Let's Talk About Core Game Mechanics

[Mustrum_Ridcully]

true, but I am not aware of any games where the average result leads to the more unusual outcomes. You either hit or miss, depending on the target value, but you neither critically fail nor critically succeed, so it generally will be among the most 'normal' and definitely most common outcomes possible.

From my understanding that was the OP's complaint, that a bell curve makes the more exciting / interesting stuff less likely to occur compared to a linear distribution.

One thing I am considering for 2d10 or 2d6 roll systems: If you get the same number twice, it's considered some kind of crit.
Whatever that would mean, exactly, remains to be decided. (Of course, that still means the highest number is the best, obviously, but you might get a more spectacular result with a 16 on a 2d10 then with a 19.)

 

[Bill Zebub]

I would say that is a very common rule in general, now. Even GURPS says not to roll Driving skill for a trip to the shops, and with the now-common design concept of (generalising) “don’t roll a skill unless there is an interesting outcome for both results“ it seems to be becoming a default in GM advice (and good advice to be sure).

It's the "both" part that often gets missed. And, admittedly, is challenging for the GM. E.g., knowledge checks.

 

[Micah Sweet]

Sure, you could make a game where the skill values overpower the RNG in most cases. Just taking some numbers off the top of my head, you could have a system where a novice has a skill value of like 10, a master has 50, and you roll d10+skill to see if you succeed. Maybe even an exploding d10 to reflect that you might accidentally luck into doing the right thing. This would provide a fairly narrow window of difficulties where random chance determines if you succeed or not, but in most cases you'd know ahead of time if you're good enough.

Bell curves accomplish a watered down version of this. Let's say we're playing GURPS (3d6, roll under or equal to skill) and you have skill 13 and I have 15. On a normal task, your chance is about 84% and mine is about 91%. We're both pretty good at our stuff and fairly likely to succeed. But let's say the task is difficult and has a -4 penalty. Now you're down to 37% and I'm at 62%. I'm still not certain to succeed, but my higher skill helps me absorb more of the increased difficulty. Now, I'm not saying GURPS has the perfect skill system – far from it, the skills are too narrow, the costs are too high, and it's far too generous with penalties for everything. But that particular bit is pretty nice in reflecting how high skill has diminishing returns on normal tasks but is very helpful when things get difficult.

I think I've mentioned it before but I'd have loved to see a version of Pathfinder 2 where proficiency level (trained, expert, master, legendary) gates automatic success rather than allowing the attempt in the first place. They do have the Assurance skill feat you can take, but since that only counts proficiency bonus and not stats, items, and other buffs it's about equal to guaranteeing rolling a 5 or so.

I prefer narrow skills provided the costs to acquire them are reasonable. It's all about setting logic to me.

 

[...m...]

One thing I am considering for 2d10 or 2d6 roll systems: If you get the same number twice, it's considered some kind of crit.
Whatever that would mean, exactly, remains to be decided. (Of course, that still means the highest number is the best, obviously, but you might get a more spectacular result with a 16 on a 2d10 then with a 19.)

...i've wondered similarly with three-die systems and whether it's worth separating critical effects from overall success or failure; i.e. two 1s or two 4/6/8/+ (depending upon dX rolled) manifest bane or boon effects regardless of overall success or failure, but only triple 1s or triple 4/6/8/+ present critically-rare automatic failure or success...

...this would also offer the balance of making boons/banes and critical rolls more likely at untrained levels (3d4), which taper off in favor of predictable success at greater levels (3d6/8/+)...

...for context, the system i've been noodling with essentially uses 5-30 bounded-accuracy DCs (similar to 5e), with 3dX chains serving as a proxy for proficiency: 3d4 (untrained), 3d6 (proficient), 3d8 (expert), 3d10 (master), and 3d12 (legendary) for PC career progression with 3d16, 3d20, 3d24, and 3d30 reserved for escalating monstrous/demi/planar NPC power levels...

