Ygg Action Resolution again (Concessions and Stuff?)

[Christoffer Lernö]

I'm cutting and pasting from my Ygg document here. I kind of came up with an idea that seems to work. Advantages and Concessions seem to be a little more natural as well. But I only just worked it out. If you would have a look, then that would be nice.

Basic Resolution System
This is how it is done: pick as many D12 as you have for the relevant statistic. You want to jump? Well that's Movement. Armwrestle? That's Power, and so on.

Roll these dice and for every D12 which is equal to the Difficulty (more about that later) you score one degree of success. If you roll a 12 on a die you not only get a success but also an extra die to roll.

Now round up the successes and compare them with what you need. So what do you need? Well that depends on the test. There are two types of tests: Opposed and Static.

Opposed Tests
This is when someone or something else is also trying to win. Armwrestling is a good example. To find the winner, simply compare who has the highest degree of success. The one with the highest wins. In the case of a tie, there is a second test and so on until a winner has been found.

Static Tests
If you want to jump a chasm or just see if you can make out Mount Doom in the distance, you do a static test. The GM gives the required degree of success required to make it.
A vague rule of thumb might be:

1 Trivial
2 Anyone can do it
3 You can do it
4 Not everyone can do this
5 Pretty advanced
6 Most can't do this
7 Almost noone can to this
8 Only the extremely skilled could ever do something like this

Difficulty
The difficulty describes how easy it is to do a task and how big the chances are that an action is performed at it's optimal efficiency.

1 Perfect conditions
6 Sucky conditions
11 The most random conditions

Concessions and Free Advantages
If you have a higher degree of success than you need to complete a task, then you can trade those in for other benefits. Benefits include improving the time to complete an action or the quality of it. These benefits are called Free Avantages.

The opposite of Free Advantages are Concessions. You can do concessions before or after you rolled. By making a concession (which is basically making the quality of your action worse in some respect) you get another die to roll and add to your degree of success. Note that you can mix both conessions and advantages, for example taking longer time to complete a plainting in order to get a higher degree of success to put into improving the quality of the painting.

When & What About Concessions
Concessions can be made before or after the normal dice are rolled. There is no penalty to doing it before or to do it afterwards except for the fact that you can't make concessions change things that took place before the action was performed. Say that you're jumping a chasm. You roll and you fail. Now at this point you can't take a concession saying you took off your backpack and jumped without it nor that you did it with a running start. On the other hand you might be able to get away with jumping over but losing the backpack in mid jump (for another die).

When & What About Free Advantages
Just like concessions, there are limits to what can be taken as advantages after the action is rolled for. For example, you can't make the jump then declare that you were wearing the backpack, if you weren't supposed to be wearing it in the first place. However, if it's before the jump you can say you're jumping with the backpack, but that will increase the necessary degree of success. So if you fail, you have to take concesssions to get over, and these concessions can't be that you forgot to put on the backpack.

[Mike Holmes]

I would simplify this by dropping "static tests". See this Rant. I apollogise in advance if you've already read and disgreed with this.

Simply replace "degree of success" with opposing difficulty dice rolled by the GM. Makes any test possible for anyone theoretically, and gives you the same scale of success for both "types" of rolls. As such, then, one net success means that you barely made it. One net success for the chasm means that you barely failed. Then Advantages are just spent successes, and Concessions are just purchased successes.

Straightforward.

Mike

[Christoffer Lernö]

I would simplify this by dropping "static tests". See this Rant. I apollogise in advance if you've already read and disgreed with this.

I read that (when you first posted it no less!) and I agree with you. So why the heck static tests?
Well the mechanic started out as a static test. In fact it was the damage mechanic. As such there was no need to oppose it as the defender's toughness was already kinda baked into the target number. The successes was never needed. Now sure you could make a mechanic where the armour work as an opponent, shaving off damage successes with the armour piercing value of weapon works as a target number. In fact this is the system Shadowrun uses.

My goal however has always been to keep the rolls as few and as simple as possible. That's why I rather have a single roll than opposed rolls. That's why the whole combat only has stuff rolled by the attacker and not the defender.

I actually consider the opposed tests to worst mechanic in the system as it involves two rolls. However, I could not do it any other way as I was required to (for simplicity's sake) use the damage mechanic for action resolution.

I comfort myself with the fact that action resolution isn't supposed to be a big thing in my game anyway, so it won't make much of a difference in handling time and stuff.

Simply replace "degree of success" with opposing difficulty dice rolled by the GM.

