Median based RPG...

[Blackguard]

Hello all,


I am working on an RPG that is in development right now and I have a few questions, if you please.

My core assumptions are these:

a) Skill based system
b) Central reference point is the human median
c) d100 rolls for resolution
d) Variable modifiers to the roll as per difficulty
e) Average man with average training facing an average task will succeed 50% of the time
f) 1-50 = failure, 51-00 = success

Question One
------------------------------------
What this essentially means is the average man will have an aptitude of 0, and a training level of 0 - neither advantaging or disadvantaging the roll.  Characters with no aptitude for a skill will have a negative value (-1 to -15 for humans in the system) and those with aptitude will have +1 to +15.

To reflect a base level of 'no skill' I chose -50 (remember, 0 is average training).  An average man will not be able to succeed on an average task with this modifier, though easier tasks will still have a chance of success.  The rationale is that even the most difficult skills to master have some minor tasks that almost everyone could complete.

As a result of my choice, an extremely gifted novice will start his training in a skill at:

(- 50 Base) + (+15 Apt)  + (0 Training) = -35

On an average task, the player would have to roll 66 or higher on a d100 in order to succeed:

66 + (-35) = 51 = success

After some training, this level will increase towards zero and perhaps past it, of course.  My playtesters have exhibited concern that starting players have a negative value.  They draw a qualitative distinction between the above value and a 15% chance of success on a normal task, even though they are exactly the same statistically.  Perception they say is key, so should I switch to a system that starts at zero and builds?  I specifically chose to allow negative values so that values below the normal human base (-50) could be found, say in the case of injury or disability or in children.  A man with a congenital muscular defect might have a base of -75 in Jumping, for example.  With a bottom value of zero, taking into account something or someone worse than humans becomes nigh impossible - how do you get worse than 'zero' skill?

Question Two
--------------------------------
The second question concerns exploding dice.  As a means of providing for luck, I have decided to include exploding dice, but I would like some feedback on what percentage chance you folks have used in the past with this mechanic.  My friend's Rolemaster campaign allows explosion on a 8% chance (01-04 or 97-00), but this is a bit much for me, I think.  I was considering only allowing dice explosion on the extremes (01 and 00) but even a 4% chance would probably work (01-02 and 99-00).  Does the inclusion of exploding dice slow things down too much?


Question 3
-----------------------------------------------
The third question is on maximum values for skills.  My training mechanic is that you roll a d100 and add your Apt score.  If this exceeds your current training value, increase it by one (an elegant mechanic I found in HarnMaster).  This means that there is a maximum value for training of 100 + Apt and that advancement gets harder the higher you climb.  A Master of a given skill might be:

(-50 Base) + (+15 Apt) + (+115 Training) = +80 on the d100 roll for success

This means that the minimum roll is 81.  An average task is automatic (shy of dice explosion) and even Difficult tasks (-25) are also automatic.  A true master of a skill, which these values represent, would only have a chance of failure on Very Difficult (-40) or higher tasks.  Does this seem realistic for humans?  Other races could have higher Apt scores, allowing for a higher maximum (and better performance than a human), so the system is still open-ended in that fashion.

Question 4
--------------------------------------------

Some skills are easier than others to master.  There are essentially two options for skill development:

a) Start some skills at a higher level than others.
b) Allow skill to progress at different rates.

I have chosen to do the latter, but I have run into a difficult situation with my playtesters.  One character in an emergency situation picked up a baseball bat as a weapon.  I started the character at:

(-50 Base) + (+4 Apt) + (0 Training) = -46

and all hell broke loose because the player felt this was waaaaay too low.  The mechanic I use allows him to progress in this skill almost twice as fast as someone with a rapier, for example, but that didn't help him in the exact moment he picked up the bb bat, even if it would help him in the future.  My argument for him was that, in combat, skills faced off against each other, so the difficulty of the task is determined by your foe.  If he faced off against someone equally unskilled in Dodge, for example, he would have a 50-50 chance of bashing his opponent.  I consider this fair, but he was stuck on the number itself.  Does this sound fair to you people?


That's it for now, so let me know what you think.

[Valamir]

Addressing your first question.  I'm not seeing the difference.

If you simply slide the scale from 50 points up you have exactly the same statistical situation as you do now.

