Ruling Skill Challenges: Rule #4
by Mike Mearls
As Mike delves more deeply into the process of building a skill challenge, he realizes that it's something of a skill challenge itself. And the deeper you get into it, the more challenging it becomes.
Talk about this article here.
* - Probably meant to be Dungeon 162
However, I find myself annoyed, as a News Guide by the wrong issue number. Second time thats happened this month.
Also, This series of posts started in Dragon, then switched over to Dungeon.
Minor annoyances though. I can easily overlook them as long as the content of the article itself is good, and so far, this series has delivered for me.
I liked the advice in this article.
To many skill challenges are simply set up as a series of skill checks (its done in Dungeon as well) and that, as these articles ably point out is the worst thing to do with them.
Again good advice from Mr. Mearls.
These articles have been explaining how Skill Challenges should actually work
My question though: when you're running a long Skill Challenge (such as the example with the Baron in the article), who actually makes the check, and how many? Do you just have a single PC make the check for, say, interacting with an individual PC? Does the entire party basically "aid another?" Or do you involve lots of checks that create bonuses or penalties, but only have one actually count as the success or failure?
I'm trying to imagine a short feast scene. If this is only a Complexity 3 Challenge, than the feast in the middle would be 2-3 successes, which means only half the party would get to make checks that matter. Or should all Skill Challenges basically be Complexity 4 or 5?
I kind of like making each "success" a complex skill check, so that lots of people can make lots of checks in order to determine if they've succeeded with a particular event. So the rogue and the warlock breaking into the Baron's room doing the feast get to make a bunch of stealth and thievery checks (maybe along with a few bluff checks) to get in, get the info, and get out--but in the end that only counts as 1 or 2 successes. Does that sound about right to people?
My question though: when you're running a long Skill Challenge (such as the example with the Baron in the article), who actually makes the check, and how many? Do you just have a single PC make the check for, say, interacting with an individual PC? Does the entire party basically "aid another?" Or do you involve lots of checks that create bonuses or penalties, but only have one actually count as the success or failure?
...
This is an excellent question. I've no doubt the DMG2 will address exactly this sort of problem, though we can hope for an answer before then. I generally either have the characters roll initiative (if it's fast paced like a battle) or just go around the table. Each player has the option of making a primary check, secondary check, or passing. This approach might be problematic if you have people abusing the pass or secondary skills to minimize the chance of a failure. I have a solution for that as well, but that's another story...
It's certainly possible to just let the group appoint a representative for each check, but I don't think you'll get the overall participation you're looking for that way.
I'm in a rotating DM game and summarized what I've learned from these columns so far:
*Plan for lots of diverse skills. The first two articles examine this in detail.
*The skill challenge should cover a long time period in-game.
*Players should actively decide what they want to accomplish and should then be given a subset of the challenge skills which they can choose from to try and accomplish their task.
*When the situation calls for a skill check all characters must make one (although they don't all need to use the same skill). If the majority of players succeed they gain a success, otherwise the party gains a failure. Each player may also (in addition to making their own check) attempt to assist one other character.
*Don't end the skill challenge at 3 failures. Rather, begin applying penalties at that point, which get worse with subsequent failures.
*Design options that can give bonuses or cancel out failures.
*Don't end the skill challenge at a specified number of successes. Rather have each individual success linked to a specific action taken by the group.
This will all make a lot more sense if you read the latest article with the full example. These are major changes to how skill challenges function.
Finally, the articles do not mention this, but the DC chart has changed yet again. We are only given the DCs for a level 9 encounter: Easy 14, Standard 19, Hard 24. These do not match any previous version of the DC chart. Furthermore, the errated version of the DC chart has been removed from some of the online erratas. This tells me that the designers are paying more attention to this and trying to actually fix the math errors that plagued the first two versions (which were automatic failure in version one and automatic success in version two). The designers haven't released the new chart yet, but I think it is going to be something like this:
Easy DC equals 10+1/2 level (rounded down as always).
Standard DC equals 15+1/2 level.
Hard DC equals 20+1/2 level.
Practically, this gives a character using a skill that he is untrained in that uses a "bad" stat a roughly 50/50 chance of success at the easy difficulty. Characters with a good stat or training have a roughly 50/50 shot at standard difficulties. Characters with a good stat and training have a roughly 50/50 shot at hard difficulties. (Add in magic item/racial bonuses to improve your chances further.) This seems to be a good system because all characters must make a check so the benefit of maxing out one skill at the expense of others is no longer an optimum decision.
