D&D 5E Do you use the Success w/ Complication Module in the DMG or Fail Forward in the Basic PDF

[iserith]

1e was the first place I saw it, in the write-up for bend bars/lift gates*.

The problem of allowing infinite retries of the same approach - i.e. lots of dice rolling - until a high-enough roll comes up.

Not sure it can, as the one-roll solution is already about as elegant as it gets. 3e tried going the other way and assuming a 20 would eventually occur, with the take-20 mechanic. This, while admittedly just as elegant, brought about the (IMO unacceptable) side-effect problem of every task becoming a purely binary you-can-do-it-or-you-can't setup: any uncertainty was removed.


* - I originally mistyped that as "bend bard/lift gates", which brings to mind a few interesting visuals...

If there is no cost to retrying other than time and the task isn't impossible, then there's no reason to roll. Just narrate the result.

The "one and done" approach is inelegant in my view because it doesn't seem to be well-rooted in the fiction. To the extent that I recall the conversation, @Charlaquin showed why that is way upthread. While I'm sure you've justified it in your mind these many years and are used to it, I find it to be an odd fit for someone seeking "immersion." As well, the uncertainty as to the outcome occurs before the roll as established by the DM. The roll resolves the uncertainty into success or failure (or progress combined with a setback). I'm not really sure what you mean by a binary in this case or why that is bad. That's just how it is in D&D - you win or you lose or you win with a setback.

 

[Lanefan]

Indeed, accounting for all sorts of factors external to the raw difficulty of the task and your character’s capabilities.

That’s just not true unless you believe in pure determinism. If you roll a 7 on a d20, there are objectively 19 other numbers you could have rolled, 13 of which would have been higher.

Of course there are.

But so what? The only thing that matters is what you did roll; and if you want to beat yourself up over what you could have rolled then have fun with that, but for me what's done is done.

Just like handing in a school exam: you can beat yourself up afterwards over the questions you should have got right but blew; or you can say screw it, it's done now and can't be undone, and move on.

 

[Lanefan]

Side note:

How so?

I really wish the new forums had a function to include quoted text when quoting someone.

The hard way to do it is to cut-and-paste the previous quote and put it in a quote block above what you're quoting now, as I've done here.

Tedious, yes, but functional.

 

[Maxperson]

To be fair, this is a great argument for either using 2d10 or some other multi-die setup rather than a single d20, or making most tasks use multiple checks to determine degree of success/failure. The d20 is much too swingy to model reality well.

I agree, but I've come to love d20 too much to want to switch.

 

[Lanefan]

If there is no cost to retrying other than time and the task isn't impossible, then there's no reason to roll. Just narrate the result.

This is my big beef here: "the task isn't impossible" does NOT always equate to "given long enough, success is guaranteed".

Sometimes, even though you in theory should be able to do it you just ain't gonna get it done in practice, no matter how long you give it; further, there's always a chance that someone with in-theory-lesser abilities is gonna come along and embarrass you by showing you how it's done. Fact of life.

 

[iserith]

This is my big beef here: "the task isn't impossible" does NOT always equate to "given long enough, success is guaranteed".

Well good, because that's what the rules are saying since it carves out exceptions for when, in the fiction, the same approach cannot be tried twice. The example given is trying to use the same lie again after failing to do so. That task now becomes impossible because that makes sense in the fiction and thus becomes the exception to the rules on retries. But with the "one and done" method, it's more like "you can't possibly succeed now because... uhhhh... I don't like retries so that's the best you can ever do ever?"

 

[Charlaquin]

No! The best effort was not less than it could of been. Best = best. Period. Only the roll was lower than it could have been, which is irrelevant if you only get 1 roll.

If you can’t see the obvious contradiction between “the roll represents the best your character can do,” “the roll could have been higher,” and “the best your character can do could not have been better,” I don’t know how to carry on this conversation. It’s self-evident that something here isn’t adding up.

Let me rephrase. None of that matters objectively after the roll. Objectively, all the other numbers are irrelevant to how well you did or could have done on that roll. Subjectively, yes, you care about it and it matters to you.

Sure. The fact that you could have rolled higher doesn’t matter for the narrative or mechanics under the “one roll represents your best attempt” model. And that is specifically what bothers me about it.

Sure. My point is that binary isn't inherent to the system. It's not just pass/fail.

Ok.

 

[Charlaquin]

I really wish the new forums had a function to include quoted text when quoting someone. I keep having to scroll up and see which part someone is replying to.