Instead of Normal Base -50, Congenital Defect -75
You'd Have Normal Base 0, Congenital Defect -25

Instead a final value of -35 needing an 51 or higher (15%)
You'd have a final value of +15 needing a 101 or higher (15%)

That to me IS (as your friends suggest) more intuitive, AND has an interesting benefit of 100% meaning you were 100% successful, 50% meaning you were 50% successful and 150% meaning you were 150% successful.  A degree of success/failure whose application is immediately intuitive and mathematically useful.

[Blackguard]

Thanks for pointing out that I was an idiot.

HarnMaster's flaw was that it started at zero and worked up and couldn't do otherwise.  I attempted to make a similar system that could start below zero to account for disabilities and the like and didn't realize that it was only training that started from zero and climbed.  What I did with my base value and my target value were 100% independent of what I did as far as training was concerned.

Yeesh.  I feel like a fool.

[damion]

Your not an idiot. I can actually see your point in starting where you did.

Hello all,


Question Two
--------------------------------
The second question concerns exploding dice.  As a means of providing for luck, I have decided to include exploding dice, but I would like some feedback on what percentage chance you folks have used in the past with this mechanic.  My friend's Rolemaster campaign allows explosion on a 8% chance (01-04 or 97-00), but this is a bit much for me, I think.  I was considering only allowing dice explosion on the extremes (01 and 00) but even a 4% chance would probably work (01-02 and 99-00).  Does the inclusion of exploding dice slow things down too much?

Considering that most systems with exploding dice do so WAY more often than 8%, and you want something lower, I wouldn't worry about the slowdown. A more important question is if you WANT there to always to be a chance of success. Do you WANT a person without legs to always have a (small) chance of jumping, for instance?  

Question 4
--------------------------------------------

Some skills are easier than others to master.  There are essentially two options for skill development:

a) Start some skills at a higher level than others.
b) Allow skill to progress at different rates.

I have chosen to do the latter, but I have run into a difficult situation with my playtesters.  One character in an emergency situation picked up a baseball bat as a weapon.  I started the character at:

(-50 Base) + (+4 Apt) + (0 Training) = -46

and all hell broke loose because the player felt this was waaaaay too low.  The mechanic I use allows him to progress in this skill almost twice as fast as someone with a rapier, for example, but that didn't help him in the exact moment he picked up the bb bat, even if it would help him in the future.  My argument for him was that, in combat, skills faced off against each other, so the difficulty of the task is determined by your foe.  If he faced off against someone equally unskilled in Dodge, for example, he would have a 50-50 chance of bashing his opponent.  I consider this fair, but he was stuck on the number itself.  Does this sound fair to you people?

Yea old defaulting problem. I've never seen a good solution to this in this style of system.
In this case it depends. If you take it as  you said, if the opponents dodge skill is a penalty to his roll, it seems fair enough to me.  Or to consider it another way, he should have a 54% chance of hitting a wall, which can't dodge.  You took aptitude into account, not sure what else you can do.

[Blackguard]

Or to consider it another way, he should have a 54% chance of hitting a wall, which can't dodge.

Not really.  A wall being an inanimate object, it would just be assessed a difficulty rating like a normal skill test.  Of course, being a common sense referee, I would assess hitting a wall as Ridiculously Easy and even a beginner should have no problem.

The 50-50 chance I gave him still assumes he is swinging at someone trying to Dodge him.

[Christoffer Lernö]

There is so much wrong with a system like that that I feel like crying.

A normal person average of training succeeds 50% of the time?

Lucky I don't live in that world because people would die pretty quickly after being born seeing as they only have a 50% chance of breathing.

I know, you're trying to make the system by boosting it by 50%, which in a sense is an improvement from say Rolemaster when you have 5% chance or less to succeed.

What's the most wrong?

The randomness. In your system, a skill is about half as important as luck. This might be true for playing poker (without cheating), but as a model for other things: "go to the toilet", "take a shower", "compute the probability of rolling 2 on 2D6", "making lunch" it would be a scary world indeed if that was the case.

I've ranted enough about this. The short story is: Sometimes things are random, sometimes they are not. When things are random it's usually about 'how well did I do it' rather than if I succeeded or not.

This mechanic is only useful if the characters are cripples.

