D&D 4E Heavy Concrete Data on 4e's Skill Challenge System (long, lots of tables)

[Skywalker]

It would be great if somebody followed up on this and asked:

1) How many party members can use Aid Another if there is no time pressure? i.e. why won't there always be a +8 from Aid Another on all checks assuming a 5 PC party?

2) Is Aid Another possible when there is a time constraint?

I have come across three instances so far:

The common examples we have encountered so far are (noted earlier in this thread):

1. Skill Challenges in Combat Encounters. Often PCs are pressed by Combat and are unable to engage in Aiding or even Skill Challenges.

2. Time Pressure. I have seen Skill Challenges where in order to be able to use their Skill in the Skill Challenge, the PC can't Aid Another. The normal restrictions to Aid Another still apply. For example, infilitrating a camp. One group creates a diversion, while another sneaks into a camp and may not assist.

3. Other Limits. I have seen Skill Challenges where some Skills are restricted as to how many PCs may Aid Another i.e. only 2 PCs may Aid Another when breaking down a door.

 

[CSK]

I originally posted a table in the other big thread that showed the likelihood of achieving 2K successes before K failures for various chances of success on a single role. The whole system was very skewed and produced poor results.

However, if you make the simple change of requiring K successes before K failures, almost everything falls nicely into a reasonable model. I won't post the whole table here, but suffice it to say that if you have a 50% chance of success on a single roll, then you have a 50% chance of getting K successes before K failures for any value of K.

Also, as your success rate on a single roll increases, you likelihood of passing the challenge increases in a fairly nice way, especially in the 55% to 70% range.

The only trick then is to arrange the DCs for each level so that a standard party will have roughly 50% to 70% chance to pass any given roll. That's a matter of looking at average characters in a well rounded party at each level.

Funnily enough, I don't have the books (I'm waiting for a 2nd printing so I can avoid some of the silly typos), but I am a math geek! For those interested, the easiest way to compute the probabilities (at least for me) is the following formula.

To achieve N successes before K failures with prob of success p on any given roll, we have

P(n,k,p) = Sum_{j=0}^{k-1} C(n-1+j, j) * p^n * (1-p)^j

OR for those who prefer a closed form and speak Mathematica

P(n,k,p) = (Gamma[1 + k] * Gamma[n] - (1 - p)^k * p^n * Gamma[k + n] * Hypergeometric2F1[1, k + n, 1 + k, 1 - p]) /
((-1 + n)! * Gamma[1 + k])

Enjoy!

 

[Hussar]

^ makes my brain hurt.

But, I still think it misses the point. Yes, if you have 50/50 odds and you require twice as many successes as fails, then you are going to fail far more often than not. That's pretty elementary I think.

But, the point is, the system is rewarding groups who work together as a team to improve their odds from 50/50, to much higher percentages. If you get your odds to 90/10, for instance, then your chances of overall success are pretty high.

I'm still thinking that this is actually a feature, not a bug.

 

[Stalker0]

Is this really a bug? Isn't this a feature? You have a system designed specifically to reward cooperative play. If you want to succeed, you need to work together as a group (Aid Another, taking actions that don't contribute to Success/Fail but give bonuses to others) and your chances of success dramatically increase if you choose to work as a group.

Additionally, it seems to me that many of the bonuses - rituals, Aid Another etc. are mostly under the control of the players, not the DM. So, it's up to the players to come together, work as a group and solve skill challenges.

What am I missing?

It boils down to this...do you feel if a player uses a less than optimal spell, that the entire combat should be lost?

For example, the difference in two people and 3 people aiding another is a HUGE difference. With a party knowledgeable the rules, they should kick the crap out of the guy who wants to roll his skill check instead of aiding.

 

[Terramotus]

pg 75 "Group Skill Checks" specifically discusses everyone using aid another and one person making the skill check. "Sometimes a skill challenge calls for a group skill check." That's why I said it's vague.

Absolutely incorrect.

The others make checks to help the lead character, in effect aiding that character, but their checks provide neither a success nor a failure toward resolving the challenge.

This is not Aid Another. Rather, it is the same as the use of the Perception skill check option listed in the Urban Chase. It duplicates the Aid Another actions mechanics, mostly, but it is a skill check.

Aid Another is a standard action, listen separately, that is not a skill check. It can provide bonuses to skill checks, but it is not a skill check. The Skill Challenge system "require the players to make rolls at specific times."

In other words, the DM gets to decide if he wants the players to be able to use something similar to Aid Another when he designs the skill check. The players don't get to do an end run around that and cheat out of having to make skill checks by using Aid Another.

This is so that, in a chase scene, when it comes time for the Wizard to make his Athletics or Acrobatics skill to avoid being caught he can't just decide to Aid Another, shrug his shoulders when he fails, and somehow magically not get caught.

I called WotC customer service today and they said that yes, you can use Aid Another on skill challenges.

Without trying to sound offensive, I doubt they know what they're talking about. That's not what the RAW say. I'll wait for something official, or something from an actual designer.

