Stalker0's New Skill Challenge System (Version 1.0)

It is my belief the current skill challenge system doesn't work properly, and so I have done an extensive rework of the system, using a large amount of calculations and mathematical models. The system is presented below, and in the next post I will write down a detailed account of each part of the system and explain how and why it came out as it does.

First, some ground rules:
1) If you don't believe the current system is broken, then please do not mention it in this thread, as this thread is for those who do believe it.
2) This thread is to discuss my system, not for you to talk about yours. If you would like to post your own system, I would be happy to take a look and comment in your own thread. But please keep comments focused on what is presented here.

And without further ado, my system. Assume that any thing I don't mention here is unchanged from the standard skill challenge system.

GOAL: A standard party taking on a challenge equal to their level, where each of the players gets to utilize one of their best skills, should have around an 80% chance of success in accomplishing the challenge.

The skill DC and Complexity tables are the heart of the new system, carefully crafted to provide balanced skill challenges for a party. Note that for most parties, there are only 3 levels of complexity. Complexity 4 and 5 are NOT designed with normal parties in mind.


*These complexities are only designed for parties that have 1 really strong skill user, not for general parties. The increased number of rolls prevents the one player from overshadowing everyone in the skill challenge while still providing a strong benefit.


Allowed Skills: When a DM sets up a skill challenge, he can choose a number of skills that the players may use for the challenge. These are known as “allowed skills”. He may also allow the players to choose their own skills, provided they can provide a good reason for their use.

Non-allowed Skills: A player may wish to use a skill that the DM has decided not to allow for the challenge. If the DM wishes, he can allow the player to use the skill, but only for the purpose of aiding another, and with a hard instead of easy DC (see below).

NOTE: This is a change from the DMG, where the DM can allow a non-allowed skill in with a hard DC. This mechanic replaces that.

Critical Success: Any skill check that rolls a natural 20 during a skill challenge automatically counts as a success. In addition, if that roll +modifier would make the DC, you gain a critical success. You can use your critical successes to use the skillful recovery mechanic (see below).

Aid Another (EASY): Instead of rolling a regular skill check, one character in the party per round can choose to aid another. When choosing to aid another, your skill check does not count as a success or failure for the challenge. Instead, a success provides a +2 bonus to the next person’s roll. Once you successfully use aid another, you can only aid again if you use a different skill. You can score critical successes on aid another rolls. Note that aid another rolls with non-allowed skills have a Hard difficulty.

NOTE: This replaces the aid another mechanic present in the DMG when it comes to skill challenges.

Skillful Recovery (HARD): You can choose to use skillful recovery as an immediate interrupt whenever you or one of your party members fails a skill check. Make a skill check, using one of the allowed skills for the challenge. If you fail, then the original failure stands. If you succeed, you negate the failure. That skill check neither counts as a failure or a success. You can use skillful recovery a number of times equal to the number of critical successes you have obtained through the challenge. You cannot gain a critical success on a skillful recovery roll.
Ex: In a negotiation with the duke, the fighter makes a comment that angers the duke. The charismatic rogue uses a hard diplomacy check and makes a comment about the duke’s dining hall. The duke, already impressed with the rogue from a previous comment he made (critical success) thanks him kindly, and forgets the fighter’s comment.

Daredevil Stunt (Hard): True skill masters can choose to make hard checks in order to gain more critical successes. Choose any allowed skill that normally has an easy or medium difficulty. You can now gain a critical success if you roll an 18, 19, or 20. However, only a natural 20 is still an automatic success. A daredevil stunt still counts as a success or failure for the skill challenge.
Ex. The party is climbing a steep cliff. The fighter, with his massive strength, training, and skill focus in athletics, has no trouble climbing the wall. He decides to use a daredevil stunt to climb the wall in half the time. He rolls an 18 and gets a critical success. He can now watch his fellow party members, and offer aid if they have trouble using skillful recovery.

General Questions: I’m going to try an answer a few questions I can see popping up with this system.

