frustrations with my resolution mechanic

[btrc]

Player A has a Firearms Ability of 6 and a Body Nature of 4. He rolls 6d8, and adds up all that he rolls under 4. So, he rolls a 6, 2, 1, 8, 4, and 1. The 6, 8, and 4 are discarded leaving the 2, 1, and 1. The 2 is 2 under, and both 1s are 3 under. Adding it together (2 + 3 + 3 ) nets the character an MoS of 8.

I do see problems with this mechanic though, which is the same for any dice pool mechanic I can think of. Critical successes become much more common, and critical failures are too rare to consider (or more common if it works like Storyteller's botch mechanic). With this particular mechanic, it can also be a little math heavy at higher skill levels. However, if I use a mechanic where each die can only provide a single success, I have the problem of not having a large enough MoS range to tweak with the various modifiers possible in combat.

You may be onto something here, but I think there are too many steps: roll 6d8, compare to Body, add up successes, compare to target number.

What about setting up the threshold for "success" as being a certain number of "roll unders". Something like:
Short range - 1 success
Medium range - 2 successes
Long range - 3 successes

Called shot: +1 success
Spending a turn to aim: -1 success

With the proper dice pool, this might work out. Throw in some stuff like "unskilled use of an ability always rolls 1d8", "you can never count more successes than your current Mind score", "each extra success in combat is +1 damage", etc.

This system would give you automatic success thresholds (if you apply modifiers that make you require a total of zero successes to hit), automatic failure thresholds (if you need more successes than you have dice to roll), and the "smart fighter" bonus (can't get more successes than your Mind score). It also self-limits in terms of modifiers. If you're rolling 3d8 and you need 3 successes to hit, you won't take that called shot because that would be impossible under the current circumstances.

I'm not totally sold on the "roll Ability dice and try to get under your Nature" idea, but if it works in practice, run with it.

It's kind of lonely here in this thread now. I seem to have driven away all the narrativists (they're very shy around numbers). Maybe I should drop off this thread and see if they come back...;)

Greg Porter
BTRC guy

[Tommi Brander]

One thing is not quite clear to me: why do you have a combat mechanic that is completely different from the rest of the system?
It implies a certain focus on combat. You should come up with numerous extended conflict resolutions or use a single one for all of them, if you do not want to focus on combat.


Oh, and for the record, the people in the Dark Ages didn't consider the world flat.

[Michael]

You may be onto something here, but I think there are too many steps:

You're right, that's one of the reasons that I didn't like the mechanic.

What about setting up the threshold for "success" as being a certain number of "roll unders"

This is basically the standard mechanic for a dice pool. The problem I have with this is that it essentially gives you a MoS range from 1 to 10 (if skills range from 1 to 10), and reliably from about 1-5. That means that any modifiers are going to have a much more significant impact on task resolution. It would be like trying to make a Sim using Storyteller, the resolution results are just too broad.

OTOH, with a 2d6 + Nature + Ability mechanic, the MoS range can be from 3-27+, which is more wiggle room for something as detailed as combat.

This system would give you automatic success thresholds...automatic failure thresholds.

I'm not too big on automatic successes and failures. I always want failure to be a possibility, as well as success.

It's kind of lonely here in this thread now. I seem to have driven away all the narrativists (they're very shy around numbers). Maybe I should drop off this thread and see if they come back...;)

Funny, but that's quite alright. In the past two years of designing this game, it has been very rare that I've crossed paths with another Sim designer. Most people I discuss design with are either rules-lite, free form, or LARP, and I wonder why that is. Where's all the Sim people at? Especially now, when I'm getting close to the number crunching, a Sim community is definitely one I would want to be a part of.

Don't get me wrong, the non-Sim people have been a great asset. They're a great help when it has come to helping me define my variables and how they should interact. Furthermore, they've regularly challenged my system on a conceptual level which has only made it stronger. The original system I started out with was basically a d20 clone, just more broken. Now, I believe it more accurately accomplishes what I want it to (though there's still plenty of tweaking to be done).

One thing is not quite clear to me: why do you have a combat mechanic that is completely different from the rest of the system?

Actually, the mechanic is the exact same.

Like I said earlier in the thread, combat is more Sim, while conflict is more Dramatic. Sim players (in my experience) get enjoyment out of the multiple parameters that can affect combat, but not as much so from the more social elements. This is especially true when combat uses miniatures and grid maps, making combat more "tangible". Not as many Sim players really want to have to make multiple Socialize checks at various difficulty modifiers depending on various elements. Furthermore, if that were the case, players would have to learn not only the detailed rules for combat, but also for every single individual ability that they plan to use. Sim systems are complex enough without extensively quantifying every element of the setting's universe.

