PDA

View Full Version : A Task Resolution system I have been thinking about



jseah
2014-05-17, 12:13 PM
EDIT2 for clarification: This isn't meant to be used to substitute the d20 system. Some extra modifications to skill points and ability score calculations are required to balance it. (eg. calculating bonuses from 0 instead of 10, reducing skill points and increasing difficulty numbers)

I didn't like the way d20 is so swingy, and the way that stuff falls off the RNG so quickly (and end up auto-success or auto-fail).

So I came up with this system, and asked about it some time ago. Now I have a major improvement towards making it feasible. Would appreciate comments / pointing out of possible shortfalls.


The system at its core is the simple rule: Chance of success = Skillpoints / Task Difficulty

A modification to add an auto-fail threshold subtracted 1/10th of the task difficulty from the effective skillpoints.

Advantages
Unlike d20, this one allows simple tasks to achieve auto-success quickly, while also stretching out the RNG for more difficult tasks.

No one will fail to open the door if he has the key. Unless they're drunk or 5 years old.

But it might take a burglar a few tries to pick the lock (skill 5 vs DC15). And it takes far more investment of skill to auto-succeed at this harder task.

And a large safe might be impregnable to standard burglars, requiring an extremely long time, special tools or a very good thief (DC80), and not even master thieves would be able to breach it on the first try.

This poses some of the obvious problems to playability. It's hard to calculate.

So the new modification is as follows, that makes it really easy:

(Forum games version)
Roll 1 dice of Task Difficulty number of sides (eg. DC37 means roll 1d37)
Add task difficulty / 10, round down (eg. DC37 means add 3, so you'd do 1d37 + 3)
If you roll equal or under your skill number, you succeed.

(Tabletop version)
As above but Task Difficulties are restricted to common dice numbers (2, 4, 6, 8, 10, 12, 20, 100) and their multiples.
So to do a DC30 task, you would roll 1d10 and multiply by 3 then add 3.

EDIT:
Many thanks to Rhynn who pointed out a math error. (Add to the roll, not subtract)

Comments?

Yora
2014-05-17, 01:03 PM
It's quite complicated and I don't really see what advantage it would have. What exactly is your problem with your current skill system?

I recommend reading this advice (http://angrydm.com/2012/12/five-simple-rules-for-dating-my-teenaged-skill-system/) on skills to get some better perspective on what you want and expect from a skill system to provide.

Grinner
2014-05-17, 01:11 PM
First, it's very clever. :smallsmile:

Second, I would calibrate the numbers very carefully. I think you would want to avoid anything that would give fractions in the results, for the sake of simplicity.

Third, you may want to consider including something like Dungeon World's mechanic. That is, it gives one of three possible results: Succeed, Succeed with Complications, and Fail*. Just to keep things interesting.

*Or perhaps something like Fail with Complications?

Rhynn
2014-05-17, 04:47 PM
First method

The system at its core is the simple rule: Chance of success = Skillpoints / Task Difficulty

A modification to add an auto-fail threshold subtracted 1/10th of the task difficulty from the effective skillpoints.

Second method

(Forum games version)
Roll 1 dice of Task Difficulty number of sides (eg. DC37 means roll 1d37)
Subtract task difficulty / 10, round down (eg. DC37 means subtract 3, so you'd do 1d37 - 3)
If you roll equal or under your skill number, you succeed.

Third method

(Tabletop version)
As above but Task Difficulties are restricted to common dice numbers (2, 4, 6, 8, 10, 12, 20, 100) and their multiples.
So to do a DC30 task, you would roll 1d10 and multiply by 3 then subtract 3.

Honestly, your first method (deriving a % chance by division) looks less complicated to me than the second method; you just need a calculator and d100.

Either way, this is just too complex. (Also, I assume you don't mean skill points, but rather total modifier?)

Also, limiting the set of DCs seems hugely awkward and artificial to me. That's a good sign you should re-think this. Consider just switching to a d100 system where you derive a base success % somehow that you can record on the character sheet.