One thing to keep in mind - the DCs I used aren't changes to the core rules. I'll talk more about this in the next column when I design a skill challenge as requested by readers, but I don't see the DCs as canonical, in the sense that you *must* use those DCs.
The elevated DCs I used are there for a couple reasons:
1. In most cases, the PCs have the option to chose a lower DC skill check.
2. The really high DCs sometimes show up when the group can gang up on a skill check with the aid another action.
3. Since failure is a sliding scale, I made the DCs higher than normal to compensate.
Really, though, when doing DCs for your home game use the table as a starting point. It's a great place to pull a DC when you need one on a moment's notice or as a guide to what the numbers should look like, but they aren't hard coded into the game.
Some of the other techniques I pulled out, like party checks, are also features of this skill challenge that you could use in other challenges, but they aren't necessarily the One True Way to do things.
Think of skill challenges almost like you might think of building a campaign setting or the story line behind an adventure. There are a lot of different methods you can use, both mechanical and literary. The choices you make depend more on the effect you're going for rather than a rigid formula.
I've mentioned this before, but previous attempts I've made have simply been about 5 minutes of dice rolling and success counting, with some brief narration of actions by the players. They know it's a challenge, they narrate something (sometimes simple "I scout ahead, looking for a passable trail through the dense forest" and sometimes interestingly complex: "In my scouting, I stumbling upon a band of gnolls fresh from the kill..."), and they make a dice roll, and it goes in the success or failure column.
It's phenomenally boring. So what is it like for other people? That's the part I have the most trouble with: feeling like it's worth the time taken to roll dice, rather than simply doing a freeform interaction (or worse, saying "Somehow you figure out that EvilDude is behind it, and you confront him at his lair beneath the Chuck Wagon").
I set out a battlemat, drew the tower, placed a few bandit guards, and had the PCs move in initiative order. Each time the PCs failed a check, more bandits came out of the tower, the bandits started moving closer to their hiding positions, and so on.
This set up also allowed for a lot of near misses. For instance, when a PC made a Stealth check, I spoke in-character for the guards and said stuff like:
Guard 1: "Hey, did you hear something?"
Guard 2: "No. You're awfully jumpy tonight."
Guard 1: "I think someone's out there." (Points toward grove of trees the wizard is hiding in.)
Guard 2: "We're in the middle of nowhere. The only thing out there is a raccoon or an owl. Now shut up and get another log for the campfire."
The nice thing was that I could use that framing device even for successful checks - the players didn't know if guard 1 would convince guard 2 to check out the grove.
I think that one of the important things about using skill challenges is figuring out when to use them. Sometimes, a Streetwise check to find the thieves' guild is just a skill check.
Somewhere around the fifth or sixth skill check, my daughter narrated the appearance of a pack of gnoll marauders (to my great surprise), just to have something interesting happen.
I think part of that is that the challenge itself had nothing particularly unpleasant riding on it (a surge or two? meh), but I also didn't really do much with it, either.
I'll have to work on making each die roll a tense moment - have something important riding on it.
The key lies in building tension into each die roll. One of the best skill challenges I ran involved the party trying to sneak up on a tower that some bandits were camped inside of.
I set out a battlemat, drew the tower, placed a few bandit guards, and had the PCs move in initiative order. Each time the PCs failed a check, more bandits came out of the tower, the bandits started moving closer to their hiding positions, and so on.
This set up also allowed for a lot of near misses. For instance, when a PC made a Stealth check, I spoke in-character for the guards and said stuff like:
Guard 1: "Hey, did you hear something?"
Guard 2: "No. You're awfully jumpy tonight."
Guard 1: "I think someone's out there." (Points toward grove of trees the wizard is hiding in.)
Guard 2: "We're in the middle of nowhere. The only thing out there is a raccoon or an owl. Now shut up and get another log for the campfire."
The nice thing was that I could use that framing device even for successful checks - the players didn't know if guard 1 would convince guard 2 to check out the grove.
This sounds excellent, but how did the challenge end? Did the guards notice the PCs and combat started on the third failure, or just when it seemed appropriate? Did the PCs have to make 12 checks to get into the tower, or enough checks based on how fast they moved across the grid? When and how did you resolve the encounter?
I think the Skill Challenge is an excellent framework for encounters, but some vagueness is hidden behind codified rules in the book. It leaves me wondering: Are challenges are supposed to be more free-form for when they start and end (rather than X successes or Y failures)? How exactly do you calculate "appropriate" DCs for your party? It's stuff like this that I hope to see addressed in your next article (hopefully with a more "standard" example), or perhaps in DMG2 (though including revisions in the errata so everyone can get them for free would be nice ;p).