I hate the way the quote function works now in general. Every post is a battle with the forum software to make it work right.

So, this is in reply to “reduces room for improv”. I allow the PCs to create or change aspects of the world via improvised storytelling during play. This ties into what I’ll talk about wrt “volcanic” worldbuilding, but basically, I only decide the general difficulty of a thing ahead of time. I often don’t even keep an NPC statblock as written if what the PCs come up with changes what I want from the encounter. So, because I haven’t decide exactly the inner workings of a lock ahead of time, I can decide in response to a bad roll, that it is a trick lock and they need to investigate it in order to unlock it, or I can simply say, “it’s odd, you feel like it should have worked.” And then the players, directly or indirectly, determine that it’s a trick lock by speculating about it and then investigating further. Then, when they roll a high investigate check or arcana check or whatever, the nature of the lock and what stumped them initially is revealed, either allowing a newcheck or even negating the need for another check.

likewise, I might say, it’s an odd lock, how do you approach it? And their response might lead to a tools check, an investigate, and an arcana, and which rolls succeed or fail might determine, with input from the players, what is happening in the fiction.

Gotcha. Yeah, I’m familiar with that style of play, D&D just isn’t the game I go to for it.

I imagine that is both due to our differing preferences and their effect on how a given resolution system plays out, and a driver of why we prefer different things. The causality of preference is loopy like that, IME.

For sure.

Not just that, but that is part of it. Rather it is about building enough of the world for it to feel complete and alive, but leaving room for your ideas about the world to change, and being willing to throw out pages or even chapters of information you’d prepared because soemthing the PCs said or did made some other idea into a much better idea for this particular campaign.

That means that in my Eberron campaign, there is a noble house in Breland and Thrane that is engaged in a shadow-war with The Twelve in order to break their monopolies, and it means that I don’t actually know yet whether Erandis Vol is actually a villain or not, and even if I did “know”, my answer could change next session when I get a strong indication that my BoV Paladin PC is looking for a tragic story to redeem some bittersweet ending out of, or is looking for an enemy that is visceral to her story and who she can just brutally murder without remorse.

Gotcha. Yeah, I prefer a setting bible approach. Leave room for future development, absolutely. But don’t change things once they’re established.

IME, avoiding front loading creates a smoother player experience, as well.

Not in mine. Certainly I don’t find the “one roll represents your character’s best effort” approach to action resolution to be smooth by any definition of the word.

Whereas for me, D&D, especially 5e, runs just as smoothly either way.

I’m not saying D&D can’t run smoothly in what you call a volcanic style. Just that I would rather use a different system that is built for that style if that’s what I’m looking for.

 

[Charlaquin]

To be fair, this is a great argument for either using 2d10 or some other multi-die setup rather than a single d20, or making most tasks use multiple checks to determine degree of success/failure. The d20 is much too swingy to model reality well.

Only if modeling reality is your goal. It’s not mine, and I find the swingyness desirable. Remember, the way I run it, rolls are supposed to be dangerous things you want to avoid having to make when possible and shore up your chances of failure on when they can’t be avoided. A check or save is a gamble, it should have a fair amount of swing to it, and reducing that swing should require you to either play to your character’s strengths, consume limited resources like inspiration or spell slots, or both.

 

[Charlaquin]

1e was the first place I saw it, in the write-up for bend bars/lift gates*.

The problem of allowing infinite retries of the same approach - i.e. lots of dice rolling - until a high-enough roll comes up.

If infinite retries are possible (i.e. there is no meaningful cost or consequence for failure), the most elegant solution IMO is not to roll. You can assume they retry until they get it, skip the rolling, and narrate success. On the other hand, if there is a cost or consequence for failure, retries will naturally be limited by the player’s own ability and/or willingness to pay that cost or risk that consequence.

Not sure it can, as the one-roll solution is already about as elegant as it gets. 3e tried going the other way and assuming a 20 would eventually occur, with the take-20 mechanic. This, while admittedly just as elegant, brought about the (IMO unacceptable) side-effect problem of every task becoming a purely binary you-can-do-it-or-you-can't setup: any uncertainty was removed.

Uncertainty is only removed in cases where infinite retries would be possible - i.e. cases where there is no cost or consequence for failure, which in my opinion shouldn’t be uncertain anyway.

* - I originally mistyped that as "bend bard/lift gates", which brings to mind a few interesting visuals...