1) I’m running a party of 4, and one of the guys is a skill monkey!! He has everything in bluff, and he simply bluffs his way through everything. How can I handle him with your system?
First of all, is it really a problem? Afterall, if the guy has thrown a lot of resources into being a skill guy, then there’s no problem with him reaping the benefits. However, if you still feel it’s a problem, you may consider reducing skill items in your game. Magic items that add +4 or +6 to a skill can actually have a very large impact on the math of a skill challenge. Or, use the special complexity 4 or 5 I added in my table. This will reduce that one person’s impact on the skill challenge without greatly increasing the difficulty of the challenge overall.

2) I’ve looked at your skill DC table, and it’s a bit more complicated than the standard one. In general I don’t like tables that much. Are those exact DCs really that important?
The short answer is…often yes. The skill challenge system is intuitive and useful, but it has a problem with wide variance that can’t really be fixed without a drastic change to the math, which few people would like. My updated system curbs a good deal of it, but it still exists. Even a change of +1 can have a decent impact on your party’s chance to win a skill challenge. Whenever building a skill challenge, take half a minute and right down the proper DCs for your party. While as the DM you ultimately know your party far better than I do, these numbers will give you a solid win rate with an average party.

3) Hey I like this system and I would like to make it better. What can I do to help?
The biggest thing people can do to help is to provide me live data from their games. First and foremost, what kind of skill challenges is your party experiencing? What skills are being used? What are the skill modifiers players have as they increase in level? Right now my model is all theoretical, and it won’t be complete until I see it used with actual characters in actual games.

4) No offense, but how can you know the system is broken? After all, the game has barely been out yet. How can you be so sure the system doesn’t work?
While I obviously love to do theoretical models, I greatly believe in real experience being the final word. So why not this time? The reason is while math can’t tell you everything it can give you a good indicator, and in this case the math is blatantly saying the system is broken. Here’s a quick and powerful example. Take a team of 5 1st level characters through a complexity 5, level 1 skill challenge. Allow them each a +9 to their best skill (+4 mod + skill training) and allow them to use those best skills throughout the challenge. The party’s win rate? 7%.
Now if that win rate was around 50%, I’d still question it, but I would certainly run the system through its paces before making any final judgments. But it’s not even close, its 7%!! That’s just not acceptable, so a replacement system is needed.
If you would like to look at the math for yourself, go to this thread: http://www.enworld.org/showthread.p...779#post4278779

5) My group also thinks the skill challenge system is broken, but we don’t like a lot of houserules. Is it possible to use only some of your system, like the skill table for example?
If you would like to take only a part of the system, then take the Skill DC Table and the Complexity Table. With those two, your well on your way to a balanced skill challenge system. The other parts of the system help to tighten the numbers up, so without them you might notice a large difference between a complexity 1 and a complexity 3. However, you could run it with just those two parts.

6) I was looking at your skill table. Should I use this for all skill checks, or just skill challenges?
My table is specially designed to handle skill challenges. When dealing with regular skill checks, use the Tables provided in the DMG.

In this post I will go into much more detail about the system. This is mainly nuts and bolts stuff, you don't need to read this to use the system.

Let’s Begin: The Problems with the Skill Challenge system.
1) A party handling a skill challenge of their level has a very low chance of succeeding. This is of course the heart of the issue. If a party is handling a skill challenge of their level they should at the bare minimum a 50/50 chance of winning. But in general, players are supposed to win, so even 50/50 would be regarded by many as too low. As currently stands, those numbers stand at around a 10% win rate or lower, which is completely unacceptable.
2) Skill Challenges have a huge variation in win rate based on DC and complexity. This is a problem hidden in the math of the skill challenge system. For example, let’s say the skill challenge is perfectly balanced at complexity 3 at a set DC. The win rate is exactly where you like it, and everyone is happy. If you add even +1 to the DC you can throw off the win rate by 15-20%. Change the complexity and you can change that number even more. If a party is just slightly weaker in skills than another party, they can literally go from a decent chance of beating the challenge to a very poor chance with the most minor changes in skills.
3) Increasing complexity can actually make a challenge easier depending on the skills of your party. This one isn’t necessarily a “problem” as much as it is unintuitive. For example, if your party on average will succeed on each individual skill check of a challenge 70% or more of time, you will actually increase the party’s win rate by increasing the complexity. However, if the party has only a 65% chance, then their win rate will drop by increasingly complexity.