And what about when the modifier is +0 ? Division just gets you a straight-up 0% chance - or a negative %, after (DC/10) is deducted!

And finally, how are you going to handle opposed tests?

Let's look at the math (the most important thing, really).
Modifier +5, DC 20
Standard: 15+ succeeds, that's 30% chance.
First Method: (5-2)/20 = 15%.
Second Method: d20 - 2 vs. 5, 1-7 succeeds, 35% chance.
Third Method: Same as above.

Modifier +10, DC 20
Standard: 10+ succeeds, 55%.
First Method: 10/20 = 50%.
Second Method: d20 - 2 vs. 10, 1-12 succeeds, 60% chance.
Third Method: Same as above.

Modifier +15, DC 20
Standard: 5+ succeeds, 80%.
First Method: 15/20 = 75%.
Second Method: d20 - 2 vs. 15, 1-17 succeeds, 85%.
Third Method: Same as above.

So, right there you can see that, with DC 20, your methods produce steady differences compared to the baseline - in different directions, which means they are not equal at all. Your basic assumptions appear to be incorrect.

Modifier +5, DC 25
Standard: 20+ succeeds, 5%.
First Method: (5-2)/25 = 12%.
Second Method: d25 - 2 vs. 5, 1-7 succeeds, 7/25 = 28%.
Third Method: Can't do it, but DC 24 would get: [1d8]x3 - 2 vs. 5, roll of 1-2 succeeds, 2/8 = 12.5%.

The forum game method is more than twice as likely to succeed as the tabletop method and the first method.

Modifier +10, DC 25
Standard: 15+ succeeds, 30%.
First Method: (10-2)/25 = 35%.
Second Method: d25 - 2 vs. 10, 1-12 succeeds, 12/25 = 48%.
Third Method: Again, DC 24 would get: [1d8]x3 - 2 vs. 10, roll of 1-4 succeeds, 4/8 = 50%.

Look at that; in the standard method, +5 vs. DC 20 and +10 vs. DC 25 are equal, but your method produces huge differences between them.

Modifier +15, DC 25
Standard: 10+ succeeds, 55%.
First Method: (15-2)/25 = 52%.
Second Method: d25 - 2 vs. 15, 1-17 succeeds, 17/25 = 68%.
Third Method: DC 24: [1d8]x3 - 2 vs. 15, roll of 1-5 succeeds, 5/8 = 62.5%.

So in +10 vs. 25, first method beats standard, but in +15 vs. 25, standard beats first method... and we can see that if the DC and the modifier rise by the same amount, the chances of success go up for your methods.

Modifier +0, DC 10
Standard: 10+ succeeds, 55%.
First Method: (0-1)/10 = -10%. Doesn't work!
Second Method: d10 - 1 vs. 0, 1 succeeds, 1/10 = 10%.
Third Method: Same as above, 10%.

You can see already that the modifier itself has a huge effect on the chances of success, regardless of the actual difficulty of the task.

Modifier +5, DC 10
Standard: 5+ succeeds, 80%.
First Method: (5-1)/10 = 40%.
Second Method: d10 - 1 vs. 5, 1-6 succeeds, 6/10 = 60%.
Third Method: Same as above, 60%.

Look at that - an easy task (DC 10) with an okay bonus (+5) and the first method is only half as likely to give a success.

Modifier +10, DC 10
Standard: 1+ succeeds*, 100%.
First Method: (10-1)/10 = 90%.
Second Method: d10 - 1 vs. 10, 1-10 succeeds, 100%.
Third Method: Same as above, 100%.

* D&D 3E only has automatic failure for attacks and saves. If this is either of those, 2+ succeeds, and that's 95%.

Modifier +15, DC 10
Standard: 1+ succeeds*, 100%.
First Method: (15-1)/10 = 140%.
Second Method: d10 - 1 vs. 15, 1-10 succeeds, 100%.
Third Method: Same as above, 100%.

