Six Impossible Things Before Breakfast
This is not the article that most people will have expected to see in this space.
It’s supposed to be the fifteenth shelf of the Essential Reference Library, but that’s taking a lot longer to complete than expected – up to 3 hours per item just to gather and describe all the links, with 66 items to be done. You’ll see the work just oozing off the page when it finally appears!
So this is unashamedly a filler article. But it’s a goody.
How Hard Is It?
- “I leap up, grab the chandelier, swing from it to the balcony, roll, and dive out the window in one smooth motion.”
- “I use the pickax as a lock-pick.”
- “I stare into the fog, straining for any hint of the enemy – and when I see him, I silently strike!”
- “I execute an 11-G Immelmann with barrel roll to break the missile’s radar lock, then blast it out of the sky.”
- “I use pressure from the CO2 extinguisher to keep the antimatter in the air so that it doesn’t touch anything solid.”
- “I use the mirror to see around the corner and ricochet a shot off the far wall of the corridor to take out the bad guys.”
- “I break the impossible code. What do I have to roll?”
These are all cases in which the GM is perfectly entitled to requiring a skill check or attack roll of some kind.
And that means that these are all cases in which the GM is required to assess how difficult the task is.
‘So What,’ you may be thinking. ‘That’s easy enough to do.’
Every game system has some sort of skill resolution system in which a character’s capabilities are tested with a random mechanism to determine success or failure with the probability of success being determined by the “difficulty” (or some analogous assessment parameter) of the task.
And, all too often, game designers think it’s simply a matter of creating a list of difficulties and associated modifiers or adjustments – the details will vary from game system to game system, but the principles remain unchanged. In D&D 3.x, the DC table looks like this:
Very Easy | DC 0 |
Easy | DC 5 |
Average | DC 10 |
Tough | DC 15 |
Challenging | DC 20 |
Formidable | DC 25 |
Heroic | DC 30 |
Nearly Impossible | DC 40 |
Most game systems provide no more guidance than that. A few are better – Pathfinder, for example, has specific DC standards for each skill, providing a solid basis for an assessment of DCs.
‘So What? What more do you need? you may be thinking. ‘Just pick the category that seems right, read off the DC, and get on with the game.
If only it were that easy…
The Benefit of Expertise
The definition of any specific task in terms of difficulty category isn’t as easy to pin down as you might think. There are two inherent problems to be overcome.
At The Easy End
One definition and measure of expertise is that tasks that were impossible to even contemplate become achievable, even routine.
That’s why gaming bloggers like myself revisit old topics from time to time – even if our gaming skills haven’t improved, even if we haven’t thought of any brilliant new tips or mechanisms, our expertise at explaining things will hopefully have improved, so that we can make clear what we failed to communicate on our previous attempts.
A skilled artist can capture a recognizable likeness with just a few casual pen or pencil lines. A beginner may be unable to do so with hours of painstaking effort. Personality, Expression, and Mood are captured and manipulated automatically by the expert.
“Easy” is a relative term that means different things at different standards of expertise.
At The Almost-Impossible End
And, at the same time, “Almost Impossible” also changes in content with expertise, as the artist example shows.
With the boundaries that seemed so simple suddenly rendered vague, the task of assigning an appropriate DC to any specific task suddenly becomes far more difficult.
Expertise Cap Vs Catch-all Net
There’s another hidden issue that lurks in the tall grass to catch out unwary GMs. Is the highest category capped – are there some tasks that are simply not possible without a certain standard of expertise – or is it a catchall for any task that is not ruled impossible outright?
Low-Skill vs High-Skill
To be fair, the reason that there is inadequate documentation in most RPGs is because at low levels of expertise, the impact of these effects is relatively minimal, and the category labels can be taken more or less at face value.
Only at high skill levels do the distorting effects of relative competence become overwhelmingly significant.
Even at moderate skill levels – competent to earn a living at a particular task, no matter how poor the living standards might be – the distortion, while present, is relatively easy to ignore.
But any game that lasts long enough will eventually butt heads with the problem.
When?
This will vary from game system to game system, and from difficulty class to difficulty class. As a rule of thumb, the precise details start to matter when characters achieve a skill standard such that it requires only an above-average roll to succeed at a task of that difficulty class.
A d20 system has linear die-roll probabilities – so that’s succeed on a roll of 18 or 19 or more, or succeed on a 3 or 4 or less if the goal is to roll low. 3d6 has a dumbbell probability curve, buying the GM more time – succeed on a roll of 15 or less, say, or succeed on a roll of 6 or better if the goal is to roll high.
The Hero System is a roll-low system; Skill + Modifiers < Roll equals failure. ‘Modifiers’ are the equivalent of DC.
DnD is a roll-high system; Roll + Skill ≥ Difficulty equals success.
For example, lets look at DC 20 on the d20 system. When does this difficulty category begin to display distortion?
roll (18) + skill (x) ≥ difficulty 20, so x is 20-18=2.
Two ranks in a skill – that’s when it makes a difference whether or not a task is classified as DC20 vs DC25. But, because of stat bonuses, a single skill rank is probably enough.
Let’s look at it another way: what DCs are subject to distortion effects due to relative competence at a total skill of, say, 7?
roll (3) + skill (7) ≥ Difficulty, so DC 10 is affected.
roll (18) + skill (7) ≥ Difficulty, so DC 25 is massively affected.
You can employ similar testing with any game mechanic. The results are still the same – a small difference in the difficulty class to which a task is assigned can have a massive impact even at relatively low skill levels.