You could consider that the stat set by the GM comes from such a roll already. As I haven't given details for how the GM gets required degree of success, don't assume it won't look like that. In fact even the older version with karma resolution mostly relied on the GM rolling up target numbers instead of choosing them. It's something I'm gonna stay with so your suggestion actually makes a lot of sense from that point of view... I can let the GM use this method to generate the required DoS. However - it will tend to be the same for every character and the GM can always override his roll (I was thinking of it as a GM-aid, not as an actual mechanic the GM had to follow)

Makes sense?

[Andrew Martin]

You might want consider the changes made in the recent versions of the WW storyteller system. Use a fixed target number, say 7, as it makes scanning the dice for successes easier.

[Christoffer Lernö]

You might want consider the changes made in the recent versions of the WW storyteller system. Use a fixed target number, say 7, as it makes scanning the dice for successes easier.

Aaaa Andrew you evil thing you! First I don't know what a "recent" version of WW's system would be. Have I seen it already in the WW books I have or not? There is no way to tell from your cryptic passage. And I don't have the possibility to browse through any material here in Taiwan (all my stuff is back in Sweden).

And except for scanning the dice, why give up movable target numbers when  I actually use the target number as a way to turn the randomness dial.

[Le Joueur]

And except for scanning the dice, why give up movable target numbers when  I actually use the target number as a way to turn the randomness dial.

Why?  Because you already have one?  That'd be the number of dice rolled.  Harder task?  Take away a few.  Easier?  Add a couple.  No 'target number dial' needed.  It's redundant anyway.  (Much the same way that GURPS escalating point-costs is redundant.  A sigmoid graph like theirs already has diminishing returns built in.)

Fang Langford

[Jonathan Walton]

And I don't have the possibility to browse through any material here in Taiwan (all my stuff is back in Sweden).

I feel like Keanu Reeves.

"Dude!  You're in... Taiwan?"

I should have noticed that before.

I alternate between living in Raleigh NC ("home"), Oberlin OH ("college"), and Beijing ("home #2").  A while back, me and Erick Wujcik were planning to start brainstorming about creating a Chinese-language RPG tailor-made for a Chinese audience who thinks "jue-se ban-yan you-xi" are just computer programs.  It fell through in the end, because both of us were busy with other things, but that's still something I'd love to at least THINK about at some point.  We might even be able to bring Erick back into the discussion (if he isn't working on "RECON: Modern Combat" or something else).

Any interest?

Later.
Jonathan

[Christoffer Lernö]

And except for scanning the dice, why give up movable target numbers when  I actually use the target number as a way to turn the randomness dial.

Why?  Because you already have one?  That'd be the number of dice rolled.  Harder task?  Take away a few.  Easier?  Add a couple.  No 'target number dial' needed.  It's redundant anyway.

Err? What?

Parameters: Randomness, Character "skill" rating, degree of success needed.

Player rolls as many dice as the skill rating gets an extra die on each 12 against the randomness number and counts total number of successes.

Working the skill rating up and down will alter the number of successes as well as the chance of boosting a skill beyond it's original level. Randomness number makes success increasingly less certain, but most of all it mixes up the levels.

Notice that the extra die on 12 gives you at least a 40% chance of getting another die. If the target number is 1 (given success) that's a 40% chance of getting a degree of success of 7 or more. Compare that to someone having 1 die who only has a 8% chance of boosting the degree of success from 1 to 2.

For opposed rolls, we have at randomness (target number 1) a mixing that allows a strength 3 person (like my mother) win over a strength 7 person (someone built like Arnold Schwarznegger) 0.004% of the time. If we put in a randomness of 11 instead, my mother has a 23% chance of success. So the randomness has a very real impact on things, and it is separate from the other dials and work in quite a different way.

I could introduce another dial, namely the chance to reroll. Right now it's set to 12, but one could imagine it being moved to 11-12 and so on depending on the situation. However right now I don't feel a need for such a dial.

What this mechanic allows me to do is to separate armwrestling (randomness 1) from playing cards (randomness 11?). Both are opposed tests but the former has a lot less of randomness. Basically I introduce a randomness dial which I then use to tune the randomness of situations.

But maybe I misunderstand your post completely Fang.

[Andrew Martin]

Pale Fire wrote:
> Aaaa Andrew you evil thing you!

:)

>  First I don't know what a "recent" version of WW's system would be.

Trinity and Exalted have the latest versions. You've all ready got the reroll on max die value, and you've dropped the botch on "1". So just stick with the fixed target number (and measure everything in dots) and you'll have Storyteller 2 for D12 real soon! :)

> And except for scanning the dice, why give up movable target numbers when  I actually use the target number as a way to turn the randomness dial.

It's quicker. And quicker play is more fun, because it's less time spent on mechanics and more time roleplaying (assuming all else is equal).