Wasn't your appoach supposed to address the ease of getting automatic successes?

What about impossible tasks?

Modifier +0, DC 25
Standard: 0%.
First Method: (0-2)/25 = -8%.
Second Method: d25 - 2 vs. 0, 1-2 succeeds, 2/25 = 8%.
Third Method: DC 24: [1d8]x3 - 2 vs. 0, nothing succeeds, 0%.

Modifier +5, DC 30
Standard: 0%.
First Method: (5-3)/30 = 6.66%
Second Method: d30 - 3 vs. 5, 1-8 succeeds, 8/30 = 26.66%.
Third Method: [1d10]x3 - 3 vs. 5, 1-2 succeeds, 2/10 = 20%.

Wow! Look at that second method go! And again, compare to the previous example; the modifier and DC both went up by +5 here.

Modifier +10, DC 60
Standard: 0%.
First Method: (10-6)/60 = 6.66%.
Second Method: d60 - 6 vs. 10, 1-16 succeeds, 16/60 = 26.66%.
Third Method: [1d20]x3 - 6 vs. 10, 1-5 succeeds, 5/20 = 25%.

So, if you double the modifier from +5 to +10, you'd have to double the DC to get the same chances of success... an increase of 30, in this case.

At this point, I just made a spreadsheet (I can't be bothered to jam the third method in there since it's so wildly variable; the die type and the multiplier change and that has a huge effect on the results. Here's some success chances from it (standard / first method / forum method):

+10 vs. DC 30 = 5% / 23.33% / 43.33%
+20 vs. DC 30 = 55% / 56.66% / 76.66%
+10 vs. DC 40 = 0% / 15% / 35%
+20 vs. DC 40 = 5% / 40% / 60%
+20 vs. DC 45 = 0% / 35.55% / 53.33%
+30 vs. DC 45 = 30% / 57.77% / 75.55%

+1 vs. DC 10 = 60% / 0% / 20%
+2 vs. DC 10 = 65% / 10% / 30%
+3 vs. DC 10 = 70% / 20% / 40%
+4 vs. DC 10 = 75% / 30% / 50%

+1 vs. DC 5 = 85% / 20% / 20%
+2 vs. DC 5 = 90% / 40% / 40%
+3 vs. DC 5 = 95% / 60% / 60%
+4 vs. DC 5 = 100% / 80% / 80%
+5 vs. DC 5 = 100% / 100% / 100%


The above should illustrate another big problem: your proposed systems are pretty much opaque to players (unless they're plugging numbers into spreadsheets). They'll have little to no idea about their chances of success - and more importantly, the DM won't.

Finally, the big flaw in your math (aside from the methods not being equivalent) is the dividing; the ratio of the modifier and the DC has an enormous effect on the chances you get. This really screws low modifiers, even at low DCs.

This is why knowing statistics is essential to designing task resolution systems for RPGs.


(If anyone spots any errors in my math above, or in my understanding of the OP's proposed resolution methods, please point them out and I'll make corrections.)

jseah
2014-05-17, 08:53 PM
Oh crap, I made a mistake. It's supposed to be add 1/10th of the difficulty to the roll.
Thank you very much for pointing that out.

Yes, skillpoints is supposed to include any bonuses.

-----------------------------

The forum games and tabletop versions aren't meant to be equal to the original or each other.
The original idea WAS to use a calculator. I was told that needing a calculator at the table was too much. =(

Both of them sacrifice accuracy compared to the original (the round down of the 1/10th) to avoid needing a calculator.

The tabletop version further sacrifices resolution (d10 makes success chances come in 10% chunks).

-----------------------------

The % chances weren't meant to be the same as d20. That was sort of the point.

Also, zero skill being negative success chance just means an auto-fail. Having less than 10% of the difficulty means you can't attempt it.


I suppose assessing how useful an extra skill point is will be harder. How useful that point is will depend on the task you are trying to perform. If you know your desired task, then you also know how difficult it is, and therefore how useful a +1 will be.