And that's not what I'm designing for. LRD is intended to be Sim/Drama...a kind of "best of both worlds".

Oh, and for the record, the people in the Dark Ages didn't consider the world flat.

I know, Aristotle was one of the first who figured it out, which was what? 2,300 years ago? I was being metaphorical.

[btrc]

OTOH, with a 2d6 + Nature + Ability mechanic, the MoS range can be from 3-27+, which is more wiggle room for something as detailed as combat.

Of course this also means that your fixed elements (nature & ability) can become more important than all other factors. If I read it right, you can be rolling 2d6+15 as a maximum case. It doesn't seem to be kosher to be trying to minimize the spread of fixed modifiers, when you have a 1-15 spread in the "fixed modifiers" of nature & ability.

And there is always the "self-examination" phase, which is where you ask yourself "How accurately can I simulate this reality with my level of game design skill?". Anyone who feels confident that they can model 1% granularity (a percentile dice system) is either kidding themselves or a helluva better game designer than I'll ever be.

What is the minimum useful level of variance you think your system should have? 3%, 5%, 8%, 10%? I'm personally comfortable with the 8-10% level as a meaningful amount of difference, but the number you pick will be the benchmark for the simplest possible combination of dice and dice types you are willing to accept.

And don't ignore the intangibles. There is something to be said for rolling a big handful of dice on occasion. There is more expectation and tension than with say 2d6, and even if it takes a little longer to resolve, the extra enjoyment may be worth it. Part of tense situations in Shadowrun might have been trying to make that key roll against a difficulty of say 11, where you rolled a handful of d6, kept the 6's and rerolled them trying to get a 5+ on the rerolls. The "okay, I have a chance to succeed!", followed by the "did I succeed?". The extra rolls took time, but added dramatic tension. An extra roll, counting successes or 6's or other simple and fairly quick tasks might be worth it, and might also give you the granularity you are looking for.

Here's another optimum to aim for. For any given combination of skill and tasks, a real-world success curve is generally a half-bell curve shape, starting at a peak of 100% chance and trailing off to effectively zero chance. If you take a combination of nature and ability and start at or near 100% success chance in difficulty, will increases of difficulty in 1 point increments generate the slow-fast-slow dropoff that most real-world tests will follow?

This only matters of course, if that is the kind of results curve you want. The nature of your gameworld and dramatic sense of resolution may work better with a different curve, but the question remains as to whether it will be the same curve shape for both novice, intermediate and expert skill users?

Greg Porter
BTRC

[Michael]

btw, I hope you don't see me as being argumentative. The debate is rather helpful as it has helped me better understand the elements that work, and better adjust the elements that need work.

It doesn't seem to be kosher to be trying to minimize the spread of fixed modifiers, when you have a 1-15 spread in the "fixed modifiers" of nature & ability.

I'm not totally sure if I read that right, but if you mean fixed modifiers as environmental effects and "fixed modifiers" as nature and ability, I wouldn't say I'm trying to limit the range of environmental effects. It's more like I want to be sure I have a wide enough spread for nature and ability so that when I get to designing the various environmental effects I don't have to "retro-fix" natures and abilities to be compatible with them.

And there is always the "self-examination" phase, which is where you ask yourself "How accurately can I simulate this reality with my level of game design skill?". Anyone who feels confident that they can model 1% granularity (a percentile dice system) is either kidding themselves or a helluva better game designer than I'll ever be.

You're absolutely right. There's no way I'd ever design a d% system. One of the major criteria I am using for determining granularity (at least when considering roll over/under mechanics) is what is the odds of the lowest possible result, and the odds for the highest. It really helps with designing the critical success/failure element.

What is the minimum useful level of variance you think your system should have? 3%, 5%, 8%, 10%?

I suppose that's really debatable depending on the designer and system. For what I'm doing, and what I'm trying to build, I could be comfortable with a 3% variance, though it might not be necessary for many situations. It's not too much of a concern as long as I feel like I have small enough increments to work with. Technically, I actually could do a d% mechanic, which would ensure my increments were small enough, though if I were to do that, I'd only end up doing everything in 5% increments, anyway (and not a true d% system).

For any given combination of skill and tasks, a real-world success curve is generally a half-bell curve shape, starting at a peak of 100% chance and trailing off to effectively zero chance.

Crap.

I had just typed out a whole counterargument as to how a 2dX roll-over mechanic could be used to simulate that curve when I realized I screwed up my math and it actually didn't work. That slow-fast-slow dropoff bit blew it out of the water.

This is what I meant by adjusting the elements that needed work.