[Christoffer Lernö]

Trinity and Exalted have the latest versions. You've all ready got the reroll on max die value, and you've dropped the botch on "1". So just stick with the fixed target number (and measure everything in dots) and you'll have Storyteller 2 for D12 real soon! :)

But that would spoil the main point of the mechanic which is

...I actually use the target number as a way to turn the randomness dial.

Why is this important? Because this system is an adaption of the karma mechanics of higher than stat wins, lower than stat loses equal is 50-50. Turn the dial to 1 and you essentially have this system.

In the old system however, you had to add a "random events" modifier (which I subsequently dropped, but I felt it still left something to be desired not having it) which made the difficulty different between different persons.

Basically you simulate this by cranking up the randomness dial. With 1 at the dial, you're assuming perfect conditions. Turning it up makes more random stuff enter the equation. Stuff that the GM hasn't calculated into his/her "required degree of success". Like weather conditions, footing and whatnot.

Actually you could use the same mechanic with the randomness dial turned up. However, it takes more time than to just roll a simple dice and read off results.

Sure, speed is an issue, but in the cases I'm thinking of (comparatively rare action rolls) it's more important that results are reasonable. In addition, by moving the dial to 1, there is no way you can fail on a die. That means at "skill rating" 6 you succeed on anything which has 6 or less for difficulty which in turn gives you a solid indication on when you have to roll and when you don't have to.

You only have to roll when:
* Randomness dial above 1: roll
* Skill rating less than required degrees of success: roll
* An opposed test rolls same or equal degree of success.

Actually I'm surprised it works so well. But maybe I'm overlooking something, or are you only suggesting the dropping of target number as a way to speed up play?

If it's the former, I have to seriously look into it. If it's the latter, game testing will tell if it needs to be optimized or not.

[Le Joueur]

Hey Christoffer,

So good so far...

And except for scanning the dice, why give up movable target numbers when  I actually use the target number as a way to turn the randomness dial.

Why?  Because you already have one?  That'd be the number of dice rolled.  Harder task?  Take away a few.  Easier?  Add a couple.  No 'target number dial' needed.  It's redundant anyway.

Parameters: Randomness, Character "skill" rating, degree of success needed.

The player rolls as many dice as the skill rating gets an extra die on each 12 against the randomness number and counts total number of successes.

Working the skill rating up and down will alter the number of successes as well as the chance of boosting a skill beyond its original level. Randomness number makes success increasingly less certain, but most of all it mixes up the levels.

When you alter the number of potential successes, you change the "randomness," don't you?  "working the skill rating...down" "makes success increasingly less certain" with a static target number, perhaps more simply that introducing dynamic target numbers on top of dynamic die pools.

As for your parameters, they pretty much list the "randomness" factors.  Two is duplicative; three is quite redundant.  What I am saying here is you're basically putting three randomizers into one package and then complaining elsewhere about too much reliance on randomness.

Let me separate the three randomizers, maybe that'll make it more clear:

Randomness Number

With a minimum number of successes equal to one, make the size of the die pool static and you get most game systems; GURPS' die pool is three, always three; their Randomness Number is affected by many factors.  Seems random enough to be popular.[/list:u][/list:u]

Die Pool Size (Character Skill Rating)

The target number is static and the minimum number of successes is one.  Affecting factors add or delete dice from the pool.  Better or worse 'skill' also sizes the pool.  Since you don't seem to be a fan of 'always having a chance,' this should be ideal; too many negative modifiers and all the dice go away.  I still don't see how this is in any way 'weaker' than any other on this short list or why it makes any difference (other than complexity) to add another randomizer to it.[/list:u][/list:u]

Degree of Success Needed

I'm surprised you hadn't mentioned this one earlier.  This is actually a randomizer in its own right; die pool size is static, target number is static, how many successes are a number affected by various factors like skill and situational modifiers.  For example, your skill is 3 (well skilled), you want to do something without 'aiming,' you roll the 'seven dice;' you take out up to your skill number in failures and one success for not aiming.  The resulting 'balance' determines your fate; mostly success is 'generally' winning (with some consessions), mostly failure would be pyrrhic victory.  This works just as well as either of the other two randomizers.[/list:u][/list:u]The point is, you're using

three randomizers all at once.  Why?  Any one of them will do the job; more just adds complexity and makes the game more difficult (assuming complexity and difficulty are related).

For opposed rolls,

The "randomness number" on opposed rolls would be the number of successes the opponent generated.  This almost raises the number of randomizing factors to four.  Is complexity a design goal?