But yes, it's harder to estimate than d20 (+5% flat).


What exactly is your problem with your current skill system?

My current problems with the d20 system: (and why I decided to make a new one)

If you have any chance of success right now, in 20 points or less, it will become auto-success, regardless of difficulty.
Sometimes needing 20 points for auto-success is too much.
Sometimes it's too little.

Hard tasks are downright impossible to attempt, requiring huge investment to even have the first chance of success.


What I wanted was that a character should have to be ridiculously good in a skill in order to make difficult things look trivial (auto-success). Hard things should still be hard even for characters who have lots of points, they just succeed faster.

At the same time, easy things should be cheaper to master than harder things. While not removing even a small chance for someone who has to have a guidebook open next to them.

Thinker
2014-05-18, 02:33 AM
Your system's problem is still that it is really complicated to pull off on the fly. You put a lot of effort into this and that is plain to see. It is innovative, but not intuitive. It might give you what you want, but it is too complicated for you to ever get there in real play.

To resolve your problems with the swingy nature of d20, switch to using 3d6 instead. You get a nice bell curve for your rolls and don't have a hard time calculating what the player needs to roll for his character to succeed. You can also look into requiring more passes or failures to complete a task so as to reduce the importance of a single roll. Lastly, you could make it so that there are more degrees of a single roll - a slight failure is a success, but with a caveat; a slight success is a success; a major success is a success with a boon to the party; and a major failure is a total failure.

As for the auto-resolution nature of the system, that is a design feature. Once a character gets so good at doing something, it doesn't matter the circumstances anymore, that character can do it all. If you don't like it, consider adding a fate system to your game - once per session, the GM can make a character fail a task. You could also bring a deck of cards and make the players draw from the deck for every action that isn't an opposed roll - on an ace of spades, the character fails automatically and on an ace of hearts the character automatically succeeds.

Rhynn
2014-05-18, 05:45 AM
Oh crap, I made a mistake. It's supposed to be add 1/10th of the difficulty to the roll.

Okay, that makes the first and the second method equal. It doesn't matter whether you add (difficulty/10) to the difficulty roll in the first or subtract it from the modifier in the second/third, the effect is the exact same. (The formula for getting the success % is the same for the first and second method.)


The forum games and tabletop versions aren't meant to be equal to the original or each other.

But with the change, the first and second are equal. The third is obviously still going to vary (for instance, +10 vs. DC24 is 1/3 success with the first and second methods, 25% with the third).


Also, zero skill being negative success chance just means an auto-fail. Having less than 10% of the difficulty means you can't attempt it.

So you're going to produce more autofails, and tasks will be impossible with a modifier of +0 or less - something that's very much not true in D&D 3E. Any time your modifier is less than +1, you know you're going to fail any task. In fact, any time your modifier is equal to or less than (difficulty/10), you're going to fail automatically (for instance, +1 vs. DC 10; a 60% chance in the basic system!).

Is there a reason that you want all tasks to be impossible at modifier +0 (the untrained human average) ? Because under your system, an untrained average human (or a trained below-average human, or someone suffering negative modifiers to the attempt) can never succeed at anything. Even a slightly trained or above-average human (+1) can't succeed at very basic and simple tasks (DC 10, like swimming in calm water, getting along in the wild, a routine performance, handling an animal...).


What I wanted was that a character should have to be ridiculously good in a skill in order to make difficult things look trivial (auto-success). Hard things should still be hard even for characters who have lots of points, they just succeed faster.


At the same time, easy things should be cheaper to master than harder things. While not removing even a small chance for someone who has to have a guidebook open next to them.

But what you've done is created a system where the absolute value of your modifier is the most important thing, not the difficulty of the task, to the point that easy tasks are frequently impossible; a DC 5 task (a 5 foot long jump, hearing people talking through a door, staying in a saddle, etc.) is impossible at modifier +0, a pathetic 20% chance at modifier +1, 40% at +2, 60% at +3, 80% at +4 ...