 

[gonesailing]

^ makes my brain hurt.

But, I still think it misses the point. Yes, if you have 50/50 odds and you require twice as many successes as fails, then you are going to fail far more often than not. That's pretty elementary I think.

But, the point is, the system is rewarding groups who work together as a team to improve their odds from 50/50, to much higher percentages. If you get your odds to 90/10, for instance, then your chances of overall success are pretty high.

I'm still thinking that this is actually a feature, not a bug.

Hussar there is a yes,but in the math. Reread the original post, especially the conclusions. In order to get the super-high probability of success certain conditions must be met above and beyond just "working together".

- The Skill Challenge and DM must allow "Aid Another". This is hardly a given. Certain skills just don't lend themselves to it

- Super skill monkeys are necessary. Not just someone optimized for skills, but someone who puts everything into ONE skill, and hopes that is the one needed.

Basically the fact that the percentages go so easily and quickly from less than 10% to approaching 100% proves that the system is unmanageable for a DM. I NEED to be able to use my own small bonuses to manage success and failure in a (mostly) predictable fashion. I can in fact do that in Combat encounters.
I want the underlying math in the system to be stable, so that I don't need to think about it too much.

 

[Hussar]

It boils down to this...do you feel if a player uses a less than optimal spell, that the entire combat should be lost?

For example, the difference in two people and 3 people aiding another is a HUGE difference. With a party knowledgeable the rules, they should kick the crap out of the guy who wants to roll his skill check instead of aiding.

Overall though, the bonuses are not so huge. You're only talking a +2 difference. A single magic item, stat, or other modifier has an equal or greater bonus.

Now, I agree that you would want people to aid. I read though, that it's not an either/or thing. Am I mistaken in that. I thought Aid was a minor action, meaning that you can Aid and still make your skill check.

But, then again, if you look at combat, I'd kick the crap out of the guy who decides to use a dagger instead of his main weapon "just cos". Aid another is a good tactic. We have a system that is rewarding teamwork.

In game situations are going to limit when and where you can Aid Another and other bonuses you can accrue. There is a wide range of situations where you won't get the +8, but might only get +4 because of whatever. And, that modifier will likely change for many skill rolls as well. You might only get 2 people helping you the first time, because the other three are engaging that troll. However, the wizard backs up and helps the next guy, giving a +6 to the next roll.

And, it might add a great deal of tension to the scene as well. The party could try to do it with a straight up roll the first time and fail. Now they cannot fail again, so they drop everything to aid the next four rolls, meaning that the tactical situation becomes much more dangerous.

 

[Stalker0]

Overall though, the bonuses are not so huge. You're only talking a +2 difference. A single magic item, stat, or other modifier has an equal or greater bonus.

That +2 difference....equals 25% extra win rate for the party.

The reason is its not just the +2 difference, its the fact that you aren't have to make a check that determines a failure or success.

In the most optimal scenario, the person with the highest modifier is making all the checks, while the party is providing them a big bonus. Basically aiding eliminates all of the "medocrity" in a skill challenge.

Btw, I'm not saying aid isn't a good mechanic. The importance of teamwork is....welll...important!!

The problem is the system is horrible at handling aid another. The variability is way too large. It doesn't encourage teamwork, it absolutely demands it with a blunt object.

 

[gonesailing]

And, it might add a great deal of tension to the scene as well. The party could try to do it with a straight up roll the first time and fail. Now they cannot fail again, so they drop everything to aid the next four rolls, meaning that the tactical situation becomes much more dangerous.

I think you are thinking of a single skill check not a skill challenge that requires X successes before Y failures. These things can take many many shapes and sizes and not all lend themselves to Aid Another. The system has to work regardless

 

[Hussar]

Hussar there is a yes,but in the math. Reread the original post, especially the conclusions. In order to get the super-high probability of success certain conditions must be met above and beyond just "working together".

- The Skill Challenge and DM must allow "Aid Another". This is hardly a given. Certain skills just don't lend themselves to it

- Super skill monkeys are necessary. Not just someone optimized for skills, but someone who puts everything into ONE skill, and hopes that is the one needed.

Basically the fact that the percentages go so easily and quickly from less than 10% to approaching 100% proves that the system is unmanageable for a DM. I NEED to be able to use my own small bonuses to manage success and failure in a (mostly) predictable fashion. I can in fact do that in Combat encounters.
I want the underlying math in the system to be stable, so that I don't need to think about it too much.

But Stalker0 doesn't account for magic items or rituals or class abilities. The lack of a skill monkey and Aid Another can be taken up that way as well. There are a number of rituals and magic items that give skill bonuses, frequently more than a simple +2.

In the most optimal scenario, the person with the highest modifier is making all the checks, while the party is providing them a big bonus. Basically aiding eliminates all of the "medocrity" in a skill challenge.

Isn't that pretty much saying that in the most optimal scenario, the party is "taking 20", to use a 3eism?