I'm on the cusp of either adding more attributes to my tabletop project, or ditching them entirely.

I think I'm starting to suspect that 90% of why I *have* attributes is my deep-seated belief that I can fix the D&D attribute approach.

But I keep finding places where I'm basically using cruft to let people use their best attribute, or removing the modifier.

And I've already made the decision that if you are trained in a skill, it removes the penalty for low attributes under the theory of You Are Good At The Stuff You're Good At.

So at this point, attributes mainly exist to give default modifiers for skills in which you are not trained, and things not covered by skills.

And as I'm currently using a d20 RNG, the impact of them is kind of fiddly.

If it seems like the answer should be obvious... the game kind of does need the space occupied by the attributes to exist as part of overall character definition. The kind of story I'm looking to emulate, these are characteristics that matter.

So I don't know. Maybe I'll just rethink using them on a scale based on D&D attribute mods and how they intersect with rolls, arithmetically.

Oooh. Here's an interesting idea.

So, if you're adding an attribute to a d20 roll and it's 3, that makes a 3 in 20 chance you roll where it's going to matter, right? And 5 is 5 in 20. And you quickly run out of spaces for attributes to matter because there's only so many numbers

Imagine if your attributes are not added to a roll, but instead if your d20 roll is equal to or lower than your attribute, you can re-roll and only have to take the second roll if it's better.

Your attribute still has roughly the same chance of coming into play, but one of the complaints about the d20 system's "plateau of chaos" is how frustrating it is to get super low rolls on the thing you're known for.

Which is not eliminated by this system, you can still fail!

You can have 10 in Strength in a world where the absolute strongest person most people will personally know has a 3, and you can still fail. But it's very unlikely.

I like this? Previously I've been trying to both avoid that aspect of the chaos plateau and yet preserve failure chance by having it so if you roll a 1 and still succeed, you take a Risk token, and you fail on your next 1 that would have succeeded, and that's fiddlier.

One problem with this is that it only incorporates attributes into the success rate, which means every character would be equally easy to hit in combat, but I already didn't like how attributes were being incorporated defensively.

(It was +attack attribute -defense attribute)

The reason defenses were subtracted instead of changing the target number is this game uses a range of results instead of a target number. Unskilled, skilled, and expert each have their own success table, with six results.

And I cribbed that from Strike! but put it on a chart of 20 numbers because I want higher levels of skill to impact the expand the range of numbers that give exceptional results, not just give more likely to succeed.

The six results are:

Red, fail with a cost or complication.
Orange, fail.
Yellow, partially succeed or succeed with a cost/complication, can choose to "fail gracefully" rather tahn pay cost.
Green: Success
Silver: Success with fringe benefit.
Gold: Success with better benefit.

So for example,

Maybe Unskilled gets red on a 1 - 3, 15% chance of making things worse when you try something not your wheelhouse. Gold on a 20 only.

Skilled gets red on a 1 or 2, Gold on a 19 or 20,

and Expert gets red on a 1, Gold on a 18 or 20.

Character sheet design would have the result tables printed on the bottom so everybody has them right in front of them at the table, obviously.

And yeah, a 10% chance of making things worse for a skilled practitioner is pretty high... but that's before accounting for attributes, which with attributes as an additive bonus would reduce that to 5% or 0% with a positive attribute, and I already had skills "wipe out" negative

So with the idea that attributes mean a chance to re-roll low rolls, an attribute of 2 (which would be most people's second best attribute) means you have a 10% chance of a 10% chance of making things worse when you do something you're skilled at, or 1%. Not impossible or likely!

And "neither impossible nor likely" is kind of where I want critical fumble type results to be, for the stuff you're good at? Like, oh. That would be MEMORABLE.

So this leaves me with the question of how to handle defense.

I've already toyed with the idea of a D&D-ish tactical system where the default assumption is you "hit" (using the D&D definition of a hit where most are glancing blows) and the question is how well.

I'm not doing that with this system, though the range of yellow to gold all counting as at least a glancing or awkward hit means hits will be well more common than misses.

So now I'm wondering... defense rolls? Which is the opposite of eliminating hit rolls, but I feel like players like a sense that they get a chance to do something about incoming hits. Minions, mooks, and anyone without defined combat abilities don't make them, save time.

Oof, though, on the other hand, few things would be more frustrating than getting a really good hit and watching it go away because of a game runner roll.

