Ilbranteloth
Explorer
Would you prefer the 3E model? That the modifier could easily be +20, such that anything an expert could possibly fail at (DC 22, in this example) might still be entirely beyond the capability of someone untrained? There were plenty of complaints about the 3E model, where the specialist was the only one who even had a chance.
How could you possibly address this issue, without breaking from the 1d20 mechanic entirely? Reliable Talent is one approach, but people don't seem very happy about it. What are some alternatives?
Really, I think the system as designed addresses most of the issues, with a few minor modifications. I don't think I explained my thoughts very well in my earlier post, so:
1. Your passive score essentially provides a floor for your skill ability. That is, most of the time, below this threshold you don't significantly risk failure without other mitigating circumstances. That doesn't mean you can't, but most of the time you probably won't need to make a check if the DC is less than this. This is a number that can help your consistency when determining whether a check is needed or not.
2. When making a skill check, if you have proficiency, then you know you can succeed, it's really a question of how long it might take and if there are any complications. Most of the time you can use the difference between the DC and the roll as a measure of time. This eliminates the need for retries. I also consider other complications/consequences if you fail by 5 or more, 10 or more, 15 or more, etc.
3. If you are non-proficient, the check is closer to a success/fail check. Again, retries generally don't work.
4. I usually don't allow a non-proficient character to attempt something that has a DC of 20 or higher. This isn't a hard and fast rule, but a guideline - if the task is a complex one and the DC is 20 or higher, then I think you'd need actual training to succeed at it. For example, if you claimed to be a minstrel, and were asked to play a particular piece that's a favorite of the king's on the lute, then no amount of luck will allow an untrained individual to succeed. That doesn't mean that they might not be able to find another way to bluff their way through the scene, they just won't be able to play the piece.
5. Rolling a check requires an action, generally above and beyond the normal circumstances. So walking through a forest is usually a passive Perception check, and the checks are based solely on the passive score. Going slowly and keeping your eye out for trouble? A passive check with advantage, and I'll usually take into account your full capability (20+). For a Perception check this means you'll probably get a sense something is wrong, (the "hint" approach), but will need to do something to make a check. Climb a tree to get a better vantage point? Then I might ask for you to roll.
6. The DM asks for a check dependent upon the actions the character takes. In many cases, the action resolves the question itself with a success, or with no chance for success, so no roll is needed. If the character has no chance for success, there's no point in a roll.
7. Advantage to a passive check (+5 modifier) does not increase your capability, only your chance for success. That is, your measure of capability is 20 + ability and/or proficiency modifiers, not circumstantial modifiers.
--
There are a number of factors as to why I prefer to do things this way.
1. I want to eliminate re-rolls. Not by an arbitrary rule that says "no rerolls," but because the rolls already take into account that you will continue to retry until successful, unless there's some other factor.
2. I want to reward proficiency. The chance of outright failure is lower with proficiency, although the circumstances matter here too.
3. I want to better define the guidelines of when to roll. We already hand-wave all sorts of skill checks. By basing most rolls off of passive scores there's a guide to work with. You can still opt not to make a check even when the DC is higher than the passive score. I do that all the time, because their score might be increased due to advantage, but frequently because there is no consequence to taking some extra time. This directly ties into the frequent advice given about when to ask for a skill check, but utilizes the passive score to help with that guideline when needed. I also take into account the character's full capability (20 + modifiers) here.
--
Now we have altered the math in a few ways because we feel that 5e makes things too easy. That's our personal preference. As mentioned before, I think that most of the published DCs are too low. It's not a hard and fast rule, and is largely irrelevant since we don't use the publishes APs. But when people complain about expertise and reliable talent, particularly in relation to passive Perception checks, this is the first thing I'd consider. Again, the detection DC for traps in ToH seems way too low for a dungeon of that nature.
We also felt that the idea that somebody untrained could be significantly better than somebody trained in some cases, and also that the difference between no modifiers and expertise got too wide. In addition, the way proficiency is defined is inconsistent across different aspects of the game. Using this for skills only is simple, using it across the game like we did is more complex.
So we use the following system:
Non-proficient: Ability modifier, with a maximum bonus of +1 (we'll actually probably use a maximum bonus of 1 less than proficiency at a given level).
Proficient: Ability modifier or proficiency modifier, whichever is higher.
Expertise: Ability modifier plus proficiency modifier.
--
For reliable talent itself, it's more or less redundant for me since it's essentially the same as a passive score. The complaint still seems like one of degree of benefit. That people don't like how many things it makes an auto-success by rule. That, to me, is a math issue. Either reliable talent should work differently, such as a +5 bonus (which is essentially advantage, although it could stack with advantage) would probably do the trick for most people.