Why I Hate Skills

[Thomas Shey]

I wasn't talking about only D&D 5e, just using the dominant game system as the example. But obviously you can craft scenarios for a game where the statistical difference would be obvious.

I'd argue its obvious in more games than not once you get away from the D20 and D100 systems. Unless the dominance of the D&D sphere makes the rest of those irrelevant, I don't think this is so much "crafting a scenario" as noting that its a point that makes no sense throughout much of the hobby.

 

[Bill Zebub]

As I noted, bonuses on 3D6 systems are not symmetrical throughout the range, and as such can end up playing a great degree of significance in a situation. For example, a +3 modifier on a roll-low 3D6 system like Hero or GURPS can turn an 8- roll into an 11_ roll. That may not look like a lot at a glance, but it turns a 25% roll into more than a 50% one. Its not nearly as big a difference at a high end (a 14- roll to a 17- is not nearly as significant. But you spend a lot more time wandering around the range where the impact is much stronger.

Yes, as I said upthread, to accomplish the same kind of statistical bias in dice pools the software would have to 'weight' the d6's, not change the number of dice or the target number.

 

[Bill Zebub]

I made a mistake in calling out 'discrete' skills. Really what I was thinking, and mis-described, was systems where the differences are statistically smaller. The archetype I had in mind was D&D, where having a skill typically means a 10% to 20% advantage across the levels at which most people play. But discrete skill systems could of course have more variance, and non-discrete approaches (backgrounds, attribute based, etc.) could be smaller. Apologies. Thanks for the pushback.

 

[VHawkwinter]

I am all for 100% success at some things. That's how I play Shadowdark, for example: a thief is always going to pick a lock unless it's an unpickable lock or if they're trying to do it in difficult conditions (during a fight, for example).

That sounds good. Does Shadowdark have an explicit list of these things you can do without rolling, based on some list of character traits or something? I've never played Shadowdark.

I like to hand out explicit and growing lists of autosuccess abilities as part of character creation (handled through skills). When I run/play 3.x, those lists need to be made by making notes off of the skills chapter relative to your skill bonuses, but in my system I intend to have it be both a viewable screen attached to your character sheet (like your personal spellbook which will include all the spelsl you have and none you don't) and also have it automatically checked against when you go to attempt a task, by telling you what difficulty benchmark of task you autosucceed at in that category when you click to roll in some user-friendly way; and if you autosucceed then instead of logging a roll and bonuses to the group task-check logs it will log that you autosucceeded at x type of check with a DC of y or below, based on your character sheet bonus of (whatever) to the group task-check logs (because it's being built as an app, and will integrate as many of the non-maps-and-minis VTT convenience features as I can manage).

 

[Thomas Shey]

Yes, as I said upthread, to accomplish the same kind of statistical bias in dice pools the software would have to 'weight' the d6's, not change the number of dice or the target number.

At which point I think people would have to be less sensitive to low rolls (since I'm assuming they can still see the totals) than any set of players I've ever seen. Its one thing to not be able to tell a +2 from a +5 in a D20 roll; its a very different thing pretty quickly to not tell it in a 3D6 roll. It doesn't take long for people to tell that about half the rolls they should be seeing should be 15 or better, and if they don't see that they're going to start having questions. There's just not enough swing in there to not notice.

 

[Bill Zebub]

At which point I think people would have to be less sensitive to low rolls (since I'm assuming they can still see the totals) than any set of players I've ever seen. Its one thing to not be able to tell a +2 from a +5 in a D20 roll; its a very different thing pretty quickly to not tell it in a 3D6 roll. It doesn't take long for people to tell that about half the rolls they should be seeing should be 15 or better, and if they don't see that they're going to start having questions. There's just not enough swing in there to not notice.

I really don't think the number of dice matters. A 15% bias is a 15% bias.

What's tricky with 3D6, or any bell curve dice pool, (and why it could only be a computer) is that the number of dice that need to be biased, and the amount of the bias required to approximate any given %, varies with the target number.