Also, if defense rolls are equivalent to hit rolls, that means it's about even odds that an average hit goes away, which is kind of grim?. But on the other hand, exceptional hits are exceptional so it would be less likely they'd be overturned.

Maybe I'll start from the standpoint that there is no attribute-based defense, defenses are always exceptional special abilities a la Uncanny Dodge

...defense is subtracted from DAMAGE ROLLS, negating a hit if the damage is reduced below 0, and ignored if the damage roll is maximized.

So a dagger doing 1d4 damage has a 25% chance of ignoring defense (good at slipping through armor and glancing hits with a knife still wound; "the winner of a knife fight is the one who dies at the hospital").

This is the kind of thing where no matter how much math you do, it's very hard to tell how it will play out until you play it. I have a big old note at the top of my combat rules to the effect that they're more experimental and work-in-progress than the rest of the game.

Mmm. Okay, I have doubts about this but I kind of like it? Another wrinkle: flexible weapons (whip, flail, chain, etc.) don't ignore defense on a max damage roll but you can layer another effect onto the hit. Disarm, entangle, move, hit the eyes, etc.

This would make whips (1d4 damage) very unlikely to seriously harm an armored opponent but exceptional for control.

And my sweetspot for how weapons intersect with the combat mechanics is to have reasons for weapon choice to be consequential so it's not just "this is the weapon my character can use that has the biggest dice".

To that extent, I frequently toy with the idea of doing away with weapons having individual damage values and making the difference in the narrative layer of combat (E.g., you can't wrap a greatsword around a person, you can't cleave a shield in two with a whip.)

....I think this is what I'm doing, at least for now? As I've said, combat is the area that requires the most playtesting to know if and how it holds together.

...it would make damage *very* swingy for heavily armored people and tough creatures, though.

But part of the reason for the narrative/effect layer of combat is to allow and encourage players to find other ways of dealing with tough foes than whittling down their HP.

Laying aside the defense question for now and going back to the re-roll based on attributes.... so if success outright, green or better, is a 10 or better for you... that's a 45% chance of not getting outright success.

And if you have a Strength of 10, which is a ridiculously high number for a mortal human being at this scale, then if you fail to succeed outright (roll a 9 or lower), you're re-rolling because it's less than your Strength, so your failure rate is 45% of 45%, or about 20%.

Your odds of winding up with a red result of 1 or 2 on any roll are still 1% (10% of 10%) because your attribute in excess of your roll doesn't affect things in any way.

20% chance of not succeeding outright is still interesting.

The one thing about this is that as you are not re-rolling unless your result is very low, there's 0 chance for an attribute to affect the *degree* of success. Below truly superhuman attributes, you'll never re-roll even a yellow result. Do I like that? Only skill affects degree.

One aspect of the system that is a reason I wanted to keep attributes at the outset of this thread is that being defined as strong in an attribute-level area in this system gives you a limited use ability to Invoke the attribute, a la, "I'm strong, so I do the strong thing."

I.e., there's a rare resource called Hero Points you can use to make certain narrative declarations, but your attributes give you a per-adventure/episode version of Hero Points under your control that must strictly relate to the attribute.

So if you are marked down knowledgeable, you can make a roll on your knowledge skills knowing you are somewhat insulated from the worst result, or you can 1 to 3 times per outing declare a good result.

I think I like this. I think I like where this leads.

This should be "but" rather than "or". You can mix both of those things freely. And I think under this model you should also be able to Invoke after rolling if the result isn't what you needed or hoped for. https://twitter.com/AlexandraErin/status/1294285355441631233

Part of game balance in D&D is resource attrition, and if you can't refresh your attribute invocations mid-adventure then you run into "don't use that, that's the GOOD potion... whoops, the game is over" territory.

On the other hand, if you can correct a terrible roll by fiat but only a limited, low single digit number of times... that presents an interesting choice when something goes badly wrong at the start of an adventure.

And this also further address the chaos plateau problem in a way that, again, does not eliminate the chance of failure, but preserves the sense of agency that is what players really don't like about the chaos plateau: "I specifically specified my character can do this thing."

I'm at a frustrating tipping point here because I have the bones and musculature and a lot of the skin of what I think is a REALLY good adventure game if you're using narrative/roleplay conflict resolution for combat... and I'm trying to meld that into a tactical combat system.