Thanks a ton Mike!
PS. I absolutely love the write-up for this weeks article! I'm scheming how I can work that into my own campaign ASAP. :D
This sounds excellent, but how did the challenge end? Did the guards notice the PCs and combat started on the third failure, or just when it seemed appropriate? Did the PCs have to make 12 checks to get into the tower, or enough checks based on how fast they moved across the grid? When and how did you resolve the encounter?
There were two things going on:
The players basically decided when to end the challenge by attacking. Alternatively, they could've just spied on the bandits and slipped away. In the adventure, they had to prevent the bandits from ambushing a caravan. They could've learned the bandits' plans and escorted the caravan along a safe route, attacked them, or something else.
In the actual game, the PCs managed to avoid failure and sneaked into a great ambush position. The failures worked like this:
1 Failure: More guards wake up and put on their armor. A patrol starts walking the tower's perimeter.
2 Failures: All the guards are now awake, and a second patrol starts.
3 Failures: The bandits alert their leader, and he organizes them.
The PCs hit two failures, but were able to ambush and defeat the leader before he woke up. In addition, the leader is alone in his room until the bandits hit 3 failures (the guards leave him alone for fear of waking him up for a false alarm).
At 1 or 2 failures the PCs could've escaped, but at 3 the bandits would've tried to chase them down.
Thanks for the reply.
Playing in a rotating DM game I've seen quite a few skill challenges and some have worked better than others.
We had one DM run a skill challenge that was meant to be a chase. In retrospect the DM probably made the mistake of drawing a grid, which prompted an Eladrin to use their teleport power instead of a skill and pretty much ended the skill challenge before it began.
I used three skill challenges when I ran an adventure. The first went okay, the second not so much and the third actually worked really well.
The first was a wilderness chase where the players were trying to move through a mountain pass while avoiding bands of goblins and kobolds. I allowed the PCs to use skills to find sidepaths, outrun the enemies, move through more challenging areas, etc. While the encounter had a satisfactory conclusion (the party got a group of goblins and kobolds to run into each other, allowing the party to proceed without further trouble) there were a few problems. Some of the checks, such as outrunning a hunting party, I would have everyone roll an athletics check at a very low difficulty because if one character was too slow it would hold back the entire party. For things like nature checks I would have everyone roll a high difficulty because I figured that if one character succeeded they would pass the information on to the others and that would be a success. This led to a third type with a moderate difficulty where one character would attempt to climb a cliff and lower a rope for the others. So there was only one roll and its success or failure determined whether a success or failure was accrued.
The problems with this were two-fold. First, the enemies were too passive. They were basically just waiting for that last failure to come up in order to pounce. So the encounter ended up not being nearly as suspenseful as it could have been (at least the background music helped). The second problem, was that the PCs constantly had the same set of options and quickly figured out what they were good at and just kept repeating it. Again, this would have led to a very static encounter if I hadn't been able to improvise some additional flavor to alter the situation.
Having read your article, if I was to do it again, I would have had the skill challenge represent the whole journey through the pass, (which took a few days anyway) rather than just what was supposed to be a dramatic chase near the end. I would have employed a larger set of skills and come up with penalties for gradual, rather than abrupt, failure and added random events. Most importantly, I would have used your method for having all the PCs roll and giving success if a majority succeeded (while letting them choose the skill). It would have been a far better encounter.
The second one I ran was a negotiation that I basically realized part way through wasn't working properly as a skill challenge. I abandoned the skill challenge aspect and just let the PCs RP their way to success.
The final one went quite well, but it occurred during a combat. I think the reason it went well, was because all the players had something to do (they could fight or use skills). Further, failure just meant the fight continued until the PCs got the successes they needed. Further, each check was tied to a specific action the pcs were taking and it had a big impact on the battle. The tension definitely went up as the fight/skill challenge went on because the PCs were losing hp and the enemy becoming increasingly vicious as it tried to stop the PCs from destroying its portal.
After my turn DMing, I've seen a few more skill challenges from other DMs. We had one DM do a very good skill challenge that involved the players planning to break into a building. It went quite well, although I think it would have gone even better if some of your ideas, in particular having everyone roll each time and increasing the tension by having progressive failure. He also attempted a negotiation skill challenge and that one fell apart the same way mine had. He also ended up scrapping the challenge half-way through and rping it out to its logical conclusion.
