Letting players choose HOW to roll skill check etc. (1d20 vs 2d10 vs 3d6)

[Greywander]

I've been thinking about making a hack of the 5e rules, and one of the modifications I'm considering is a change to the core d20 mechanic. What I want to do is give the players more freedom over their dice rolls and create more tactical choices over how they roll their dice.

Under the system I'm considering, the default or standard roll would be 2d10. This skews moderately toward the average, and makes it impossible to roll a natural 1 while still being possible to roll a natural 20. The average roll is 11.

Players could also choose to roll cautiously, which would be a roll of 3d6. This skews heavily toward the average, and makes it impossible to roll a natural 1 or 2, but also to roll a 19 or 20. The average roll is 10.5.

Finally, they could roll recklessly, using the classic 1d20. All outcomes are equally likely, and the full range of values from 1 to 20 is available. The average roll is 10.5.

The more I think about it, though, the less use I see for the "standard" roll. Mathematically speaking, you'd always want to roll either 1d20 or 3d6. If your average roll is enough to succeed on a roll, then roll cautiously. If an average roll is insufficient, roll recklessly. The "standard" roll has an average roll that is 0.5 higher than the other two, but I don't know that that would justify keeping it around. The only reason I can think of to make a standard roll is if you don't know if an average roll will succeed or not.

One thing I'm noticing is the treatment of 1s, 2s, 19s, and 20s. A cautious roll removes the possibility of getting any of those values, while a standard roll only removes the 1. I'm wondering if it would be a good idea to incorporate those values into a critical failure/success system, and if that would be sufficient to justify the existence of a 2d10 roll option. If rolling a 1 was really bad (like, "your weapon breaks," or, "you injure yourself"), it might justify rolling 2d10 instead of 1d20 just to avoid rolling 1s.

Edit: I forgot to mention how this system would most likely interact with the advantage/disadvantage system of 5e. Probably the way I'd do it is that you roll one extra die and take the highest/lowest dice of the correct amount. For example, making a cautious roll with disadvantage would roll 4d6 and take the lowest 3. This means that rolling cautiously would inhibit the effects of (dis)advantage, while rolling recklessly would amplify it.

[Eldan]

We're currently looking at a German system called Splittermond. It's for the most part terribly generic and I don't find it very interesting, but it has an interesting dice mechanic:

Players roll either 2d10 (normal) or 4d10 (risky).

There are critical failures, a result of 2 or 3, and critical successes, results of 19 or 20.

If you roll 4d10, you take the two higher dice, but if any two dice show 1s or a 1 and a 2, you fail. It makes criticals much more likely.


Perhaps something like that might work for you? Players can roll more dice, to take more risks?

There's something like an advantage system, too, though it works with fate points. Players can, amongst other things, spend one to reroll their second-lowest die.

[Yddisac]

This conversation has happened elsewhere on the board, and people with more time on their hands and a better grasp of probability than I have explained this better, but basically, retrofitting anything other than a d20 onto D&D is generally a bad idea (see, e.g., http://www.giantitp.com/forums/showthread.php?567531-Using-3d6-instead-of-d20). If you want different core mechanics, you're better off seeking different games than wrecking the balance of an entire system built around one core mechanic by swapping it out for another.

You've also pointed out several issues that changing the core mechanic can have. As it is, players can severely mitigate the effects of disadvantage at will. That means disadvantage is far less useful than it should be, and all mechanics that provide disadvantage are knocked out of balance. Exhaustion isn't so bad, and fear isn't such a good spell anymore. The probabilistic effects of +1 weapons, updates to proficiency bonus, etc. will swing wildly along with whichever roll players choose to use. There are other problems that others have addressed better than I.

As an abstract idea for a hypothetical game system, I'm still wary about the idea because of how wildly the probabilities would swing (it'd be that much harder for GMs and content creators to balance things), but I do like the element of choice involved. It's certainly a workable idea in a vacuum, and if you were developing a totally new game system, I'd be curious to see how it went. Retrofitting it - or anything besides the d20 the game was built around - onto 5e, however, is likely to cause more harm than good.

[Greywander]

If you roll 4d10, you take the two higher dice, but if any two dice show 1s or a 1 and a 2, you fail. It makes criticals much more likely.
This is an interesting idea, and I could probably work with it to enable you to both succeed on a check and trigger a "fumble", things like, "you hit, but now your weapon is stuck in their armor." I'm not sure how it would work with (dis)advantage, though. Maybe a standard roll would be 2d10 and a risky roll would add 2d8 or 2d12, so with (dis)advantage I could just add another d10 and take the highest/lowest two d10s.


As an abstract idea for a hypothetical game system, I'm still wary about the idea because of how wildly the probabilities would swing (it'd be that much harder for GMs and content creators to balance things), but I do like the element of choice involved. It's certainly a workable idea in a vacuum, and if you were developing a totally new game system, I'd be curious to see how it went. Retrofitting it - or anything besides the d20 the game was built around - onto 5e, however, is likely to cause more harm than good.
I followed that thread and read The Angry GM's article. He does good work. Yeah, I agree that it would be very tricky trying to fit this to 5e as-is, and what I had more in mind was to create a spin-off system that was merely based off of 5e, kind of like Pathfinder was to 3.5e. (Although, let's be honest, it's a long shot that I would ever complete such an endeavor, but sometimes you just have to work things out, or else they bounce around in your head forever.) In such a system, I'd probably want to balance it around rolling 2d10, while also understanding how 1d20 and 3d6 would interact with the system.

For (dis)advantage, another option is to have it always add an extra d20, and then you take the better or worse of the two rolls. So if you roll cautiously with disadvantage, you would take the worse of 1d20 or 3d6. Although again I'm sure if someone worked out the math, they'd probably find that it's always better to roll recklessly or cautiously when you have (dis)advantage, again probably depending on the target number needed.

I'm not married to this idea, I just wanted to see if it was possible to give the players a bit more choice. It just seems like players have no control over whether they succeed or fail at a task, they can only choose which tasks to perform and then let the dice decide if they fail or succeed. Conceptually, it seems to fit, but you're right that it will be kind of spotty to retrofit it to a system designed for 1d20. That said, you still can roll a d20 if you want to.

[Mangles]

I don't think the 2d10 option is good (or rather it is too good) as you never risk anything for the reward. However allowing 3d6 or 1d20 in certain situations is interesting. Especially for characters that take advantage of critical hits. Orcs and Champion fighters will risk a lot by losing crit chance. Conversely though any class that doesn't have extras to crit could always take an alt option to remove risk.

Perhaps only allowing this in certain cases would be a better idea. Like only on proficient skills, weapons, tools and never with disadvantage.

[Vogie]

I do like the idea of being cautious by using a 3d6, especially when players are being stealthy or relying on subterfuge in other ways... it reminds me of the "guaranteed success if you take 20 minutes to do so" mechanic in other systems.

But trying to break up a normal roll into smaller pieces seems off.