I don't blame you, I'm blaming about 1000 RPG which stuck with this shitty type of mechanic.

[Andrew Martin]

e) Average man with average training facing an average task will succeed 50% of the time.

I feel (like Pale Fire) that this "core assumption" is wrong. I'd maintain that the success chance is more like 99% or better. If one is incorporating opposition like a combat, then sure, roll 1D100, 01% - 50% for one side to win, and 51% - 100% for the other side to win.

[Mike Holmes]

There is so much wrong with a system like that that I feel like crying.

This mechanic is only useful if the characters are cripples.

I don't blame you, I'm blaming about 1000 RPG which stuck with this shitty type of mechanic.

Wow, scathing. Doesn't address the poster's question. But scathing. Doesn't consider that the term "Average" might mean something completely different to you than it does to the poster. But scathing. Doesn't consider that we know nothing of he poster's needs for his system. But scathing...

Good scathing...

Mike

[Matt Wilson]

Hey Blackguard:

Do you have anything in mind to determine not just whether a person succeeds, but how well they succeed?

For example, is every 10% by which I make it a greater degree of coolness? If you posted it and I missed it somehow, sorry.

I haven't seen many games with d100 that work they way I want a game to, so keep up the work.

-Matt

[Walt Freitag]

I feel (like Pale Fire) that this "core assumption" is wrong.

This can be corrected, though. I feel it's more a problem with wording than with the concept. (The concept is a simple linear scale, which I feel has severe shortcomings as a model of probabilistic outcome, but is certainly popular and successful enough in commercial systems -- e.g. d20 -- not to be 'not worth discussing.' Training + skill + mods + etc + d100 >= 101 is mathematically equivalent to the more standard percentile system of rolling d100 under the sum of training + skill + mods + etc.)

An average adult human with average training facing an average task will succeed most of the time. That's because the vast majority of average tasks that people perform succeed. And this is no less true for highly trained and/or specialized and/or professional skills. A gourmet chef, a blacksmith, a plumber, a computer programmer, a surgeon, a hired killer who failed at professional tasks even 10% of the time, let alone 50%, would quickly be fired or out of business.

So what kind of situations have a 50% chance of succeeding?

- People using skills near the limit of their ability. An average person making lunch succeeds most or all of the time, but an average person attempting a recipe from the Charlie Trotter Kitchen Sessions cookbook would probably make a botch of it about half the time. An average surgeon removing an appendix succeeds most or all of the time, but an average surgeon faced with an emergency trauma case might only pull the patient through half the time.

- Nonprofessionals trying to "do-it-yourself" with tasks normally performed by professionals.

- Person in an equal contest against another. Charlie Trotter can succeed in making the recipes in his own cookbook most or all of the time, but will only be able to cook as well as Iron Chef Masahiro Morimoto about half the time.

- Students or trainees. Maximum learning occurs when the student works on tasks near the limit of their ability, so teachers adjust the difficulty of tasks near there. (I'm talking about in-class exercises where the teacher can add the assistance necessary to reach success eventually, not about homework or test tasks which once upon a time were adjusted in difficulty to an average success rate of 75% aka. a "C", but are now adjusted to about an 88% success rate, aka a "B+".)

- Adventurers. Those wacky folks who take swords and rope everywhere they go seem to always be trying things at the limit of their abilites: leaping chasms, climbing cliffs, picking locks, reading ancient inscriptions, and so forth.

This is what often leads to a problem or an apparent contradiction: a typical adventurer with typical training attempting a typical adventuring action like picking a lock is not, by any stretch of the imagination, an ordinary person with ordinary training attempting an ordinary task. It's in no way comparable to a baker attempting to make a bagel or a politician attempting to deliver a speech. (It's much closer to a baker attempting to deliver a speech or a politician attempting to make a bagel.) No wonder it's difficult to design one resolution system that can handle the baker, the politician, and the thief with equal verisimilitude.

Starting skills at zero base is justifiable when they're extraordinary ones (an average person with no training would be completely unable to perform them). But even then you have to be careful between what's a separate skill, and what's an improvised use of an existing skill. Does picking up a different, less than ideal, tool with which to apply a skill mean that the entire skill is negated? In the case of the baseball bat, if the character had never used a hand-held swung weapon before, then setting the skill at zero might have been justifiable. But it's clear the player regarded it as an improvised use, under adverse conditions, of some other weapon skill. Perhaps you should have played it as the character's short sword skill with a -30 for a poorer weapon for which only a portion of his trained techniques would apply. Or if the character had a skill with a blunt weapon like mace or hammer, the minus could have been smaller than that.