We've also had another DM who attempted a skill challenge to have the PCs enter a lair undetected. This involved climbing down a slippery passage using athletics. Athletics was the only option unless someone fell, in which case other skills were used to catch the person and negate the failure. Most of the party wasn't particularly athletic and we all fell, alerting the guards and taking quite a bit of damage in the process.
At any rate, that has been my experience with running and playing skill challenges since 4e came out. Overall everyone at my table has felt there is an enormous amount of potential but we haven't been able to get it quite right. Your article has made me quite excited to give skill challenges another go. The suggestions really get to the heart of the problem we've been having.
I do have a question for you. Can you elaborate on how assisting works in your skill challenge? You say:
"The PCs can each use a different skill, provided each skill is allowable for that action. Each PC can also aid one
other PC. One PC can receive aid from more than one ally."
I assume that assisting is in addition to making a check themselves. How do you run this though? Lets say you are traveling in your current quarter. 1 PC makes a stealth check, 3 make athletics checks and 1 a streetwise check. Do you allow the characters making different types of checks to assist each other or only the ones using the same skill? What do the players roll to assist? Do they use their initial roll and if it beats the assist DC do they succeed and give someone else a +2 or do they make a second roll? What DC do you use for the assist? (The DMG has a static number, but it becomes automatic by level 30, so I'm assuming most people house rule a different DC.)
Oh, and by the way, I really appreciated your analysis of the skills by class and role. I had done a similar thing and found it very illuminating.
I just sent in an email with a suggestion for a skill challenge for Mike to design in the next article. I'm going to post it here as well since I think it could generate some interesting discussion. (Apologies if this is inappropriate!)
-----
The challenge: the party has been unexpectedly transported to a ruined city deep in a trackless desert. They must find their way back to civilization without succumbing to hunger, thirst, environmental hazards, or just plain getting lost.
I'm finding that this is a difficult challenge to design because it pretty much precludes the use of social skills. The example skill challenge from Mr. Mearls' article makes heavy use of social skills, but in this situation, there just aren't any NPCs to interact with. When you remove social skills right off the bat, it's much harder to meet his guideline of including "2 + the number of characters in the party" skills in the challenge.
Additionally, I have a question about his advice for skill selection. The advice in the first article regarding "common" vs. "rare" skills, and skills by role, seems appropriate for a module designer who is making up the challenge for an unknown party. A DM crafting a challenge for his own table generally knows the composition of the party and exactly which skills they do and do not have available. Why not simply pick skills for the challenge off the list of skills the party actually knows? Or perhaps most from the list of known skills, with one or two not on that list to force the occasional untrained roll?
Best,
Ryan
PS -
Here were some of my initial thoughts about what skills I might choose for this challenge, showing why I find it difficult when the social skills are thrown out.
The overall structure of the challenge would include a section in the ruined city followed by a section in the desert proper.
The main tasks or "subchallenges" I had in mind were as follows:
1. Find food and water. (both in the city and the desert)
2. Find useful equipment (maps, tents). (city)
3. Avoid/survive hazards like sandstorms, sunstroke, predators, etc. (mainly desert, maybe a little bit in the city)
4. Navigate successfully out of the desert or to a settlement in the desert. (desert)
Skills and which tasks they apply to:
Nature (primary for #1, #3, #4)
Athletics (primary for #3)
Endurance (primary for #3; mitigate failures for #1, by being able to eat and drink less)
Heal (mitigate failures in #1 and #3)
History (city only, primary #1 and #2)
Perception (primary for all tasks)
So there I have six skills which aren't too much of a reach... but if the party includes more than 4 players, we need more skills to meet the "# of players + 2" guideline. I've also only found 2 skills for the "navigation" task.
I can squeeze in Religion and Arcana if I add additional factors to the basic setup of the challenge. Say that the ruined city is ruined because it was cursed to be swallowed by the desert and its inhabitants cursed to get lost in the desert so that they always circle back to the city. Now maybe Religion might help figure out ways around the curse. Say that the city used to have enchanted roads leading to it from the outside world. These roads are now swallowed up beneath the sands, but by using Arcana a character might be able to detect their faint magical traces.
So I've had to modify my setup/setting in order to accommodate more skills. That's not necessarily a bad thing, but it's something it might be interesting to discuss in a future article...
I do have a question for you. Can you elaborate on how assisting works in your skill challenge? You say:
"The PCs can each use a different skill, provided each skill is allowable for that action. Each PC can also aid one
other PC. One PC can receive aid from more than one ally."
