D&D 4E Showing the Math: Proving that 4e’s Skill Challenge system is broken (math heavy)

[Torquin Trebuchet]

Well, it shouldn't take statistical analysis to convince the average D&Der that rolling 11+ four times before you roll 10- two times is going to be a hard road. Hopefully we've all rolled a d20 enough to understand that much, at least - but perhaps hope springs eternal when it's your d20.

I think giving the PCs a lower roll if they really justify their skill use and come up with something creative is a good start. In my game, we ran an impromptu skill challenge when the party started losing the Irontooth fight. They had managed to kill Irontooth himself, and wanted to parlay with the remaining kobolds. I had them roll skill challenges without telling them the DC, but knowing in my mind that I would set it based on the aptness and creativity of their skill use. They made two 15s, missed a 20, sensed that things were tense and offered a big bribe (auto success) which also earned them another 15, which they passed. It was a lot of fun, and felt a lot less deus ex machina than it would have without the rolls and the structure.

Good to bring this point up, but really a very easy fix for DMs.

 

[pemerton]

Your question is answered in the first post. Initiative order is irrelevant.

This can't be right.

Suppose that there are 5 PCs taking part in a Complexity 1 challenge.

If the two weakest PCs go first, the likelihood of getting 2 fails before 4 successes is greater than if the two weakest PCs go last.

A limiting case that illustrates the point: if 2 PCs have a 0% chance of success, and they go first, then the chance of that party succeeding at that skill challenge is zero.

 

[pemerton]

Everyone still questioning which bonuses apply where can save themselves a lot of trouble by going back and reading this overlooked but important post.

http://www.enworld.org/showpost.php?p=4279427&postcount=32

There is one problem with that table: its columns 1 to 5 correspond to Skill Challenges of Complexity 0 to 4 (ie 2/1 up to 10/5).

When you ignore its column 1, which is irrelevant, the degree of swinginess is quite a bit less. Tolerable, even.

 

[pemerton]

fundamentally the idea of a skill challenge is that everyone participates, and everyone rolling pretty close to an autosuccess to boost the one guy who has a chance is boring.

You may have missed this part of the post:

Now -- what about aid another. If aid another is "I roll to help out" that's boring, if aid another is players actively bolting on other points to the lead PCs act "My lord, it is only just that we do this!" says PC1, PC2 nods, interjecting "and its what your grandfather would have wanted you to do boss" -- man -- how boring.

That is, properly narrated and integrated into the action, aiding another need not be boring at all.

And Abeattie: another good contribution to the thread.

 

[Wulf Ratbane]

This can't be right.

Suppose that there are 5 PCs taking part in a Complexity 1 challenge.

If the two weakest PCs go first, the likelihood of getting 2 fails before 4 successes is greater than if the two weakest PCs go last.

A limiting case that illustrates the point: if 2 PCs have a 0% chance of success, and they go first, then the chance of that party succeeding at that skill challenge is zero.

It is right, because he is only analyzing the cases where the PCs win.

We don't care how many permutations there are for failure; we only need to find out how many permutations (Y) there are for success. The number of failures is obviously (1-Y).

A 4 success / 2 failure challenge is over in 5 rolls or less, every time, and every possible permutation of success will be some combination of 4 success and (0 or 1) failures:

SSSS
FSSSS
SFSSS
SSFSS
SSSFS

Any permutation with 2 failures in it is a Fail of the entire challenge.

 

[Kinje]

Wulf,

You're both correct - to an extent.

The poster's original math ignored the order in which your failures occurred, so long as you didn't have more than the number required to actually fail.

It also, however, assumed optimal skill check numbers for all people rolling.

pemerton's point is that in cases where not all party members have an equal chance at success, initiative order can play a huge role in determining the outcome of a Skill Challenge.

Example: You have a party of six. Six people have a chance roll successes, two people cannot possibly do so (They are untrained and have an 8 in the associated stat, netting them a -1 to their D20 chance to roll a DC 20 check. Guaranteed failures.) If the two automatic failures go anywhere other than dead last, the party cannot succeed. Period. This isn't taken into account in the OP's math, because he is assuming fairly optimal conditions for the PCs, but it is a valid concern in regards to the system as a whole.

 

[Stalker0]

pemerton's point is that in cases where not all party members have an equal chance at success, initiative order can play a huge role in determining the outcome of a Skill Challenge.

Very true, I did look at this in my analysis, but I didn't post tables for it, as that's a lot a lot a lot of tables I would have to create. The potential "buffer" to this problem is aid another. Theoretically, a person with the worst skill check would be the most likely to try and aid, as his roll has the weakest chance of success. However, aid has such a huge powerful effect on the system it doesn't make a good buffer.

I have tweaked it greatly in my new system, and have found it greatly limits the variance with initiative.

 

[webrunner]

Wulf,

You're both correct - to an extent.

The poster's original math ignored the order in which your failures occurred, so long as you didn't have more than the number required to actually fail.

It also, however, assumed optimal skill check numbers for all people rolling.

pemerton's point is that in cases where not all party members have an equal chance at success, initiative order can play a huge role in determining the outcome of a Skill Challenge.

Example: You have a party of six. Six people have a chance roll successes, two people cannot possibly do so (They are untrained and have an 8 in the associated stat, netting them a -1 to their D20 chance to roll a DC 20 check. Guaranteed failures.) If the two automatic failures go anywhere other than dead last, the party cannot succeed. Period. This isn't taken into account in the OP's math, because he is assuming fairly optimal conditions for the PCs, but it is a valid concern in regards to the system as a whole.

If I know i'm a terrible speaker, and i'm in a diplomacy-based skill challenge, and came up first, I'd either hold my action, come up with something i could do (to give a + bonus), or expect to blow the whole thing.

For a more specific example: If you're trying to bluff your way past some guards, and the truthful paladin comes up first, and says "No we're really just here to murder your lord".. I'd expect that to become a combat situation very quickly.

 

[Wulf Ratbane]

Wulf,

You're both correct - to an extent.

The poster's original math ignored the order in which your failures occurred, so long as you didn't have more than the number required to actually fail.

It also, however, assumed optimal skill check numbers for all people rolling.

pemerton's point is that in cases where not all party members have an equal chance at success, initiative order can play a huge role in determining the outcome of a Skill Challenge.

I'm comfortable with best case analysis. Based on the context of his posts, I certainly couldn't have anticipated that proving that the Skill Challenges are even more horribly skewed against the PCs was his intent, or even necessary.

I guess I'm a "Glass is 25.62175% Full" kind of guy.

 

[Regicide]

If the two weakest PCs go first, the likelihood of getting 2 fails before 4 successes is greater than if the two weakest PCs go last.

Yes, another reason why skill challenges are busted. All PCs must roll a skill and they must do it in initiative order.