D&D 5E Overcoming Bounded Accuracy

[slobster]

If an incompetent blacksmith has a 10% chance to make a horseshoe (not just getting a lucky 20, succeeding just because it has a low DC), why does the competent blacksmith still have a 40% chance of failure?

In that specific example I'd allow the blacksmiths to take a 10, thus solving that problem, but in general I see your issue with the system.

It's worse if we take two characters with the same attributes. Their skill training is irrelevant unless they're rogues. Depending on the skill and your play style, a change in one out of seven rolls might matter once a session or may not come up during the whole campaign.

Which comes back to the question of why we're tracking skills at all. I like skills and think they add something to the game, but they don't provide enough of a mechanical benefit to distinguish characters.

You could have skill training for everyone work like skill mastery for rogues (minus the +3 stat floor). It would mean you have to give some new presents to rogues, but would it be a step in the right direction, in your opinion?

 

[RangerWickett]

Part of the D&D Next design is that you don't make people roll for things that are easy. I mean, you don't need rules to scale probabilities by level for something like that. Just use common sense.

"I want to make a horseshoe" should not require a roll from a competent smith. I mean, maybe you call for a check, but it's to determine "is it a normal horseshoe, or an excellent horseshoe?" Failure is not an option.

"I want to run along that tightrope, tuck into a ball, and smash through the window" should not require a roll from a high level rogue. I mean, maybe a check is the difference between "you land normally, or you get a bonus to your attack roll because that was awesome." Failure is still not an option.

"I want to give a speech that rallies the nation to take up arms against the demon lord" should not require a roll from an epic level bard. The check determines whether they just fight with swords, or if you manage to create some sort of psychic gestalt to temporarily empower you as the avatar of the nation's willpower.

 

[slobo777]

That's absolutely true but, to me, just speaks to the problem. The attribute is the more important part and even that gets drowned out by the d20 in normal cases.

I feel that, if we're talking about the difference between your competent and incompetent character, one of them should be forced off the d20. If an incompetent blacksmith has a 10% chance to make a horseshoe (not just getting a lucky 20, succeeding just because it has a low DC), why does the competent blacksmith still have a 40% chance of failure?

It's worse if we take two characters with the same attributes. Their skill training is irrelevant unless they're rogues. Depending on the skill and your play style, a change in one out of seven rolls might matter once a session or may not come up during the whole campaign.

Which comes back to the question of why we're tracking skills at all. I like skills and think they add something to the game, but they don't provide enough of a mechanical benefit to distinguish characters.

Cheers!
Kinak

A D20 check models exciting, random, combat-like things ok. Where speed is of the essence, and/or anyone could get lucky, or there is lots of distraction.

I would advise not using it for many other situations.

For instance, in an arm-wresting competition, an 18 strength should beat a 16 strength nearly always.

In any sport-like contest, the better athlete or acrobat will nearly always win. In the recent Olympics, many competitions, especially those involving solo athletes, were won by someone who was a clear favourite to start. Their current form was known to the nearest second, no room for something like D20+Dex or D20+Str opposed rolls.

As a DM, when you think you have that kind of situation, there isn't a huge amount of advice printed so far beyond "Auto success when you think it is appropriate". But there are some things you could do:

1) Threshold skill, below which you cannot attempt an action, or you have disadvantage.

2) "Needs 2 successes out of 3" for something that might take a while.

. . . probably plenty of others too. I wonder if anything like this will make it into the published rules?

 

[pauljathome]

So, assuming that I don't have idiosyncratic tastes, and there happen to be a significant number of gamers like me, what modules or simple rules changes could be added or made to remove bounded accuracy?


Any other thoughts?

I think that this is so inherent to the rules that it isn't worth house ruling. The odds of getting a decent and balanced game just seem too low to me. You'll have to tweak all NPCs, all PCs, all modules, etc.