I assume that assisting is in addition to making a check themselves. How do you run this though? Lets say you are traveling in your current quarter. 1 PC makes a stealth check, 3 make athletics checks and 1 a streetwise check. Do you allow the characters making different types of checks to assist each other or only the ones using the same skill? What do the players roll to assist? Do they use their initial roll and if it beats the assist DC do they succeed and give someone else a +2 or do they make a second roll? What DC do you use for the assist? (The DMG has a static number, but it becomes automatic by level 30, so I'm assuming most people house rule a different DC.)
Yes, the assist attempt is in addition to the check and uses the same skill as the check being assisted. In your example, if two PCs wanted to help the guy making the Stealth check, those two would make Stealth checks to assist, and their regular skill check for the challenge.
Right now, I'd use the static DC, but it isn't a bad houserule to use the check's DC -5 or -10.
I've definitely found that skill challenges are a bit of a trial and error thing at this point. I like the sound of the one you ran during a battle. It'd be fun to play.
Why not simply pick skills for the challenge off the list of skills the party actually knows? Or perhaps most from the list of known skills, with one or two not on that list to force the occasional untrained roll?
That's probably the best way to pick skills for your home group, but I've also found that many DMs rely on what seems sensible or logical for a situation, rather than what might be best for the group. I was hoping that the discussion would illustrate how some skills that seem like obvious, common skills for challenges are actually rather rare for PCs.
I really like your example, particularly how you look at justifying adding Arcana and Religion to the challenge by adding some cool, flavorful elements to the situation. The idea of a city cursed by the gods, that had some sort of arcane element to its roads turns a common idea into something unique and interesting. I think that's a good approach to take.
If as you say, many DMs risk falling into a trap by picking the "sensible" or "logical" skills for a situation, maybe it would be a good exercise to work in the other direction: instead of starting with the situation, start with a list of skills you want to use, and then build the situation and the challenge around those skills. If you take that approach, it's easy to have a skill list that matches your party - that's your starting point!
In practice I imagine you would combine the two approaches, modifying both the situation and the skills in a sort of feedback loop, as I did in my example.
I still think some sort of "no social skills" example would be good for the next article. And/or, start with a random list of skills and work backwards... kind of like an improv theater challenge.
Of course mixing up the skills used in different phases of the challenge would help, but still... in Mike's slaver city example, someone with a Stealth or Athletics boosting power could use it in most of the subchallenges.
Any thoughts about ways to handle these sorts of powers? Just let the player get the repeated auto-success - after all, she used up a slot for the power?
All that said, I'm a fan of letting people use their non-combat encounter powers. Taking that power, means they didn't take a combat power and they should be able to use it to succeed at skill challenges to help them avoid combat. A DM shouldn't try to prevent characters from using their non-combat powers any more than they would try to stop them from using their combat powers.
Of course, once the action was complete they would need to do a short rest before they could use it again. So you travel in a quarter, arrive at your destination and you might not get to take a short rest before an encounter begins at your destination. Alternatively if you are taking multiple actions in a row, you wouldn't necessarily have a chance to get the power back.
Good point. That factor should help the choice to use the skill boosting power over and over a bit less of a no-brainer.
All that said, I'm a fan of letting people use their non-combat encounter powers. Taking that power, means they didn't take a combat power and they should be able to use it to succeed at skill challenges to help them avoid combat. A DM shouldn't try to prevent characters from using their non-combat powers any more than they would try to stop them from using their combat powers.
I totally agree that you don't want to shut out use of these powers completely; I intended that suggestion more rhetorically than seriously. I'm wondering if there is some middle ground between "sorry, can't use it!" and "ok, you use it again and succeed again" over and over.
I'm not even worried so much that these powers would "break" the challenge, so much as they might tend to push players back into the passive, do the same thing over and over mode that Mearls has warned against. That's already going to be a potential issue if you have one skill that outshines the rest; how much more tempting will it be to use that skill over and over if you also get an effective +10 each time you use it?
Going back to my "lost in the desert" example for a moment, I hadn't really nailed down the exact structure of the challenge, but let's say during one phase, the party has to face two skill challenges per day: one to navigate and avoid getting lost, one to forage for food and water. Perception can be used in both challenges. Would it make sense to tell the player, "you can use Perception for both challenges, but you can only use your Perception boosting power for one or the other (even though it's an encounter power)?" He can still benefit from the power, but now at least he faces a potentially interesting choice.