Tackling the Problem: Our basic assumptions
Before we can do any math to fix these problems, we need to know what the solution is. In other words, how often should a party beat a skill challenge? I took my own personal intuition, and asked many other people I game with. We each came to around the same conclusion:
We will assume that a party of 5 is facing a skill challenge of their level with complexity 5. All checks for the skill challenge will be medium difficulty (the standard skill challenge). Each player will have the ability to use their best or close to their best skills for this challenge. In other words, we are assuming skill training and a high ability score (probably +4 or more). So at 1st level, each player will roll at a +9 in general.
With this assumption, we felt that a party should succeed at that challenge 80% of the time. However, at the same time, we thought that each individual check should succeed around 70% of the time. We all know how it feels when your DM gives you a skill DC, and everyone looks at each other across the table thinking that DC is absolutely crazy. With a standard skill challenge, every check should have a reasonable chance of succeeding, so that players don’t get frustrated.

Pen and Paper: Our major limitation.
An important part of this process is to remember that we are playing a pen and paper game. Players are doing math in their heads and looking up charts in books. I could create an absolutely beautiful mathematical model that would run skill challenges perfectly across multiple skill levels, and you would never want to play it because it would simply be too darn complex! So while we are fixing the system, it is important to remember that the end product must be as easy to use as it is clean in its final results.

Now that you all know the goal, our assumptions and limitations, I will go through each section of my rule system and explain in detail how it solves the problems I have outlined above.

The Complexity Table:
This table has undergone a large revision. I have change the success and failure numbers, now have a footnote for two of the complexity numbers. So what is going on?

Curbing the variance: As I mentioned, one of the problems with the current system is a change in complexity has a huge change in the win rate. The reason is simple, the more rolls you have, the larger the variance. If a take a 70% chance of succeeding at any 1 roll and multiply it across 4 rolls, then I multiply it across 12 rolls, you’ll see a very large difference in the final number. The reality is that’s a basic property of the math we are working with, and I can’t change it without breaking our pen and paper limitation. But I can limit it.
The first thing is to reduce the variance in total rolls. In the original system, you could roll as many as 5 rolls for complexity 1 and 17 rolls for a complexity 5. That’s just too large a difference to create a consistent system. With my system, the rolls go from 5 to 11, a much tighter fit. We then change the ratio of successes to failures. This helps tighten the gap even further. Now when I set each individual success to 70%, the difference between complexity 1 and 3 is only 4%. That’s much better than 12-15% we experienced before.

Fixing the Inversion Problem: Fortunately, this is an easy fix. With the new success/failure ratios, this is no longer a problem. At any success rate you set, a higher complexity will equal a harder challenge….but hopefully not too much harder.

Questions you may have:
1) I understand the new table, but the highest complexity feels like too few rolls to me. Is there a way I can keep the length of the old complexity 5 but still get the good math of the new system?
The best way to do this is to do a big skill challenge in parts. For example, run a complexity 3 skill challenge for your party, and then a little later run a new complexity 3. Give them rewards or failures at the end of each one. This will give you added length, but allows the math to “reset” which will provide you a better challenge system over all. Remember that complexities 4 and 5 are designed for special circumstances, and are NOT recommended for most groups.
2) Complexity 3 has 7 successes and 5 failures. Couldn’t I just make a challenge with 14 successes and 10 failures and get the same results but have a longer challenge?
The answer is no. Probability doesn’t work that way. Even though the ratio of success/failure is the same, the actual results would be very different.
3) I don’t understand the footnotes in the table. What about complexity 4 and 5 is so special?
In general, I don’t recommend these complexities for a standard party. The reason is the average party’s win rate drops greatly with these systems. Their purpose is for a party that has players of general aptitude in skills but has one very very strong skill user. Some DMs may find this skill user takes too much of the spotlight in their skill challenges. By adding more rolls, the big skill user will contribute a smaller portion to the party’s success, but will still have a big impact. In addition, while the win rate drops with these complexities, the big skill power of the one character brings the number back up.

