Community

Is there a question tucked away in there somewhere?

I was going to use KotS Area 8: Sir Keegan's Tomb as an example where several different skills can be used to gain successes and get all the characters involved, which seems like a good way to run these to me.

Then I noticed that particular skill challenge doesn't follow the rules. Four successes before four failures...

This has been discussed ad nauseum over at EN World, and a little bit on the Rules Q&A forum here. Short answers: yes the math is totally broken, no we don't have any clue how it got through WotC playtesting, and yes it can be fixed, but only with great difficulty. By far the best tested and most mathematically sound fix yet proposed can be found here.

I agree that at lower levels, it is much more difficult to beat a Skill Challenge that at higher levels. Complexity has little impact, except for at lower levels.

The practical suggestion for DMs is to manage the skill DCs based on how well your players are min/maxed. If you want them to get on with the show, go with easy DCs and circumstance bonuses. If you want the encounter to be a nail biter, go with moderate DCs with easy DCs if for the win. Hard DCs are for obscure uses of skills or the really far-fetched.

Not wanting to create the specific case math function for the higher complexity cases, I created an Excel sheet to simulate the success vs. failures in a Skill Challenge.

Hopefully, the math below helps DMs to weight their options and work with the system.

The simulation results below do not account for skill boosting magic items, but it does assume that the character is increasing the ability bonus by +1 per 8 levels.

Level, 1, 6, 11, 16, 21, 26, 30
Difficulty, easy, easy, easy, easy, easy, easy, easy
Skill Bonus, 7, 10, 13, 17, 19, 23, 25
DC, 15, 18, 22, 25, 27, 29, 30
4 before 2, 40%, 39%, 29%, 38%, 38%, 62%, 71%
6 before 4, 37%, 39%, 26%, 37%, 37%, 69%, 81%
12 before 6, 29%, 32%, 19%, 29%, 28%, 61%, 76%

Level, 1, 6, 11, 16, 21, 26, 30
Difficulty, mod, mod, mod, mod, mod, mod, mod
Skill Bonus, 7, 10, 13, 17, 19, 23, 25
DC, 20, 22, 26, 29, 31, 33, 34
4 before 2, 6%, 11%, 7%, 11%, 11%, 23%, 28%
6 before 4, 2%, 4%, 2%, 4%, 4%, 16%, 24%
12 before 6, 1%, 2%, 1%, 2%, 2%, 11%, 17%

Level, 1, 6, 11, 16, 21, 26, 30
Difficulty, hard, hard, hard, hard, hard, hard, hard
Skill Bonus, 7, 10, 13, 17, 19, 23, 25
DC, 25, 26, 30, 33, 35, 37, 38
4 before 2, 0%, 1%, 0%, 1%, 1%, 4%, 7%
6 before 4, 0%, 0%, 0%, 0%, 0%, 0%, 2%
12 before 6, 0%, 0%, 0%, 0%, 0%, 0%, 1%


The text above in in "comma delimited format". To make sense of it, copy it into a word processor and convert text to table using the comma as you delimiter.

I'm actually starting to wonder whether the table on P61 and the reference in the example on P42 are errors left over from a previous revision. The table on P61 is just floating there without any associated rules text around it and the reference in the P42 example seems to be the only place it talks about adding +5 for a skill check - unless there's some actual rules text about doing so that I'm missing.

If you just use the basic DCs shown on P42 and ignore the out-of-the-blue +5 in the example, the DCs for skill challenges actually work out reasonably well - but you need to bear in mind some caveats.

Suppose a 1st level character has a +8 modifier to a skill they're good at. +5 for Trained, and +3 for a good stat (making allowances for a racial +2 in there somewhere as well). A DC 15 (moderate) check requires a 7 or higher, which gives a 70% chance of success.

Now, you need an 80% success rate (4 successes, 1 failure), so they seem to be behind, but the walkthrough of skill challenges shows several ways that the odds may be tilted towards the players: Most of them allow for at least one easy check to be made, most of them allow you to make a "non-risky" check of some sort to add +2 to another character's check, and the dialogue with the Duke example shows that throwing in a +2 for good roleplaying of the check is also an accepted practice. The one easy check is almost certain to work if someone is trained in it, and all of those +2s which you can apply also help to push the odds to something more favorable to the PCs - so in the end, the 80% success rate is reasonably achievable. Not guaranteed by any means, but you've at least even chances.

Higher complexity actually helps your odds in this case - six successes/three failures means your success rate only needs to be 75%, while eight successes/four failures pushes it down to about 73%.