Still, the slow-fast-slow dropoff would only apply to a character attempting a task that is essentially beyond his/her ability and having a large enough pool of positive modifiers to make the task quite probable. At first, the modifiers wouldn't help so much, but if you lump enough on, the character starts to have a real chance at success, but then at a certain point, it the benefit of the modifiers peter out.

This, of course, would require a novice character to be attempting an extremely difficult task where everything is in his/her favor, which would likely not be terribly common, if not extremely rare. And when this does happen, I'm not sure this would be a problem. It would just be a rare situation where the character was really lucky. It's certainly happen to me on occasion.

Of course, if it really were a problem, I could always switch to a linear curve and go with a 1d12.

Part of tense situations in Shadowrun might have been trying to make that key roll against a difficulty of say 11, where you rolled a handful of d6, kept the 6's and rerolled them trying to get a 5+ on the rerolls.

I don't disagree. Pool mechanics certainly have their appeal. The main thing that concerns me is the diminishing returns that occurs with pool mechanics. Players will have no reason to master an ability if the returns crap out at the halfway point (or wherever). Granted, I could "fix" it by making the best Techniques require the higher ability levels (which I was planning to do anyway), but designing such an element as a "fix" is kind of a cheap way out.

What it boils down to is that for most pool mechanics, the range of significance is much too low for a range of skills. How can a shortcoming like this be resolved?

Also, the "flat-spots" at the intervals of 6 with the Shadowrun mechanic intrigue me. While there is a part of me that thinks the breaks are unnatural, there is a larger part of me that believes there is a way to take advantage of its uniqueness.

but the question remains as to whether it will be the same curve shape for both novice, intermediate and expert skill users?

Well, the curve shape is relative to the difficulty of the task and ability of the user. If you take a task that is of moderate (50/50 chance) difficulty for the intermediate user, it would be extremely easy for the expert and extremely difficult for the novice. Likewise, situational modifiers will have more of an effect on the intermediate user than the novice or expert (like with example before with tying my shoes, making scrambled eggs, or stitching a wound...all in the dark).

Does that answer that question or am I missing something?

[btrc]

And here I thought I was the being the argumentative one... Since I really know nothing about your world or what your design has gone through in the last 2 years, I'm just tossing out mechanics ideas and design philosophies in the hope that it will steer you towards what you are hoping to do, rather than my ideas of what is a "good" system. Speaking of systems, another possibility is the "table", a grid of "skill/nature total" cross-referenced with "modifiers" to get the "to hit" number. This allows you to generate any success curves you want, including proportional modifiers, but it requires the table for all skill use. You can add color coding too. It has its plusses and minuses, like any other resolution system.

Still, the slow-fast-slow dropoff would only apply to a character attempting a task that is essentially beyond his/her ability and having a large enough pool of positive modifiers to make the task quite probable. At first, the modifiers wouldn't help so much, but if you lump enough on, the character starts to have a real chance at success, but then at a certain point, it the benefit of the modifiers peter out.

True enough. I was using a typical firearms hit probability as my model, where you start at zero range (100% chance to hit) and go down from there. If you take novice, intermediate and experts, all their curves have the slow-fast-slow dropoff, just at different ranges. God's gift to pistols does not have a significantly better chance to hit something with a .45 at 100 meters than the average joe. His chance may be better, but it will still be in the tiny probabilities at the tail of that curve. A long barrel revolver with a telescopic sight is another matter entirely. You would see a similar curve with a tank's main gun when fired stationary at a stationary target, moving at a stationary target, stationary at a moving target and moving at a moving target. The curves still follow the slow-fast-slow dropoff, just at different ranges (I did a field trip to the Aberdeen Proving Ground research library for one of my game projects once...)

It's really quite tricky to set up a simple resolution system that lets a novice hit something at point blank without making experts too expert at long range, or which lets experts do expert stuff without making novices unable to do easy stuff. Anyone can make a system that works for a narrow range of average values. The extremes of a skill curve are where things break down.

But, as I said, this is just something to keep in the back of your mind, not a mandatory design goal. If you can make it happen seamlessly with the system, it can add a felt but unseen level of versimilitude that may make the system "feel" better, but adding a clunky step to the mechanic to make it happen would defeat the purpose.

Greg Porter
BTRC

[Michael]

Since I really know nothing about your world...

It's really not too "out there". It's like an alternate Earth, with a subtle supernatural influence that will get stronger over time. For the fantasy medieval installment, the supernatural element is mostly utilitarian and not terribly powerful. By the time players reach the contemporary installment, it will have "grown" significantly in scope. Players will be have powers that could possibly let them fall three stories and not take a scratch (for example), but there will be limitations in how much they can use their power. So, it's not like characters could consistently throw themselves off buildings and survive.

