Skills and Optimization

Sarah Darkmagic - Posted on 27 September 2011

Over the weekend I posted an article showing the skill distribution for the two Essentials players books. I added some quick analysis along with a statement that I intended to delve a bit deeper. Today I'd like to discuss why I think the current skills system leads people to be optimizers.

The Core Mechanic Encourages Optimization

In 4e, the books describe the core mechanic of the game, "the most important rule," as the following:

1. Roll a d20. You want to roll high!
2. Add all relevant modifiers.
3. Compare the total to a target number. If your result equals or exceeds the target number, the check is a success. Otherwise, it's a failure.

When I run games for new players, I'm told to explain this core mechanic. It explicitly sets up a situation where there is a binary success/failure and where players will feel the need to optimize. While there are always exceptions, such as DM-given bonuses for roleplaying, regardless of the table you play at, this is the assumed base experience.

So, if that informs the base of what you know about the game, it's reasonable to assume that a player likely will pick skills that align with the character's better abilities, the ones that give the largest bonus, especially when he or she has no idea how those skills might be used in the game itself. After all, what's important is that we roll high.

I'm a Star at Something, Right?

Even if that wasn't the core mechanic, I don't think it's particularly radical to suggest that people who sit down for the first time might pick skills based on what their character is innately good at. When I started out, I had no idea if history was going to be more useful to me than bluff (not that I necessarily had access to both of those skills), but I did know what my character was likely to be good at and just said she was also likely to be interested in learning more about those things. Also, it seems typical that players might want to differentiate their characters in a particular area. We're playing heroes after all. For instance:

  • "Thadeous has the gift of a silver tongue and the personality to milk it for all its worth." (Diplomacy)
  • "Serena's parents love to tell the story of how she walked across the laundry line to visit her friend at age 3. They knew she was a natural for the royal circus." (Acrobatics)

If everyone in a hero groups has basically the same set of skills and about the same abilities in them, what makes them unique? What would Leverage look like if all five of them were hitters? What would the Fellowship of the Ring look like if they were all wizards or rangers? Furthermore, if we are to have class skill lists, they should make it easier, not harder, to create a hero. Why auto-assign certain skills if the class isn't particularly good at the related ability, as in the case of the warpriest? But if we look at those distributions, we can see that by design it's harder to do that for some classes.

How Might Word-Based Ranks Help?

This natural inclination towards optimization, reinforced by the core mechanic, is one of the biggest reasons I like the idea of moving to word based ranks for skills. I personally would love to get away from worrying about +2 bonuses here and -1 penalties there. In my opinion, they often clutter our feat and magic item lists with these small opportunities to move the needle in one direction or the other for too small of an overall gain. Small increases in attack precision make a lot more sense given the small differences in defenses between levels. I don't think the same is true with skills. I'm not sure about the rest of the rules in that column, such as when you may roll, but I think that's a better discussion for later.

When Might Skill Clumping Be Counter-Productive?

Unlike combat, the roleplaying portion of 4e doesn't provide a way of moving the spotlight from player to player. Trying to enforce an initiative-order like system on this part of the game can make the story stilted and unnatural. In my experience, the lack of tools for moving the spotlight around makes it very easy for a player to fall out of the game during role-playing and skill challenges, especially shy or reserved players. If a character doesn't have a special skill to contribute, I think this makes it even more likely that unless the player has an internal reason for wanting to be part of it, such as enjoying play-acting or storytelling, the player will tune out. This isn't necessarily bad, not everyone enjoys this part of the game, but I think the current skills distribution can lead to this behavior in players who might otherwise enjoy role playing. As with just about anything, DM skill can overcome this.

In addition, just like certain damage keywords, some skills are more valuable than others. While this will vary from group to group, in general, heavy use of acrobatics, bluff, thievery and streetwise in skill challenges, traps, and other skill rolls may cause the spotlight to shine much more often on certain players versus others.

When Might Skill Clumping be Productive?

For some skills, we might want a number of characters competent in them. For tasks that come up often in the game, but not in a way that really spotlight's a particular character's talents, we might want a number of the people in the party to use the same skill. Perhaps they are trying to convince the hermit wizard that they aren't going to run off with his expensive magic item that will be useful in tracking down the evil goblin king who is attacking the town.

Some Ideas to Help

I've written a fair bit already, but I want to point out some tools that can help. I will write more in the future about this.


What some games do to move the spotlight around is to break the game down into scenes and having each scene have a subset of the characters present. For some groups, we could have multiple scenes taking place across the city at the same time. Perhaps one team is off trying to pass off a forged letter while another is gathering information in a bar. Even within one room we could have multiple scenes going. Perhaps a grate high up in the wall could provide the group with information of what's in the next room while another group could be investigating some runes on the floor.

The emphasis should be that the people in a particular scene should rely on their own skills and the skills used should be driven by role play. Yes, this can mean splitting the party during these parts of the game. If you keep each scene short and provide easy ways for the characters to escape any combat, this should be fine.

Non-binary Checks

Not all checks need to be binary in nature. The result doesn't have to indicate success or failure. Instead, it could represent degree of success. Let's say you're in a treasure room when someone, I'm not going to name names Mr McStealy, triggers a trap that causes the ceiling to begin descent. The DM can ask you to make a check to see how much treasure you escape with and the number you roll indicates the amount of success. Everyone will want to make this check regardless of their modifier. It's fun and we get to roll some dice. Not all non-binary checks need to be fun but they need to mean something to the players and/or their characters.

Ok, more later. But for now, I have to get some food. :)


I already use non-binary checks for most games. Effectively, I give them a lesser amount and quality of relevant data back from "close but no cigar" skill checks, and the reverse (more and better) for high scoring checks, unless the story demands otherwise.

