, but keep 1d20 as a setting.

[shemetz]

I think it's wrong to discard this idea because of numbers being a bit above or a bit below the RAW numbers - it looks to me like this can be easily fixed by just bumping proficiency modifiers or DCs or Defense by 1 or 2 (up or down).

However, I think there needs to be a lot more discussion about the impact of the shape of this curve. The article linked in https://faustusnotes.com spends 95% of the text just going through charts, and only barely mentions the actual pros and cons of this change, which I dislike - D&D (especially 5e) is heavily designed according to what feels good.

At the most basic level, this change would mean that easy things are easier and harder things are harder. The "common" probability of ~60% success at an average task (such as stabbing or shooting an enemy) may remain the same, but the extremities skew further up and down, making everything more consistent in a way that isn't always more fun.

I know a lot of people want skill proficiencies to be more consistent, and while this does it, it isn't necessarily the best solution, as it has a lot of side effects.

I'm strongly in favor of listing exactly the problems this means to solve, and listing the potential problems this can create (for example, I think it'll really hurt the idea of bounded accuracy for battles against very strong or very weak enemies).

I'm less interested in discussions about numbers for - those can be tweaked rather easily. For example, if using 2d10 ends up causing a lot more near-misses (where half damage is dealt), this is not a real problem because it can be fixed by reducing the miss threshold from 4 to 3. Similarly, the problem of critical chances changing from 5% to 1% can be fixed by redefining them to be "roll a 19+" or "roll 9+ on both dice".

[mlenser]

I agree with what you're saying. It should definitely be fully explained and explored to ensure it is a good option before I implement it. I wrote about the numbers because the math was the main reason it was removed before.

I'll write pros/cons later

[mlenser]

The goal of this change is that a character attempting a task (picking a lock, swinging a sword, remembering their history, jumping across a gap, etc) will generally have consistent results.

A jumper will generally jump ~10-15 feet (generic numbers, not to be taken as statistical math), pretty consistently. There could some some cases of a jump that is perfect and the result is longer or some cases that the character stumbles and the jump is much shorter, but I believe those cases should be much less common.

Ignoring success rates, math, and mechanics, more consistent results is the goal.

I believe 2d10 could be implemented currently without the game having very many cracks. Though we could probably refine a few things.

---

Math:

[shemetz]

A lot of these changes would be great - especially for skills (like jumping or picking a lock).

For fighting ("swinging a sword"), it's a completely different issue. An expert athlete will certainly jump perfectly almost every time, but fighting is inherently a lot more random. If a system like this is applied to combat, it will inevitably break the idea of low-level threats still being a danger.

With the current system, a CR 1/4 Goblin will hit a level 10 character about 30% of the time, and a CR 15 Mummy Lord will hit about 60% of the time. If we use a system like 2d10, these chances will have to change to something like 5% and something like 90%. The chances for the players to hit the monsters will be roughly the same and flipped, of course.

I'm not immediately opposed to this, but this will bring us back to older D&D systems, where low level threats are not a problem.

If the goal is just to solve consistency out of combat, I think there are other solutions that can be implemented.

For example, making passive skills a "floor" for active skills, or adding mechanics that only care about your modifier and don't require your roll.

There could also be meta-mechanics like Inspiration which allow the players to get consistency in moments that matter, but not always. There's a lot of tabletop RPGs (and some video games) where the players can spend "inspiration", "focus", "fate points", "essence", etc. for D&D, these could allow you to "take 15" (or 10 or 19) instead of rolling. alternatively, they could give you a bonus after you've made your roll, much like bardic inspiration.

So, in summary - I think there can be other solutions to make skill checks consistent (or even a partial solution: use 2d10 for skills and 1d20 for combat). I don't see a way where combat could be changed to 2d10 without making extremely strong/weak enemies even more/less of a threat. In other words - some monsters will be consistently trivial or consistently unbeatable.

[mlenser]

For fighting ("swinging a sword"), it's a completely different issue.

I don't think it is. A goblin can shove just as much as it can swing its scimitar. While both sides of a shove are rolled, the difference between bonuses is still amplified.

If a system like this is applied to combat, it will inevitably break the idea of low-level threats still being a danger. With the current system, a CR 1/4 Goblin will hit a level 10 character about 30% of the time

I don't believe it would. I'm not seeing how you concluded your numbers. Anydice:

Level 1-4 (Defense ~16)

• output 1d20 + 4 > 16: 40%
• output 2d10 + 4 > 16: 36%

Level 1-4 with shield or Level 5+ (Defense ~18)

• output 1d20 + 4 > 18: 30%
• output 2d10 + 4 > 18: 21%

Level 5+ with shield (Defense ~20)

• output 1d20 + 4 > 20: 20%
• output 2d10 + 4 > 20: 10%

---

This is not just about skills/out of combat. It's about every 1d20 roll in the game.

[Lamorak11]

I'll be frank, I have nothing to add about the validity of a 2d10 system.

However, I will say, for the sake of my position being out there and in case there are others that agree, that the d20 is a sacred cow that I am flat-out unwilling to give up.

If, as was stated, the d20 system would remain fully usable at the same time as a 2d10 system, then I have no issues with the addition. However, if there is a point where the d20 is no longer supported, my play group and I will regretfully have to go back to vanilla. As I am extremely invested in KRYXRPG, I really don't want this, but hopefully a solution can be found that avoids this.

