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

[Jasperak]

I wonder what is going to happen when we see Skill Challenges in published modules. Is a DM going to have to adjust them for his players. Seems kinda counterproductive to have presumably solid math behind combat challenges, but clearly disturbing trends in the math behind SK if the OP is correct with his analysis.

 

[Eldorian]

The actual skill challenge with the undead knight was much more reasonable (ie, they intend for the groups to succeed, unlike the last one where they do not). My group isn't at that point yet - did most groups fail that one?

Check out table 3, it deals with -5 DC, which the knight encounter has, due to it being "easy".

Though the OP used a +9 as his base, and that's high, you should still see at least a good chance of winning.

 

[SKyOdin]

It looks like the OP is assuming only Normal DC skill checks. How would a mix of Easy, Normal, and Hard DCs affect the outcome? If only a few points variation can swing the chances of overall success, then changing the DCs by 5 or so points between different checks in the same challenge should have a noticeable effect on the outcome.

This may have big implications for skill challenge complexity too, since a higher complexity challenge has more room to include Easy and Hard checks.

 

[Skywalker]

I wonder what is going to happen when we see Skill Challenges in published modules. Is a DM going to have to adjust them for his players. Seems kinda counterproductive to have presumably solid math behind combat challenges, but clearly disturbing trends in the math behind SK if the OP is correct with his analysis.

Though I assume the OP's math is solid, he has not provided a comprehensive analysis. I think his assumption of +4 Attribute for all Skill rolls made and that Aid Another will always be maxed out are extremes. Also, we haven't seen what an Easy or Hard Skill base would look like, even though many SKill Challenges contain them.

As such, some middle ground analysis would be very useful. Remember that every Skill Challenge will be a combination of all these elements.

 

[Stalker0]

As such, I would love to see the stats for having just 1 or 2 PCs Aiding, rather than just the extremes.

Your wish is granted sir, table 4 has been added, showing the effects of 1,2, and 3 people aiding another.

 

[Skywalker]

Your wish is granted sir, table 4 has been added, showing the effects of 1,2, and 3 people aiding another.

You sir, are a legend. Any chance of a +2 chart?

Actually I think PWR(2) and PWR(3), particularly the later show some pretty decent ranges.

It also shows just how hard this thing is to analyse given there are so many variables to take into account.

 

[Dave Turner]

They do. Basically in this case the players are using the actual DC written in the table, instead of the "real" DCs.

Ok, so help me with this scenario:

Let's assume a conscientious DM who wants to give each character a chance to shine (meaningfully contribute) in every skill challenge. Table 3 shows that, with the +5 DC footnote stripped out, a skill challenge with 5 "medium" characters has very respectable chances of success, both on the individual and overall level.

On average, a 4e PC will have 4 trained skills and, thanks to ability and racial modifiers, probably be "medium" in at least 2 of them. So in a party of 5 PCs, we'll have ten areas of "medium" competency. We might assume some overlap, like three characters who are all "medium Bluffers", but that shouldn't affect my hypo as I understand it.

Our hypothetical DM designs a skill challenge that focuses on at least one skill that one player has at "medium". For our hypo, it's a social challenge in which the DM chooses Bluff, Diplomacy, Insight, and History (we've got two medium Bluffers in the party, so the DM can kill two birds with one stone by choosing Bluff). The DCs for these medium skills are set at the moderate level according to the DMG table without the footnote, i.e. the assumptions of Table 3.

What we seem to have is 5 medium skill characters who are attempting to overcome a skill challenge. It's not a case of 1 medium skill character who is, for sake of argument, skilled in Bluff, accompanied by four "bad" Bluff characters scrambling to Aid #5's Bluff (recall that in order to Aid Another, one must roll the same skill that one is trying to Aid). We've got 5 medium skill characters, rolling independently and without Aid, who have very decent chances of success according to Table 3.

In other words, if we assume a conscientious DM who tailors a skill challenge to his party using the "unfootnoted" DCs, the problem seems to be substantially reduced. If it's reasonable to assume a conscientious DM, then doesn't Ockham's Razor point us to an innocent error by WotC with the +5 DC footnote?

EDIT: I just remembered, Stalker0, that your assumptions are for a Complexity 5 challenge (12 successes v. 6 failures). Do the probabilities worsen, for ALL tables, if the complexity starts to drop?

 

[Eldorian]

It also shows just how hard this thing is to analyse given there are so many variables to take into account.

But it's not, really. What Stalker0 is doing isn't hard. I taught a course in exactly how to do this stuff last semester, aimed at university freshmen and sophmores. A course for business majors, not mathematicians, computer scientists, or statisticians. A decent computer scientist could quickly write a program that generates all these tables in less than an afternoon. Generates all and more. Hell, some guy delurked and posted this: http://www.enworld.org/showpost.php?p=4279427&postcount=32

and that's all you really need to take from the mathematics perspective.

Edit: and no offense to the OP, he's actually trying to fix it, which may very well be hard, based on how messed up it currently is. Fixed and fun.

 

[Korgoth]

I'm stunned.

Again.

You are really flushing them out of the woodwork with these threads, Stalker. I'd say it's a lost cause, quite frankly.

You are educated stupid and evil.

 

[Stalker0]

It also shows just how hard this thing is to analyse given there are so many variables to take into account.

oh nothing a few thousand calculations can't do

But here's the take home message, the system can't be used reliably if its balance is so precarious on many things taking place. We can point to the table and say, "hey, if two people are aiding every round, the system is perfect!" But what happens to the people who don't allow aid for a skill challenge, but about the people who want to let all their players aid for a different skill challenge? The answer is nearly complete failure, or near total success

The system is just too swingy to account for those many variables, and believe me its a painful thing to reverse. In working on my new system, I have spent hours and hours toying with various ideas just to get the variation down by half! It can be done, but I had to make some strong changes to the system for that to happen.