It's mostly off-the-cuff, but it's generally like this:

An example knowledge check:

[Original Target DC: 15]

History (DC 13) Tells the hero that the Kingdom had a General named Bob and he's defeated some orcs in his time.

History (DC 15) Tells the hero about the Kingdom's greatest General, Lord Bob who defeated the orc horde at the battle of Three Pancakes which took place in the neighboring province.

History (DC 17) Tells the hero about the Kingdom's greatest General, Lord Bob who defeated the orc horde at the battle of Three Pancakes which took place in the Strawberry Valley (Geography will tell them how far it is from here) in the neighboring province.

History (DC 19) Tells the hero about the Kingdom's greatest General, Lord Bob who defeated the orc horde at the battle of Three Pancakes which took place in the Strawberry Valley (Geography will tell them how far it is from here) and what forces were used by both sides. His closest advisors are also named.

History (DC 21+) Tells the hero about the Kingdom's greatest General, Lord Bob who defeated the orc horde at the battle of Three Pancakes which took place in the Strawberry Valley (Geography will tell them how far it is from here) and what forces were used by both sides. His closest advisors are also named and detailed, along with casualty numbers (what would have been recorded) overall tactics and other trivia that could have been reasonably preserved and passed down.

Other checks work similarly, where a spot or search check "failed" by two may indicate a hint of something or generally arouse suspicion, an insight check may give the impression that the target is holding something back, etc.

One of the things I like about the skill rank system is that you can automatically assign a skill that the class would be 'expected' to be good at as an 'Expert' level (or whatever is appropriate). Thus, being of a low Int doesn't necessarily punish the Cleric, but at that Rank or Ranks that a dice roll is needed, a higher Int Cleric would be better at remembering certain aspects of Religion better than an average Int Cleric.

Maybe it is Monte and the L&L column but this topic seems to have been on my mind a lot lately. Their are tons of articles on the failings of skill systems.

In a nutshell skills are really a mechanical representation of a player saying my character is good at Skill X. I completely agree for new players this leads to optimization. As a DM my role in guiding new players is to ask how skill X fits with their over arching character concept etc.

In 4E a skill list shows what a Class is likely to be good at due to the class being a form of loose profession. This seems a backhanded way of simplifying the skill list for easy digestion. With the original PHB release the set of skills seemed to be intended to be balanced across rolls. Delving into this further there were flaws in the coverage leaving party wide gaps and duplications. The more player books that were released widened these gaps. Backgrounds seemed like a patch work solution to open more skill options to all the classes while trying to maintain a logical RP reason for a class to have access to them. In practice they seem to provide logical reasons for new players (and some old players) to do as you suggest and max out skills in areas they already have high stats.

Mechanics and bonuses from many other sources further compound this problem, creating an arm race of higher DCs to combat the greater expertise possessed by these high rollers. The gap between the capability of the worst and the best roller in a party can spiral out of control. The concept of a skill challenge then quickly degenerates to either a select view making specific checks or a useless succession of die rolls with no chance of consequence for any but the worst rollers. If they even make the check they are the worst at, even understanding the problem as a DM I don't want to be the player to cause the failure when I shift sides of the screen.

The word based ranks for skills seems like a good idea. I see the application used well in Hackmaster Basic, though that system has a great deal too many skills and I prefer 4E's Perception rather than lists of skills for the same application like Listen, Spot, Search, Touch, Taste, Scent etc.

Non-binary results quickly made my game after their application from Monster Lore and climb checks. This led to all knowledge checks then social encounters and finally into physical skills. Critical successes also crept their way into my skill checks and skill challenges, but oddly critical failures on skill checks haven't. I think that in the role-playing bits I struggle more to see my players fail with a good idea then I do during a tactical encounter. It is the only time I feel myself teetering on the desire to fudge a roll in their favor.

Non-binary results aren't really much different than the four tiered success of combat, are they? (Crit Fail, miss, hit, Critical hit). It is not a far leap from there to partial successes and partial failures involving entire skill challenges themselves.

Scenes have inched their way into LFR modules' skill challenges with limited success. I expect to see them used more with the new adventure format for Insider. In my own games they make simple blocks to guide Skill Challenge design. During game time scenes can be re-ordered or combined seamlessly based on player decisions.

All in all skills are an attempt to structure for design with mechanics something that is free form and at times chaotic, i.e. the role-playing part of a role-playing game. At best it takes the lets pretend moment and sets limits to the declaration that my wizard lifts the car off the fighter. At worst it stifles creative solutions and pigeons holes character conceptualization. Any re-vamp or fresh design of a skill system needs to be a flexible balance between the two extremes.

"Sarah thanks for letting my soap box up in your driveway. Wasn't me attention to blog here, really."

One way to make it not just about the optimizers is to have a challenge partly based on number of successes from different people. For example, you have three history items to be learned. You see who succeeds and assign them to those who made the roll and have the best background to learn that fact. The eladrin may be the best at history, but the human who comes from Dweezaria might know more about Sir Tightpants.

Also consider group skill use. Endurance, Athletics and Acrobatics easily lend themselves to this. If the party has to cross a desert, climb a cliff, or fight on the deck of a ship in a storm, these are important. DC's should be such that the lowest character has a chance. One way to let everybody have a chance and make the experts feel special is to let somebody take a penalty to help another. i.e. The dwarf paladin gives some of his water ration to the eladrin rogue.

Send feedback using the contact form or through twitter, @sarahdarkmagic.

Resources for FAQs



Syndicate content