Easy things are much harder under your system, while difficult things are far, far easier (so long as your modifier exceeds difficulty/10).

Meanwhile, the skill modifier where auto-success is achieved is only roughly 15% different from the base system; you auto-succeed DC X with a modifier of (X*1.1). So DC 30 with +33 instead of +29, DC 20 with +22 instead of +19... is that enough of a difference, really?


What I wanted was that a character should have to be ridiculously good in a skill in order to make difficult things look trivial (auto-success). Hard things should still be hard even for characters who have lots of points, they just succeed faster.

But look at the numbers - you've produced the opposite effect! Harder things are easier in your system than in the base system. With the corrected numbers (difficulty/10 is added to the difficulty die roll), standard / first / second (forum):

+10 vs. DC 30: 5% / 23.33% / 23.33%
+15 vs. DC 30: 30% / 40% / 40%
+20 vs. DC 30: 55% / 56.66% / 56.66%
+21 vs. DC 30: 60% / 60% / 60%
+25 vs. DC 30: 80% / 73.33% / 73.33%
+20 vs. DC 60: 0% / 23.33% / 23.33%

The "break-point" where your method gives a lower chance of success is really high.

There's only one test for the awkwardness - try using the tabletop approach at a table, and listen to the players. (And see how awkward it gets forcing all DCs to fit to the dice available. Getting some Zocchi dice - d3, d5, d14, d16, d24, d46, d50, etc. - will help, obviously.)


Really, you should consider switching to a different game with a better core task resolution system. GURPS uses 3d6 under skill (with the roll modified by circumstances, 18 always fails, 3 always succeeds), and Fuzion has 3d6 + modifiers vs. difficulty number as one approach (used in Sengoku, for instance). Dice pool systems (as in Shadowrun, World of Darkness, The Riddle of Steel, etc.) are good, too, and can produce non-zero chances of success and failure.


Finally, as a general tip: if you've come up with a task resolution system and haven't done a spreadsheet or other method to show you the results (the odds of success in any possible scenario), you're not done yet. Seeing the actual percentages is very, very, very important.

jseah
2014-05-18, 07:15 AM
Sorry, more forgetfulness on my part. I forgot to mention that this isn't for a d20 system, only that I need a task resolution system and didn't like the ones I know of.
- Under the system I was thinking of, bonuses are calculated from ability score 0 (therefore, average untrained human is 5)
- Skill points are a bit rarer
- Difficulty numbers are all adjusted. In particular, trivial tasks are around 5, easy is 10, normal is 20, challenging tasks are about 40 and hard tasks are 60+


Nevertheless, your advice is welcomed. To the spreadsheet! Perhaps after I have thought about it more and refined it, I will come back again for another round.

Since the success % numbers are quite obvious (to me anyway), perhaps a spreadsheet to calculate the marginal success chance increase for current skill numbers and target difficulty.

Many thanks, especially to Rhynn.


One more question. I was thinking of making opposing checks like this:
the target (or defensive / passive) side's skill point multiplied by 2 becomes the difficulty

Any potential problems?

Rhynn
2014-05-18, 07:28 AM
Using (skill)x2 as the difficulty gets you a 50/50 split for equal skills so long as the skills are under 5, but at 5+ the advantage goes to the target/defender/passive (at 40/60). The advantage gets smaller at 6, 7, 8, 9, but then resets to 40/60 at skill 10, 15, 20, etc.

If you're making a whole new system, there's just no excuse for not using something more intuitive and simple, like dice pools, multiple dice under target number, straight up d100 rolls against percentile skills (optionally with automatic success/failure windows at the high & low ends that can be varied based on some criteria), etc.

The first principle of game design comes into play here: you should be familiar with dozens (I'd say minimum 20+, preferrably much more) different games (to the point of having read the books and understood the mechanics) before you start. Otherwise you're going to be reinventing the wheel and making it all kinds of shapes until you get it sort of round.

Different versions of D&D 3E / d20 don't count as different games.