The Aid Another System:
One thing I decided had to be in my system was a way for a player with weaker skills to still make a solid contribution to a skill challenge. The reasons are many:
1) The variance in the skill challenge system is so high, that even 1 player with -2 to his best skills compared to the party average can actually drop the parties’ win rate by near 10%. This number gets higher in smaller parties, and if the bad skill player has initiative and goes more often.
2) No player wants to feel like they are dragging the team down. Part of the problem with a cooperative skill challenge is if a player doesn’t seem to be contributing as much as other players.
3) A self-correcting system. I can only make a system so good without breaking the pen and paper limit. I can only account for so many variables in characters, and my system will never be perfect for any party. Aid another allows the players to take charge of their weaknesses in a teamwork manner, which helps keep the skill challenge system consistent without a crazy ton of math.
So you may be wondering why I went with the system I did. After all, there are a 20 or more ways to do aid another (I know, I tried most of them). Here are the nuts and bolts of the system.

Infinite Retries, but only one success: One obvious flaw with allowing infinite aid another’s is that players with really high skill checks could simply aid another all the time, providing a permanent +2 to people’s rolls. However, I wanted infinite retries in there for an important reason…it scales with complexity. Any time I add a single +2 bonus to a roll a limited number of times that helps a system with fewer rolls. For example, if I added +2 to one roll out of 10 rolls, I helped 10% of the rolls. If I add +2 to 1 roll out of 20, that’s only 5%. Basically if I limit aid another, I give preference to complexity 1 over complexity 3, because complexity 1 has fewer rolls. This is a problem I struggled with in putting in an aid another system. However, with infinite retries the chances of a successful aid another roll increase with complexity, because there can be more attempts made (and more players have a chance to try it). So while it doesn’t scale completely, it doesn’t affect the variance too much.

No penalty for failure: To me this is a very important feature. If you are forced to roll a bad skill, your actually hurting the party by participating in the skill challenge. We never want that. With aid another, the worst that can happen is that you get frustrated for failing the check, but you won’t hurt the party’s chances.
Only one aid another per round: This one probably seems odd. There are a couple of reasons. One, I want to prevent the abuse where 1 player makes all of the “real” skill checks with their very high skill roll while the rest of the party sits back without a chance to fail. Two, because I’ve mentioned the skill challenge system is very sensitive to changes in skill checks. Allowing a party to acquire +4 to +8 bonus on most of its skill challenge rolls will greatly alter the win rate of the challenge.

Nonallowed Skills: In the DMG, the basic rule for nonallowed skills is that DM may allow you to make a skill check with a nonallowed skill but the difficulty becomes hard. This is a bad rule for two reasons:
1) If the character has ultra specialized in that skill, he may be able to make even hard DCs fairly easily. In that case, that skill just became the ultimate catch all skill, able to be used in any situation.
2) If the player has an average modifier to that skill, allowing him to make a hard Dc check is actually very harmful to your parties win rate for a skill challenge! The DM would actually be kinder to the party as a whole if the player got to skip without rolling.
In my system, I have rolled nonallowed skills into the aid another system. The DC is still hard, meaning a character is encouraged to use skills that relate to the challenge. But if he wants to press on, he won’t penalize the party. Further, if you do want to use your uber catch all skill, the most it will net is a +2 for someone else, and then you have to use a different skill to aid another. You get the benefit of your uber skill, but in a much more limiting manner.