For that matter, leaving the skill boosting power aside, would it make sense to tell your players, "you can use Perception for both challenges, but only for one or the other?"
I always figured the entire point of skill-boosting utility powers was for skill challenges - you don't get to make a skill roll for that "round" or "moment" in the challenge; you just grant a bonus to someone else.
Huh, I was just thinking about the powers that boost your own skills. One that boosts another player's skill raises other issues as you suggest.
Assuming we still want every player to have to make a roll, a la Mearls' "party check" idea, I think I'd treat using one of these powers on another player as an aid another attempt. So everyone has to make a skill check, plus everyone can do either one aid another attempt or use one skill boosting power on another player.
The key lies in building tension into each die roll. One of the best skill challenges I ran involved the party trying to sneak up on a tower that some bandits were camped inside of.
I think a lot of skill challenges have the problem that they stretch too little options for the players into too many die rolls.
Basically, after participating in a few skill challenges as a player and ruling some as a DM, I switched back to my old and proven way to play out non combat encounters, with the skill challenge rules only used to determine DCs and XP.
Basically the problem is, that the designer of the skill challenge is not able to prepare in advance on what the players will do or how they react to the situation - except if their options are severely restricted.
So, when I want a skill challenge in my game, I just imagine what I would do as a player and guess how many skill checks my solution would involve.
Then I know the complexity and the XP for the challenge.
In the actual game I just tell the players the situation, they tell me what they want to do, I let them use whatever skills they want (if it is remotely reasonable) and set the DCs usually at the DC for the skill I figured out before, but adjusting it on the fly if they come up with exceptionally good or bad ideas. (so basically it is a good idea to have a DC ready for every skill in the game for your skill challenge - except for the skills you think the players would never use. But from my experience, these are exactly the skills they will use... e.g. A PC trying to impress the king with an Acrobatics check...)
If they fail a check, something happens dependend on what they where trying.
e.g. failed intimidate means someone runs off and tries to warn the opponents or a failed Athletics/Acrobatics means loss of Healing Surge.
After they accumulate three failures some serious problems come up, or they fail entirely.
Since skill challenges give xp, I think skill challenges should have the same serious consequences of a failed combat encounter (besides TPK).
examples are: Party has to flee town, gets imprisoned, the good king gets killed.
Any thoughts about ways to handle these sorts of powers? Just let the player get the repeated auto-success - after all, she used up a slot for the power?
I don't mind the auto-success if the PCs get it once, but it is an issue to consider in longer term skill challenges. That's one of the advantages of creating situations and requiring party checks. A power insures some level of reliability, but you can't lean on it to overcome the entire challenge.
Basically the problem is, that the designer of the skill challenge is not able to prepare in advance on what the players will do or how they react to the situation - except if their options are severely restricted.
I think this captures the essence of the gap between running a skill challenge and designing a skill challenge. It also highlights one of the hardest parts about writing this article series.
In some ways, I think that a good approach to skill challenges might be to lay out some sample DCs (easy/average/hard), a few random events and complications (like the table in my column), and other notes about the situation (NPC personalities, maps of an area, and so on).
Then, let the players improvise skill checks. Maybe the challenge has a few ideas or a menu of default options if the players are stuck, but I think that skill challenges are most satisfying when the players are challenged to come up with good ideas.
In the challenge where the PCs had to sneak up on the bandits, at one point the fighter blew his Stealth check. In response, the wizard used ghost sound to create a distraction. I had him make a Bluff check, which he succeeded, and allowed him to cancel the fighter's failure.
I didn't write that into the challenge, but it was such a good idea that I had to go with it.
As an exercise I'm currently rewriting my Frozen Pass skill challenge and I came up with four actions that the PCs are likely to do, regardless of it being a skill challenge or not:
Travel on the Road, Travel off the Road, Set-up Camp (for taking an extended rest) and Explore.
Once I had these basic actions in mind, I was able to flesh out the skill challenge in much greater detail. I'll post it here when I'm finished for people to look at.
In the challenge where the PCs had to sneak up on the bandits, at one point the fighter blew his Stealth check. In response, the wizard used ghost sound to create a distraction. I had him make a Bluff check, which he succeeded, and allowed him to cancel the fighter's failure.
I guess giving a list of examples like this would help a lot.
The problem is that most of this knowledge is rather passive in my head, so it's difficult for me to come up with a list. But I really think this would be interesting for a lot of DMs - especially DM's which never had the opportunity to play at conventionsand learn from others.