Cooking or jumping would never be resolved by skill rolls at all unless the challenge is extraordinary, in which case you set the base at 50% to succeed at a level that a person "with ordianary training" (that is to say none, since few have formal training in cooking or jumping) would succeed half the time. So a two-yard standing jump would start at 50. Then, the degree of randomness depends on what modifiers you apply for distance. If it's +/- 10 per foot longer or shorter, you have a typically high degree of random variation for an RPG. If you make it +/- 50 per foot longer or shorter, you have much less of a random variation.

Similarly, the amount of points you get from skills training has a strong effect on the variability of the system. If training gives you a few percentage points at a time, then a lot of resolution will tend to take place in the 25-75% success range, meaning that there's a lot of chance and unpredictability. If, on the other hand, gaining 50 points is no big deal, so that the success probabilities often shift toward the extremes, then the skill level can predominate.

One drawback to a linear scale system is that if you want the numbers to stay "within the range" most or all of the time, so that most characters have at least some chance of success and failure with most tasks, then you're also going to have to live with the fact that luck will usually have a strong influence. On the other hand, if you're willing to make your scale much wider than the range of the die roll, so that "off the scale" situations where a character has no chance of failure or no chance of succcess (or very small chances that rely on exploding rolls) are common, then you can reduce the prevalence of luck. (In that case, though, you should also consider re-scaling the aptitude scores, which would otherwise appear to have a usually negligible influence.)

- Walt

[Christoffer Lernö]

Wow, scathing. Doesn't address the poster's question. But scathing. Doesn't consider that the term "Average" might mean something completely different to you than it does to the poster. But scathing. Doesn't consider that we know nothing of he poster's needs for his system. But scathing...

I'm willing to apologize if it turns out that he, contrary to my impression, isn't trying something at all similar to the conventional d100 with a skill bonus (or whatever) and a modifier for difficulties.

On the other hand. An average program for an average programmer, doesn't that seem like it should have a bigger than 50% chance of succeeding? Maybe it's my english. It's not my native tongue after all.

Like baking an average loaf of bread by an average man with an average training as baker has 50% chance according to that this system. Maybe average means "terribly bad"? Well that could work.

A terribly bad loaf of bread by a terribly bad baker has a 50% chance of working. Hmmm.. sounds about right.

Ok, this posting didn't add much to the thread. Look at Walt's post instead. :)

[Blackguard]

There is so much wrong with a system like that that I feel like crying.

vs.

I feel (like Pale Fire) that this "core assumption" is wrong.

Pale Fire, you really need to learn from Andrew.  It is all in the wording.  Worded like Andrew's, you get a polite and civil response and some worthwhile debate ensues.  Worded like yours, you get ignored.  Capeche?

I'm willing to apologize ... snip

You'll only apologize if my system turns out to be decent?  Rudeness deserves an apology, period.  The system has nothing to do with it.  If you haven't learned that yet, then you are likely 15 years old or someone with very bad social skills.  Either way, don't bother responding.


For the record, and in response to Andrew (the polite one), my definition of 'average' is very different than most:  

a) I work primarily from the apprenticeship model, and a starting journeyman actually knew very little other than basic skills until he had traveled around to work with various masters.

b) The assessment of 'average' has to take into account the amount of time the person has to complete the task.  An average guy doing an average task with no time limit and endless 'redos' will succeed virtually all the time, yes.  Add a time limit, stress, and only one opportunity to succeed and the percentage drops ... a lot.  Watch how many foobarred batches of bread come out of a really, really busy bakery sometime if you doubt this.  This is why d20 added the take 10 and take 20 concepts to their system for tasks that have all day to be completed.  I have similar concepts in mind.