 

[VHawkwinter]

I made a mistake in calling out 'discrete' skills. Really what I was thinking, and mis-described, was systems where the differences are statistically smaller. The archetype I had in mind was D&D, where having a skill typically means a 10% to 20% advantage across the levels at which most people play. But discrete skill systems could of course have more variance, and non-discrete approaches (backgrounds, attribute based, etc.) could be smaller. Apologies. Thanks for the pushback.

FYI, in 3e when playing with point buy stats, the range between "expertise" and actively bad at it at level 1 can be like from a -1 (Ability Score at 8) to potentially a +11 (Ability Score 20; 4 Ranks; +3 from Skill Focus; maybe +2 from Racial Skill Bonus) {I misremembered when I said +15 and was including an instance of aid another and masterwork tools, which you probably cannot afford until level 2}.
Then by level 3 you're working with a -1 to a potential whopping +23 (Ability Score 20; 6 Ranks; +2 Masterwork Tool; +3 from Skill Focus; +2 from another Skill Feat;You can afford to buy (or craft) a +1 Skill boosting item if built by the DMG magic item prices (The Expanded Psionics Handbook has a bunch at +5 and +10 that exactly follow the price formula, and the DMG has a handful as well, and you can , but the published examples are only in +5 or +10); Maybe a +2 Synergy Bonus from another skill; maybe +2 from Racial Skill Bonus).
(That is a player who REALLY wants to be good at that skill, and has invested as much as possible at that level into that one skill, but it is attainable - I have a spreadsheet for this, I use a table based on it to help me when I need to improvise DCs by the tasks at which they are level appropriate or 'theoretically possible to achieve').

And level 3 is certainly "across the levels at which most people play", without getting to the higher-but-not-epic levels at which my campaigns are usually built around (6 to 15). (Those +10 Skill Boost items become affordable (<15% WBL) to craft at level 9. level 12 to buy). {I'm inclined not to allow higher skill boosters than +10 even though the formula covers it, just because there are no comparable out-of-the-book items that give more than a +10.}

(The advancement rate after level 3 slows sharply, but those first three levels quickly give you access to most of the different types of skill bonus and a big chunk of your total attainable bonus by level 20. After that point it's basically just +1/lv in ranks, a skill boosting magic item which eventually caps out at +10, ability score boosting equipment, and tomes of permanent ability score boosting).

I believe 4e also had pretty quickly scaling bonuses for skills, but 4e is not my wheelhouse.

5e and 5e derivatives certainly have smaller differences in capability between characters of course (a significant thing I hated about 5e), but I just wanted to point out that your example doesn't hold across D&Ds, specifically not across the three D&D editions I would choose to play / run (3.0, 3.5, PF1).

 

[Milieu]

That sounds good. Does Shadowdark have an explicit list of these things you can do without rolling, based on some list of character traits or something? I've never played Shadowdark.

There's not an explicit list, but it says part of its core ethos is "The characters automatically succeed at what they are trained to do. Only use stat checks when there is time pressure and failure has dire consequences." Of course, there's almost always some time pressure in Shadowdark, as another part of is core ethos is that time "must haunt the characters' every decision."

 

[Thomas Shey]

I really don't think the number of dice matters. A 15% bias is a 15% bias.

I think you're making a mistake thinking its going to only be a 15% there. As I noted, a 15% lands very differently in the point in the curve. At some levels people won't notice it because it won't even be a point on the roll. At others, it would be most of the range.

Again, I think the visibility is going to be vastly different. I don't think you can weight the dice there without it being pretty obvious over time in some circumstances. There can be situations where you're rolling in a particular low probability range for an extended period (or the opposite). You might get it to pass in the middle for a while, but the middle isn't the only thing that happens there.

 

[VHawkwinter]

There's not an explicit list, but it says part of its core ethos is "The characters automatically succeed at what they are trained to do. Only use stat checks when there is time pressure and failure has dire consequences." Of course, there's almost always some time pressure in Shadowdark, as another part of is core ethos is that time "must haunt the characters' every decision."

Got it. I agree with that principle, but that's a little wishy washy for me. I want more specifics. But good to know.