I guess the real question is what are you comfortable with?  How many randomizers you build into your system are your business; it should suit your tastes.  What I'd prefer is if you did it fully knowing what you're about.  That way when it starts to get complicated and the playtesters say that the system requires too much attention on its own, you'll know why you 'put so much in.'

There really isn't a 'right amount of complexity.'  There are traditions that result in 'broken systems,' however.  Choose carefully, not reactively.

Fang Langford

[Mike Holmes]

There are traditions that result in 'broken systems,' however.

Amen.

Mike

[Walt Freitag]

I don't see any problem here that's that fundamental. I see some practical problems with the system as described, but I don't have a problem with the general idea of a system that has three basic variables:

1. The skill of the person.

2. The difficulty of the task, e.g. the skill of the opponent.

3. The performance variance which depends on the nature of the task.

For example, if I'm trying to outscore an opponent in a single game of pinball, and that opponent's mean score is 50% higher than mine, my chances of winning the game are still reasonable (I'd estimate, based on the variance of my own pinball scores, that they would be in the 10 to 25 percent range). But if I'm trying to arm wrestle someone whose mean arm strength is 50% higher than mine, my chances of winning are extremely poor. That's what Christoffer means by varying the "degree of randomness." It's a variation in how much a given difference in mean performance (skill level) is likely to swing the outcome in a given single instance of resolution.

In any single instance of resolution, it's possible to collapse #2 and #3 together by taking the randomness into account in the scale of the skill scores. Just say that the skill deficit I have in pinball representing a 50% lower mean score is a -3, while the skill deficit I have in arm wrestling representing a 50% lower mean score is a -9, and then resolve them both in the same way. But that doesn't work on a systemwide basis unless every skill has it's own separate score. If I'm basing skills on requisites, then both arm wrestling and cow tipping are strength-based scores. If I think my changes of beating an opponent whose strength is 2 points higher than mine at cow tipping should be much higher than my chance of beating him at arm wrestling, because of the greater influence of random factors involved in the former (e.g. the behavior of the cow), then I need that third variance factor as an independent variable in the system.

I also like the idea of using the variance factor in strategy and tactics. It can be shown mathematically that if one side in a contest has the power to manipulate the level of randomness from one round of action to the next, then that side can sometimes achieve a greater than 50% chance of winning the contest even if its expected outcome in each round is always negative, by adopting the strategy of increasing the randomness when it's behind and decreasing it when it's ahead. There are thirty-two head coaches in the NFL who make their living by grasping this principle. A team that has a perceived talent disadvantage will try to force both sides to play riskier strategies, try to run down the clock if it's not already behind (a "shorter" game increases the variance, though not enough to make it worth running down the clock if you're already behind, usually), and hope for bad weather. If I'm duelling a swordsman with superior skills, I'd prefer to have the fight on a frozen lake, or in a courtyard full of hanging laundry, or in a chicken coop filled with flying feathers... anything to increase the performance variance.

Okay, now from principle to practical. The problem with the system presented above is that it doesn't cleary separate out variance as an independent variable. Instead, adjusting variance is accomplished by adjusting the target number, which also has a very strong effect on the mean chance of success in unopposed contests and a strong effect on the mean margin of success (which is important in the system) in opposed contests.

This creates the problem that there are two different difficulty variables (actually three, but only two of them are in effect at a time): the number of dice rolled by the opponent OR the number of successes needed, AND the target number (number to be rolled on a die for it to be counted as a success).

Remember, once we get down to an individual instance of resolution, there's no longer any distinction between the chance of success as influenced by mean difficulty, and the chance of success as influenced by the degree of randomness, at least as far as the success or failure part of the outcome is concerned. There's only just one probability. Higher randomness just means that my chance of success if I have a mean disadvantage, and my chance of failure if I have a mean advantage, are higher than they would be at a lower randomness setting. (While a 50-50 chance, of course, remains a 50-50 chance at any level of randomness.) I could make that adjustment on the fly using just a single combined-chance-of-success knob, such as the number-of-successes-needed. But it's a little tricky.

The problem is, using Christoffer's proposed system, it's still tricky. If I want a difficult unopposed action to have a high degree of randomness, using this system, I have to juggle the two difficulty variables in a complex way. By setting the target number high I supposedly make the randomness higher, but I also have to make the number-of-successes-needed low, or else success just becomes even more unlikely... and making the randomness higher, when I'm at a disadvantage, is supposed to be a way of making success more likely than it otherwise would be. It would be easier to adjust just one measure of the chance of success like in the previous case. It's not quite so bad with opposed rolls -- a higher randomness applies to both sides, so it doesn't skew the mean outcome -- but it makes the mean margin of success lower for a higher-randomness contest, which is exactly the opposite of what you'd expect.