 

[Bill Zebub]

...i've wondered similarly with three-die systems and whether it's worth separating critical effects from overall success or failure; i.e. two 1s or two 4/6/8/+ (depending upon dX rolled) manifest bane or boon effects regardless of overall success or failure, but only triple 1s or triple 4/6/8/+ present critically-rare automatic failure or success...

The TOR system has an interesting take on this: While a Gandalf rune on the d12 is an automatic success regardless of dice total, an Eye of Sauron rune on the d12 does not mean automatic failure: the d6's could still total enough to make the TN. But....the EoS means you increase your "Eye Awareness" score (basically the Shadow's awareness of your presence/shenanigans).

 

[Thomas Shey]

I only meant that the claim Skill 80 = 80%, how easy to understand is that? is either from a system that doesn't take those modifiying factors into account...which seems overly simplistic/boring...or it does take them into account in which case the final percentage isn't 80% and it's not so easy to understand after all.

I guess. I just don't think "your skill is 80 and you have a 10% penalty because of X" is notably harder to understand there.

 

[Thomas Shey]

I prefer narrow skills provided the costs to acquire them are reasonable. It's all about setting logic to me.

The problem with narrow skills even from a realism POV (and admittedly GURPS tries to address this with defaulting) is many skills are either closely related, or are extremely unlikely to pick up at a decent level without picking up certain other skills. As a hypothetical example, you could have "Slashing Swords" skill. Now, there are things about usage and balance there that means your skill there isn't going to entirely transfer over to other weapon usage. But it isn't going to be completely independent of them, either; you probably could carry over a lot of that skill to, say, axes, and there are even some elements of your training and practice that will transfer over to something like spears, even though the balance and usage is quite different.

So if you want to do it right you either have to do a middlin' complex interrelationship thing in your skills, or shrug and either under or overstate how these things connect by your choice of skill lumping or splitting.

 

[Bill Zebub]

I guess. I just don't think "your skill is 80 and you have a 10% penalty because of X" is notably harder to understand there.

Sure, but I also see barely any daylight between that and "DC is 14 and you have a +2".

I think the original implication was that people have a better intuitive understanding of odds expressed as percentages rather than ratios (e.g. 40% instead of "8 out of 20") but in my experience people suck at genuinely understanding either.

 

[Thomas Shey]

Sure, but I also see barely any daylight between that and "DC is 14 and you have a +2".

I dunno; you still have to do some addition during the resolution process (i.e. roll the D20, add the +2, see if it makes the DC). The roll-low percentile case you can do what math there is before the dice get engaged and just see if you made equal or under. I think that's sometimes a non-trivial difference.

I think the original implication was that people have a better intuitive understanding of odds expressed as percentages rather than ratios (e.g. 40% instead of "8 out of 20") but in my experience people suck at genuinely understanding either.

Its a defensible argument. Though at the point someone doesn't have an intuitive sense for what "60%" means, they don't have a relational sense of any odds at all, so its moot.

 

[Bill Zebub]

The problem with narrow skills even from a realism POV (and admittedly GURPS tries to address this with defaulting) is many skills are either closely related, or are extremely unlikely to pick up at a decent level without picking up certain other skills. As a hypothetical example, you could have "Slashing Swords" skill. Now, there are things about usage and balance there that means your skill there isn't going to entirely transfer over to other weapon usage. But it isn't going to be completely independent of them, either; you probably could carry over a lot of that skill to, say, axes, and there are even some elements of your training and practice that will transfer over to something like spears, even though the balance and usage is quite different.

So if you want to do it right you either have to do a middlin' complex interrelationship thing in your skills, or shrug and either under or overstate how these things connect by your choice of skill lumping or splitting.

100% agree.

All RPGs are such terrible approximations of reality that it's hard to take seriously any claim of "realism" or "logic". Sure a 0.001% approximation is more realistic than a 0.0005% approximation. But....really? If somebody enjoys fiddling with the additional complexity, then have at it! Enjoy! But let's not kid ourselves that any of these are actually realistic.