c) Skill is a combination of talent *and* experience.  The automatic assumption is that a low percentage equates with being a moron - thus the ridiculous and sarcastic examples of breathing and making bread you guys came up with (sheesh).  What is actually represents is a lack of experience with the world.  Line up 100 locks of equal complexity but different construction and a thief will pick the ones he knows 100% of the time.  He won't pick some of the rest of them, though, that is guaranteed.  He won't have the knowledge or the tools.  Shy of listing all the locks and cross-referencing them with the thief's experience list - something I am not willing to do as a GM - having a percentage to represent running into novel tasks is the easiest option.

d) I don't think you guys hang around many average people.  In my office, just as an example, there are a handful of brilliant people, a dozen or so smart folks and a ton of people who need explicit instructions or you can't count on them to complete a task.  With decent training, you can generally count on the last batch to perform, but don't ever change the routine or you are sunk.  Give those latter folks a *novel* task and no instructions and not only do they not succeed ... most of them toss up their hands and start freaking.

d) I assume my GMs are intelligent, reasonable folks who can make judgement calls on their own.


and, yes, I have varying levels of success - six of them.  Now that I am exploring 100 as my goal level and not 50, the exact values will change, so I will get back to the forum on that aspect of it.  I considered the original post somewhat long and so I only included the most germaine questions.

[Andrew Martin]

d) I don't think you guys hang around many average people.  In my office, just as an example, there are a handful of brilliant people, a dozen or so smart folks and a ton of people who need explicit instructions or you can't count on them to complete a task.  With decent training, you can generally count on the last batch to perform, but don't ever change the routine or you are sunk.  Give those latter folks a *novel* task and no instructions and not only do they not succeed ... most of them toss up their hands and start freaking.

I think you'll be interested in reading this: http://www.reciprocality.org/Reciprocality/r0/index.html
Check out Chapter 1, the part on "Mappers and Packers"; it gives explanations for this kind of behaviour.

[simon_hibbs]

d) I don't think you guys hang around many average people.  In my office, just as an example, there are a handful of brilliant people, a dozen or so smart folks and a ton of people who need explicit instructions or you can't count on them to complete a task.  With decent training, you can generally count on the last batch to perform, but don't ever change the routine or you are sunk.  Give those latter folks a *novel* task and no instructions and not only do they not succeed ... most of them toss up their hands and start freaking.

Note "don't ever change the routine or you are sunk". They can perform routine tasks reliably, but can't cope with non-routine ones.

First of all, you seem to be mixing your meanings of 'routine'.

Secondly, I (and I think the rest of us here) are assuming that an average skill and routine difficulty is what we'll be rolling against much of the time. I may have pretty decent computer skills, and be handy with a rifle, but when it comes to driving I'm Joe Normal. Does this mean I'll fail 50% of the time at routine driving tasks?

Is this realy the roleplaying game where your friends get to play one of the ton of unimaginative lumpers in your office?


Simon Hibbs

[Valamir]

Simon, this all depends on how Failure is defined.  

Yeah, I'd say most people fail their driving skill roll alot (and 50% is a good easy number for a game).

But failure in this case can't be defined as crash and burning.  Forgetting to use a turning signal.  Waiting too long to turn the headlights on when it gets dark.  Driving too fast in the rain.  Following too close on the highway.  These are all "failures" to drive properly.

Similiarly a fallen souffle, or a irregularly shaped cake, or a steak that's a little too done, or broken yolks in the fried eggs, could all be examples of "failure"...IF failure is defined (in part) as succeeding in a less then ideal fashion.

So if the 50% chance of success is defined as "performed everything right" then an average person can easily be seen to fail 50% of the time (even good secretaries regularly have typos in their letters until they go back and check them).

This gets back to the side benefit I mentioned earlier about shifting the target numbers up 50%.  With Success being defined as 100% instead of 50% you have this method of looking at success embedded explicitly in the system.

If you roll above 100%...You did everything right.  You aced it.  If you got a 90% you didn't ace it.  You still got the job done well enough to impress the boss, but it isn't perfection.

In fact, the easiest and most intuitive (for American's anyway) system to reflect this would be to use the standard American grading system from school.  in my school 98%-100% was an A+, 95%-98% was an A, and 92%-94% was an A-.  A C generally went as low as 75% and below 60% was an F...from memory...but those of you who went to school in America I'm sure had similiar scales.

So do I succeed with an 82%?  Yes, but just with a B-...which carries with it an intuitive connotation of "pretty good but not stellar".