On the other hand, in principle, a truly independent adjustment knob for situational variance/randomness could make it a lot easier to arrive at a die roll whose combined-chance-of-success reflects both the mean advantage/disadvantage and the variance in a sensible way.

I don't see a way this could be fully achieved in the currently proposed system without making major changes. It can be partially achieved by (1) getting rid of the variable target number, make it always 7+ or whatever; (2) add a roll of centered dice (such as Fudge dice) to the number-of-successes-needed in unopposed rolls; the more dice added, the higher the variance; (3) add a roll of centered dice to the opponent's roll to add variance in high-randomness situations. The limitation is that there's no way to turn down the variance in opposed rolls for low-randomness situations.

Another choice would be to keep the variable target number as a randomness knob, but get rid of the unopposed roll that causes the most problems. Instead of setting the difficulty as a number of successes needed, set it as a number of dice rolled, against the same target number as the character (which represents the randomness setting), with the result being the number of successes the character needs. In other words, make all rolls opposed, just as others have suggested... but then you can keep the variable target number as a randomness knob. Just be aware that in that case, the mean margin of success or failure will be maximized when "randomness" is in the middle of its range (7+ on d12s) and then will decrease again as the randomness increases beyond that.

- Walt

[Christoffer Lernö]

Let me separate the three randomizers, maybe that'll make it more clear:

Ok. Now you're playing a wholly different game from what I'm talking about Fang.

How you can put the degrees of successes as a "randomizer" is actually beyond me.

Look, we have x dice, right? Put the target number to 1 and we only actually count the number of 12's rolled. You argue this is increasing randomness depending on the number of dice. However this is depending on how you view it.

It is true that 6 dice get 6 times the chance of "boosting" at least one step up, but the relative boost (that is to say, the % increase in expected value from the initial) is the same.

I can't see how you can say that something that does not affect the number of dice rolled (the degree of successes needed) is a randomizer.

How does that change the distribution of the results?

Aren't you mixing up predictability (there are natural boundaries where predictability goes down) and randomness induced by the mechanic? It's only the latter I seek to control.

[Le Joueur]

Let me separate the three randomizers, maybe that'll make it more clear:

Ok. Now you're playing a wholly different game from what I'm talking about Fang.

How you can put the degrees of successes as a "randomizer" is actually beyond me.

I don't understand.  I gave an example of using 'degrees of successes' as the way of interpreting the dice.  It may have been heavily experimental, but I think it could be turned into a "randomizer" all by itself.

I can't see how you can say that something that does not affect the number of dice rolled (the degree of successes needed) is a randomizer.

How does that change the distribution of the results?

Aren't you mixing up predictability (there are natural boundaries where predictability goes down) and randomness induced by the mechanic? It's only the latter I seek to control.

It changes the distribution thus:  You roll seven dice; you need to get 7 or higher to succeed on each.  If the required "degree of success" is 2 or more successes, then the distribution is 93.75% chance of net success; if it is 4 or more, the distribution is 50%; if it is 6 or more, the distribution is 6.25%.  This doesn't change either the number of dice rolled or the target number, yet the distribution (the chances) changes quite well.

How many successes 'are needed' affects the success probability.  How many dice are rolled affects the success probability.  The 'setting' for the target number affects the success probability.  That's three different 'dials' to affect the chances someone will have at success.  Each has a different character, some change variability, some don't.  The point is, when you make all three a part of your game, your asking the players (including gamemasters) to consider three different variables: How many dice do I use?  What is the target number this time?  And how many successes do I need?  Now multiply that by a factor of how many times a round of combat the participants will need to 'roll for success.'  (Double it if these rolls are opposed.)

Do you begin to see why I am cautioning the 'kitchen sink' approach?

Now let's say you lock in the target number (say at 7 or higher on 12) and the number of successes needed (1 is enough); every time a roll is needed, the player grabs his skill number of dice plus bonus dice, minus penalties.  One quick roll and the resolution sequence is complete.  The same scheme works with whatever is chosen as the singular dial to affect the probabilities.

Personally, I'm not fond of a game that requires the gamemaster to intervene in every resolution with target numbers, I have other things I want them doing rather than making all these judgment calls, but that's just personal bias.

Likewise, the main reason I posted here, in follow-up, was because I had never even thought of a resolution mechanic that locks dice pool and target number, varying probability by number of successes; it might even form the basis of a really good Fortune-in-the-Middle mechanic, if I do it right (you could save up 'extra successes' for later or borrow some).

Does that answer the question?

Fang Langford