I am stating this here as I didn't want this discussion to get to an advanced point without at least pointing out my opposition.

[mlenser]

@Lamorak11 as I wrote in the OP:

If I go through with this I'll make 2d10 the default, but keep 1d20 as a setting.

I previously used 2d10 from July 2017 until February 2019. The changes were very minimal and I don't expect them to expand far into the system.

[Paulorpribeiro]

The main con I see with 2d10 relative to 1d20 is that the +1 mod does not mean a 5% increase anymore, but actually range between 2% (if you needed a 20 and now need a 19) and 10% (if you needed 11 and now need a 10). This means that a maneuver or spell that gave +2 defense has to be rebalance. Not only those, but all spells, maneuvers and class features that give bonus/penalties to attack, defense, saves or skills will have to at rebalance or at least checked if they are not too powerfull or too weak. I think that is a massive work for the benefit it brings

[shemetz]

@Paulorpribeiro - I think many of the pieces of the system that give ± to attacks, defence, etc - are balanced around affecting a "standard" chance (of around 55% to hit), and do not easily stack to high numbers. The system purposefully makes it difficult to stack multiple such things together, and the chances of success tend to consistently hover around that range, so I don't think there are such potential problems that aren't already problems in the existing system.

By the way, another thing came to mind - with this chance, the chance for "near misses" increases from 20% to 30% (assuming 10+ is a hit).

[Paulorpribeiro]

I see, this change will only exacerbate (is that a word? Enhance might be better) the problems already in place. For exanple, getting a +2 defense as a reaction when the enemy already hits you only on a 18+ decreases his chances from 6% to 1%. Your maneuver had the value of +1 comparing to the old system. On the other hand if the enemy needed a 10 to hit you, his chances decreased from 64% to 55%, almost the 10% it was in the old system.

Another thing to consider, that I just figured out, is that advantage is worth less with 2d10 than it did with 1d20. In 1d20 it was worth between +5 and +2, higher on the mid range. With 2d10 it varies between +1 and +3, higher in the mid range. This can also be adjusted, but one more detail to keep in mind

[Marcloure]

My main concern about the bonuses stopping being constant is that the players and GMs (unless they have good statistical knowledge) will not know if it's better to use shield of faith on the plate and shield paladin or on the greataxe berserker. They may use it on one target unknowing that the bonus on it will be less effective than they expect.

The flat d20 doesn't have this issue, a +2 is equally valuable to both characters, so there isn't less effective use.

Now, this isn't a problem if we don't care about players and DM optimizing moves and Mana. If we want them to target the creatures they want to, regardless of they knowing it is optimized or not, then it's fine. In many systems with very complex dice resolution, you use a +1 bonus on who you want to have a slightly higher chance of success, even though you don't know what that chance increase really is.

[mlenser]

As itamarcu allued: we can fix the numbers. Statistical outcomes of small number bonuses can be adjusted. And if we decide to go through with this, I can outline some examples to show how the math changes. Cases like Shield of Faith, Sharpshooting Fighting Style, Shield, Partial hits, etc will have a statistical value similar to what they are today. That's not difficult or much effort.

So lets please ignore that for now.

---

One of the biggest issues I'm working through in my head right now is regardign defense and keeping lower challenge creatures a threat while not making defense nearly worthless against threats around the same challenge as you progress in levels.

I'll have to fix that for both 2d10 and 1d20, so maybe best to fix that now. I'll address that in #538

[mlenser]

Regarding ability checks and 2d10.

see Skills math

vs a Medium DC 1d20: 85% for prof, 70% for no prof 2d10: 94% for prof, 85% for no prof Expertise was 95%

vs a Hard DC 1d20: 60% for prof, 45% for no prof 2d10: 72% for prof, 45% for no prof Expertise was 75%

vs a Very Hard DC 1d20: 25% for prof, 20% for no prof 2d10: 28% for prof, 10% for no prof Expertise was 50%

So Medium DCs are ok and Very Hard DCs are not meaningful changed (3% difference). So the end result here is that 2d10 is no where near as bad as expertise was for Medium DCs or especially Very Hard DCs. 2d10 makes Hard checks easier for those with proficiency (one could argue this is problematic), but does not seemingly have the same problems with even harder DCs.

For me this is enough evidence that I feel comfortable with 2d10 for skills. It allows those with specialization the ability to have more reliability without causing the major problems of "all or nothing".

[mlenser]

Defenses and saves

See the math (a bit unorganized): https://docs.google.com/spreadsheets/d/1sGvcz9vbq9yk4kRBEY6ZYGWhGO6-t7CWudznQcWPk74/edit#gid=158786285

2d10 and partial hits allows low challenge monsters to continue to be threats into high levels far better than RAW does.

Compared to 1d20 & partial hits

• Plate & shield damage is ~90% of what it is with 1d20
• Plate damage is ~109% of what it is with 1d20
• Scale, Hide, Leather, and Mage Armor damage is ~119% of what it is with 1d20

It is reasonable that Plate & shield is extraordinarily tanky.

Compared to 1d20 without partial hits

• Plate & shield damage is ~145% of what it is with 1d20
• Plate damage is ~156% of what it is with 1d20
• Scale, Hide, Leather, and Mage Armor damage is ~159% of what it is with 1d20

[mlenser]

The problems of 2d10 that caused me to abandon it are not a problem with partial hits and partial sucesses.

[mlenser]

2d10 added as an alternative to 1d20 and set as the default option