My solution is the obvious one. Either decide that this is a deal breaker and ignore DndNext, or decide it isn't a huge deal and go with bounded accuracy.

There are a LOT of D&D games out there. If DndNext isn't the closest one to your tastes, start with a different one.

 

[RangerWickett]

How much better do you want high-level characters to be than normal folks? Is 5th level Bruce Willis in Die Hard, 10th level is Arnold Schwarzenegger in True Lies, and 15th level Leonidas in 300? 20th level is, I dunno, Neo at the end of The Matrix?

It depends. What are their attack bonuses?

Is the number all that matters? I mean, if the rules effectively capture the aesthetic of Neo being tougher than Leonidas -- his attacks are more effective, it's harder to injure him, and he takes out his foes more quickly -- do you still need big numbers?

In Game System 1, Neo has a +10 attack bonus and 20 AC, hits for an average of 50 damage and has 200 HP. A mook has a +5 attack bonus, AC 10, hits for an average of 10 damage, and has 20 HP.

In Game System 2, Neo has a +20 attack bonus and 30 AC, hits for an average of 50 damage and has 200 HP. A mook has a +15 attack bonus, AC 25, hits for an average of 10 damage, and has 20 HP.

In Game System 3, Neo has a +20 attack bonus and 30 AC, hits for an average of 5 damage and has 20 HP. A mook has a +5 attack bonus, AC 10, hits for an average of 1 damage, and has 2 HP.

Which system do you like best?

In each system, Neo can easily kung fu his way through mooks, and isn't really intimidated unless there are 20 or more mooks.

In system 1, John McClane can easily take on one mook. Maybe a mook is a guy with a gun in both games.

In system 2, John McClane's attack bonus is way too low to hit a mook who can threaten Neo. So McClane deals with terrorists while Neo fights magical super terrorists who can dodge bullets and, even when they are shot, can ignore the first couple hits.

In system 3, John McClane still can't hit the super terrorist, but if he gets lucky, he can take him out in one shot.

Which system models the gameplay style you want?

 

[slobster]

I think that this is so inherent to the rules that it isn't worth house ruling. The odds of getting a decent and balanced game just seem too low to me. You'll have to tweak all NPCs, all PCs, all modules, etc.

It should be an easy tweak though, right? Add half the thing's level to attacks, defenses, and skills. Done.

As we've said, skill DCs need some slight adjustment, but that whole thing has always been the realm of GM judgement, so it's not really a major difference in how the game works, just a minor adjustment in the calls the GM routinely makes.

I mean, I'm really excited about bounded accuracy, but this sort of house rule seems pretty easy to effect.

 

[Fazza]

I personally like bounded accuracy, I love 4e but one gripe was that higher level monsters couldn't be touched by lower level ones and that my paladin who had no interest in history somehow got better at remembering stuff from hundreds of years ago. I understood the concept and I can understand people liking it I think every 2 levels is a bit much I would prefer it was more like every 5. I fully expect this to be part of a module though for people who like it.

Also I agree with Kinak that +3 isn't enough but I expect to see skills become a bit more complex in future playtests and eventually modules for varying degrees of complexity.

 

[I'm A Banana]

The idea of adding "levels" to challenges other than monsters is actually a good idea from a "Three Pillars" standpoint, too.

How hard is that Dragon to convince? How hard is this trap to avoid? How easy can one get lost in this wilderness? How intractable are the Guardian Spirits of the Tomb of the Elders?

While I think I like bounded accuracy (though it's still gotta sell me entirely), reagardless of that, levels (and thus XP totals) for challenges that are not murderin' things is a Good Idea to me.

 

[Steely_Dan]

If the end result is the character needs to roll an 11 on a d20 to hit, does it matter if the bonus is +7 or +27?

"...but...but this goes to eleven..."

 

[Steely_Dan]

Removing the 1/2 level bonus shenanigans from all character's and monster's Attacks, Defences and Skills really helped my 4th Ed campaign.