So at this point I really well and truly grok why Strike!, for instance, has the tactical combat system that is part of its subtitle as an optional subsystem whose mechanical pieces and character resources don't directly intersect at all with the main game.

And the tactical combat system I have at least the bones and some of the musculature for blends well with what I'm doing and is the kind of tactical combat I crave... buuuut there's a disconnect in making the numbers matter.

Backing up a bit.

Okay. I'm already starting from the premise that everybody on the adventure is competent at fighting, and that everybody is Good At What They're Good At and Knows How To Use What They Have.

D&D 4E and 5E basically both have a lot of cruft to make sure everybody's attack and defense scores are in the same range, e.g., You Almost Always Use Your Best Attribute To Attack, and everybody has a different path to a decent armor class.

So there is an argument for saying having armor or being big and tough or fast and dodgy is just an aesthetic choice, not a mechanical one. But that eliminates a lot of element of "protect the squishy", aside from differing HP.

And I have tactical combat rules that makes weapon choice consequential; I would likewise like to make defensive strategy ("I'm a tough barbarian."/"I move like the wind and bend like the reed."/"I am a knight in shining armor."/"I dodge artfully.") more consequential, not less.

So just spitballing... this is a very spitall thread.

Defensive attributes. Constitution and Dexterity, to use D&D's terms for convenience though I don't use them in the game. Attribute scale they will be between 0 and 3. Same with armor. 0 to 3.

It's, for resource reasons, very unlikely any character will have 3 in all of them. Armor and Dex may even have some degree of mutual exclusivity between them. Say a character will have between 0 and 5 total defense points.

Subtracting them from hit rolls... makes better/more interesting hits less common or impossible, which is the opposite of what I want. Subtracting them conditionally on the initial hit being green or yellow (normal or glancing hit)? Could work.

This is probably my most rambling, random game dev thread ever and I'm sure it's weird to read but I usually solve these problems by thinking them through on Twitter, and I'm basically... everything in this game, I'm sure, works, except for defenses. I can't square that circle.

Strike!, which is an attribute-free system that uses a similar but simpler result table, just does 4E-style combat without any defenses outside of active measures to apply disadvantage or whatever.

And I feel like that's where I'm trending with this system but it feels like such an omission when I've taken pains to make weapons have a mechanical weight to them.

.......on the other hand, though? The kinds of stories I want this game to emulate, the kinds of stories I want to use it to tell? Armor and such really are more of a matter of aesthetic character design.

Sigh. I've reached the limits of what threading about it on Twitter can do

This is really it in a nutshell: it's hard to make defense *matter* in a way that doesn't make "Turn combat into more of a slog for the other side than it is for your side." a viable strategy, and there's no FUN in that direction. https://twitter.com/Cybren/status/1294295732183277568

I've been writing the phrase "If it's fun, it's fine." at various places in my RPG writing and I think the conclusion I'm coming to is what the author of Strike! also realized: defenses don't add much to the fun.

D&D can incorporate defense as an invisible difference of a small number of percentage points in the odds of hitting a particular character but that doesn't mesh well with my model.

Okay, I need to actually finish writing something to playtest and see how it works, so here's what I think I'm going with.

If you have points in a defense (which again, for convenience, I'll here call Dexterity, Constitution, and Armor), these don't affect ordinary hit/miss chances, that's all in the skill of the attacker and the result table.

But if you have positive Dex, you can use your reaction or a limited per fight resource (I'll test it both ways) to declare Dodge on an attack, subtracting your Dex from the success roll to hit you, not just potentially making a hit into a miss, also lessening an exceptional hit

And if you have Constitution, you can declare Tank on a hit and halve the damage? Like one time per Constitution point (again, range is roughly 0 to 3) per fight.

I think I like the idea of Dex being unlimited but using a reaction, Con being limited per points.

And armor would give you additional points of Tank, in addition to having a huge sway over the rolls you make to see if you're badly injured after a fight. (Which is the main thing armor does in the system so far.)

Ooor maybe Constitution having a larger impact on hit points is sufficient defensive value for it? Keep it simple.

Keep it simple, double over. Dodge doesn't use up a reaction (because dexterous characters will typically have multiple competing uses for those), doesn't involve arithmetic: Dex/Encounter, you can impose disadvantage on an attack.

And each point of armor you have will let you "tank" one attack, in addition to affecting how badly you're hurt after a fight where you got beaten into the ground.

I think I have squared my circle.