Critical Successes, Daredevil Stunt, and Skillful Recovery:
A common house rule, a natural 20 is a big win, even in a skill roll. I have adapted it to my system for a couple of purposes:
1) Its fun! Let’s face it, everyone loves to roll a 20. And if a skill challenge is supposed to have the drama of a combat in many cases, why shouldn’t the natural 20 love fall along as well.
2) It helps curb the gap between different complexities. I’ve mentioned several times how the win rate between a complexity 1 and a complexity 3 can vary pretty wildly. Adding this rule actually helps that. The reason is the more rolls you have in a skill challenge the more likely 20’s will come up, which helps higher complexity skill challenges. While all complexities benefit from this rule, higher ones benefit more, which bridges the gap.

Daredevil stunt rose out of my skillful recovery system. I wanted my system to have a benefit for big skill users. Aid Another tends to help skill users with poor skills for a challenge, and the system is designed around the normal skill users. But what about the people with a racial bonus, skill focus, and magic items to help their skills…where’s the love for them? The answer is daredevil stunt. Hard checks in a skill challenge are just that…hard. They aren’t made for just anyone, they’re made for people who auto make medium difficulties and still have bonuses to burn. This system allows those people to get some extra benefit to their high bonuses in the form of better chances for critical successes. This further allows them to bail out their lesser skilled brothers in the party, while looking heroic at the same time.

Skillful Recovery was actually my favorite idea out of all of this. I tried all sorts of natural 20 rules and nothing worked the way I wanted. I tried adding a +2 bonus to the next skill roll, I tried +5. The actual effect on the math was either non-existent or too strong. I tried adding a +1 bonus to all skill checks for the encounter. That was too swingy, or just outright too strong. Plus those ideas are pretty boring in general, skill challenges should be as exciting as a combat if possible. Skillful Recovery is my solution. One it provides a significant bonus, but not as high as other less exciting ideas. It has more impact on higher skilled players, which I felt was important. Yet it still allows lower skilled people to get use out of critical successes. And who doesn’t love bailing out a friend by making a good check and saving the party from failing the challenge? The overall effect of SR did what I wanted, it narrowed the gap between complexities. Further, it provides benefit to higher skilled players while protecting the party from lower skilled players.

The Skill DC Table
First presented, yet the last thing I worked on. Once all of the subsystems were in place then came time to set the DCs across levels to ensure the win rate stayed consistent for all parties across all levels. What I did was create 3 template characters. One character was standard, I assumed a starting 18 in a stat with skill training. He put 1 of his two ability points in his main stat (for skill purposes) at each interval. Then I took a bad skill character. Starts with a 14, no skill training, doesn’t put any points in that ability score, except for the ones he gets at 11th and 21st. Lastly, the golden boy of skills. Starts with a 20, skill training, skill focus, and a racial +2 to the skill. Takes every bonus in that stat, and picks up a +1 to +6 magic item for that skill when he can.
I then calibrated the numbers around these templates. I focused on the standard character, making sure the medium DCs provided that 80% party win rate I wanted. I tailored the easy DCS to the bad character, but bumped them up a bit to ensure they weren’t too easy for a standard character. Then I created the hard DCs to be a challenge to the golden boy, then pulled them down a bit so they were in reach of the standard character.
The result? Take a standard party, let them use their good skills for a challenge of their level. They’ll succeed around 75-80% of the time, and each roll has a 65-70% chance of success. Parties with high skill guys can have fun with daredevil stunts and more critical successes, while teams with lower skill guys can be helpful with aid another.

Excellent system! Thanks for doing this! I'll probably use it.

In both the original system and in what you present here, there is a lack of tactical choice. There is no risk involved in choosing the different approaches. The skillful recovery and the daredevil stunt, for instance, should involve significant risks that make it a difficult (and hence interesting) *choice* to use them or not.

This is because true choices involve the players more than no choices or false choices. (A false choice is either a too obvious one or one that makes no significant difference.) True tactical choices in a skill challenge system of this kind give the same support to roleplaying (i e they structure scenes and help the players describe what their characters do), but they also challenge the players more. This is because the RP space is more confined when you have (or think you've found) a best tactical choice instead of having multiple choices that are are tactically equivalent (i e are false tactical choices). The question becomes "How can I RP the best tactical choice?" instead of just "How do I RP any of these skill uses?". Restrains force creativity and involve the players more.