Speaking of systems, another possibility is the "table"

I thought about "the table", and it was really, really tempting at one point. Being able to customize any curve I want, and its relation to other curves would be a pretty nice for a designer. However, I'm not sure I want to force a table lookup on the players.

I was using a typical firearms hit probability as my model, where you start at zero range (100% chance to hit) and go down from there.

You have a good point there. However, another argument exists that a "controlled" environment like the shooting range can really only serve to test the impact of various ranges, and not overall skill or difficulty. I have a good example for this one (and it's one of my favorite stories). I have a friend who is a police officer. He did a spectacular job for all his firearms testing, which has earned him the "Sharpshooter" distinction. Basically, it's a little pin he gets to put on his uniform. Anyway, one day he ends up in an elevator with a crack dealer (he's not in uniform), and the dealer tries to sell him some rock. Well, being a police officer, my friend tries to make an arrest. The dealer isn't having that, so he draws and starts firing, as does my friend. Between the 13 rounds in his pistol, and whatever the crack dealer had, neither of them hit each other. My friend figures that the whole thing just happened too fast for either of them to be all that accurate.

Regardless, you're still right about the slow-fast-slow dropoff. That really has me thinking.

It's really quite tricky to set up a simple resolution system that lets a novice hit something at point blank without making experts too expert at long range,

The extremes of a skill curve are where things break down.

Like, seriously.

I'm trying to be very careful about that, but it's not easy.

But, as I said, this is just something to keep in the back of your mind, not a mandatory design goal.

Granted, but every system has room for improvement, and it's not like I'm on a deadline.

[Tommi Brander]

Well, you probably won't like this, but here it goes anyway...
In normal situations, d6-d6 is the roll. Each factor that favourably increases randomness increases positive die. Elements that eliminate negative effects reduce the positive die. Elements that reduce positive coincidences reduce the negative die. And things that increase adversial randomness increase the negative die.

So, for example, shooting someone is d6-d6. If it is foggy, d6-d8. Quite dark, d6-d10. Multiple enemies in dark, d8-d10 (might hit well, or not, pretty random).
Armwrestling would be d3-d3 or something like that, because it isn't very random. Lottery is d100-d100.

I have very simplistic systems here (you can get some inspiration there): http://s14.invisionfree.com/Tablets_of_Dleinr/index.php?showforum=9

[btrc]

I was just looking at Dream Pod 9's "Core" system and it has a threshold/success system that you could look at the curves for. It seems like an "NdX, keep best 1", but then each maximum die roll (in their case, 6's) counts as an extra success.

Example: I'm rolling 5d6 against a target number of 5. I roll 1, 3, 4, 6, 6. I keep the 6 and get 1 extra success for the second 6. So, I get 2 successes.

This might be a be coarse in the granularity, but the concepts involved might be worth looking at for an "NDX best 2" or similar system.

The system also rates skills as x/y, where X is your level, and Y is your depth of knowledge. A 5/1 would be a whiz at a narrow field, while a 3/3 would be a less skilled generalist in a broader aspect of that field. Like say knowing how to shoot one type of gun vs. a lot of types of guns.

Greg Porter
BTRC

[Michael]

I was just looking at Dream Pod 9's "Core" system and it has a threshold/success system that you could look at the curves for. It seems like an "NdX, keep best 1", but then each maximum die roll (in their case, 6's) counts as an extra success.

Yeah, I'm familiar with this. The probability curve it creates is rather interesting. It's one of the few mechanics that creates the log-linear curve that one would expect in a "natural" environment (so no slow-fast-slow, or flatline progression), except that it's inverted (it "rises" toward 6, when it should actually drop). Of course, the easiest way to invert the curve is to switch to a roll-under system, but there's other problems as well.

This might be a be coarse in the granularity, but the concepts involved might be worth looking at for an "NDX best 2" or similar system.

I drew out a table of a roll X take highest scenario. Actually, it was more of a probability of X 6s when rolling n dice. As expected, the returns are diminishing, even when keeping multiple dice. What was more concerning was how quickly the curve centered around 6s.

The system also rates skills as x/y, where X is your level, and Y is your depth of knowledge.

I totally forgot about that. Thanks for refreshing my memory. I knew I was missing something when "picking apart" the system. DP9 released a great core rules supplement that was mainly designed for CORE Command, but can be used for any of their settings. I don't own it, but I've borrowed it from a friend on many occasions. I think I'll need to borrow it again...or just buy it.

I have to admit, despite the flaws of the system, it always had the best "feel" to me.