I think you should add more risk to your approaches. It would make your system more fun and would involve the players more.

Very nice work, I think I'll use many of the ideas here.

What do you think of these changes?:

Critical Failure: Any skill check that results in a natural 1 is a Critical Failure. Not only does it count as an automatic failure in the challenge (if it had been for a skill check, not Aid Another), it cannot be countered by a Skillful Recovery, and it causes a -2 penalty on the next skill check.

Daredevil Stunts: You can now gain a critical success if you roll an 18, 19, or 20 if the skill check had been Easy, or 19 or 20 if it had been Medium.

I would like to see the math that has you arriving at a 7-10% success rate for standard equal level skill challenges in the core system.

Thasmodious said:

I would like to see the math that has you arriving at a 7-10% success rate for standard equal level skill challenges in the core system.

Click to expand...

Stalker will have to do that himself, but just looking at it, if 1st level characters have an average bonus of +9 on their best skills (Trained +5, Stat +4), and the Medium DC is 20, you'll need to roll an 11+ to succeed on an individual check. That's a 50% chance. (If any characters use skills they are not Trained in, or that have a lower Stat bonus, the chance goes down from there).

Now, with 50/50 rolls, you'd expect that if the number of successes or failures are equal, you should have a 50/50 shot at succeeding or failing the challenge. (If you flip a coin 6 times, overall you should average out to 3 heads and 3 tails, if you do it enough times.) However, skill challenges only need half as many failures as successes. So, back to the coin tosses, it is easy to see that the chance of getting 2 tails before you get 4 heads is significantly higher. And the more times you toss the coin, the less random chance and variance, and more likely to get the average result, a 50/50 split, which = failure because you only need half as many of them.

You can use [*pre] or [*code] I believe to format your tables with spaced text to make them easier to read, Stalker.

hcm said:

In both the original system and in what you present here, there is a lack of tactical choice. There is no risk involved in choosing the different approaches. The skillful recovery and the daredevil stunt, for instance, should involve significant risks that make it a difficult (and hence interesting) *choice* to use them or not.

Click to expand...

The skillful recovery mechanic is not a tactical choice as you say, nor is it designed to be. It is a reward for people who get critical successes.

However, daredevil stunt DOES have a cost. It bumps the difficulty of the skill your using to a hard difficulty, that could be a net +4 to +7 in difficulty. And it counts as a success or a failure as normal, so if a player attempts a DD stunt and fails, that's a failure for the skill challenge.

Samurai said:

Critical Failure: Any skill check that results in a natural 1 is a Critical Failure. Not only does it count as an automatic failure in the challenge (if it had been for a skill check, not Aid Another), it cannot be countered by a Skillful Recovery, and it causes a -2 penalty on the next skill check.

Daredevil Stunts: You can now gain a critical success if you roll an 18, 19, or 20 if the skill check had been Easy, or 19 or 20 if it had been Medium.

Click to expand...

I wouldn't recommend critical failures in the game, as the math doesn't account for them. For DD stunt, in general the people who are going to use them have a very high modifier in teh skill, so to them an easy or medium skill is probably an autosuccess or close to it. However, if you do want to use this rule, I would make easy 17-20 and medium 18-20, as I mainly based the math around medium challenges.

I would like to see the math that has you arriving at a 7-10% success rate for standard equal level skill challenges in the core system.

Click to expand...

Of course. http://www.enworld.org/showthread.php?p=4278779#post4278779
I have detailed the math in that thread for you. Any further questions about that, please submit them to that thread.

You can use [*pre] or [*code] I believe to format your tables with spaced text to make them easier to read, Stalker.

Click to expand...

If someone could tell me how to do that in detail I would really appreciate it. I have no idea how to use code for enworld.

Great system, Stalker0. I much prefer it over WotC's basic system. One question, however.

One of your assumptions is that you'd like to see PCs succeeding around 80% of the time with a skill challenge. I'd like to set the baseline slightly lower, closer to 65%. Could that be achieved in a rough way by just adding +1 to all the DCs you provide for the Skill DC Table?