Three Models
There are three basic approaches to setting DCs that provide the necessary guidance. In the absence of guidance within the rules, it falls to the GM to make his own choice amongst the options available.
These three models are
- The Everyman Standard
- The Competent Standard
- The Dynamic Standard
The Everyman Standard
How difficult is the task for an untrained man off the street? That is the assessment approach embodied by the Everyman Standard.
It makes the assessment of difficulty levels the most automatic, enabling the descriptive labels to be pretty much taken at face value.
In terms of design philosophy, it indicates that expertise is measured not by the inherent capacity for success but by the frequency with which success will occur for a given level of expertise – which sounds fine, on the face of it.
But this approach has a hidden vulnerability or two, as well. It de-emphasizes the difference between having no expertise and having little expertise. That can be a good thing when you have a smaller group than usual, but it means that all characters become that little bit more alike and less distinctive.
Nevertheless, this is usually tolerable in a game system/genre that doesn’t place any special emphasis on skills, which is the case for most fantasy and adventure game systems, especially if character levels are never expected to rise very high.
The Competent Standard
The second model asks the question of any given task, “how difficult would this be for a typical character who is competent in the skill?”
This avoids the hidden problems of the Everyman Standard, but at the price of introducing a second subjective value judgment – to wit, what “typically competent” represents, in terms of skill level.
Again looking at DnD, I would say that any character with 4 levels in a character class is “typically competent”. At average INT, that’s +6 Skill Points, plus any skill points expended in character construction, plus stat bonus. Call it a skill of 7 or more, in total – which is why I chose that number when looking at when classification distortion has an impact on the meaning of a DC label, earlier.
A character with high stats – +4 in bonuses – can reach that level with only 3 ranks, well within the reach of many 1st-level characters if they invest virtually all their skill points in a single skill, which is a useful logic check, because that certainly sounds like a standard of competence for a typical professional NPC.
But another way to look at this as the definition basis of the task DC categories is to state that it adds 7 skill ranks of “distortion resistance”, relative to the Everyman model.
In practical terms, it means that more tasks will be allocated lower DCs, making it easier for low-skill characters to succeed in them.
This is the “gold standard” for most moderately skill-based genres/systems, such as modern adventuring. I would also apply it to relatively simple Sci-Fi systems like original Traveler.
The Dynamic Standard
The most complex, but richest, solution is to define ‘competence brackets’ and to assess each task relative to the competence bracket of the character attempting the task.
This fully embraces the distortion effects on difficulty of rising competence. The same task might be classified as “easy” for a character with Skill 20 and “Challenging” for a character with minimal skill. External conditions and circumstances can also be taken into account with greater facility and ease – trying to defuse a bomb in an environment filled with smoke, for example – because it breaks the problem down into smaller sub-problems.
That makes this the ideal solution for high-level games and highly skill-based genres.
It eliminates the distortion problem pretty much completely, but it does require extending the official rules of most games, who don’t include predefined competence standards.
The Profound Impact
The choice of which model you are going to employ in any given game has a subtle but profound impact on the game. In effect, they redefine what a character of a given skill level is able to achieve using his skill.
Let’s take a reasonably typical task, and compare the three models. Free-Climbing a 20m cliff during a thunderstorm, say, in order to reach shelter.
The slower you move under these circumstances, the more slippery the rocks become, increasing the difficulty.
- The Everyman Model: An untrained, unskilled character with no ability beyond inherent expertise levels is going to be slow, and will find this task extremely difficult. The urgency involved makes mistakes more likely, and 20m is a long climb when you aren’t a trained climber. If I were feeling generous, I would call this DC 18, but most of the time I would consider it to be DC 20. If the character had ropes, and pitons, and so on, I would drop the difficulty to DC15 on this standard.
- The Competent Model: For someone who knows how to free-climb, even if they aren’t an expert, this is going to be a lot easier. Still not easy, given the circumstances. Without climbing gear, a DC somewhere between 10 and 15 seems about right. With climbing gear, that might drop to DC 10.
- The Dynamic Model: The two examples given above already provide two data points for the Dynamic Model. In addition, let’s consider an expert climber. They will be less likely to make a mistake, by virtue of their expertise; they will be much faster, further reducing the difficulty; and they are far more likely to have climbing gear on hand. Putting all three factors together, I would rate this as an Easy problem for an expert climber (DC 5); If, for some reason, he didn’t have climbing gear (or didn’t want to stop long enough to get it out of his pack), a DC of about 10 seems right.
Of course, if the cliff was an especially difficult climb, I might have assigned higher DCs, but for a typical cliff, these are the numbers. So, let’s now see how these different models and the resulting differences in DC translate to chances of success for three characters – a beginner with Skill 5, a competent climber with skill 12, and an expert with a knack for climbing who has skill 18.
- The Everyman Model: Let’s be generous: DC 18, or DC15 with equipment.
- Beginner: Skill 5, so needs to roll 13 or better – 40% chance of success. With equipment, needs to roll 10 or better, so 55% chance of success.
- Competent Climber: Skill 12, so needs to roll 6 or better – 75% chance of success even without equipment. With equipment, needs to roll 3 or better, so 90% chance of success.
- Expert Climber: Skill 18, so needs to roll 0 or better – but a ‘1’ is always a failure. 95% chance of success, even without equipment. Using equipment doesn’t improve his chances.
- The Competent Model: DC13, or DC 10 with equipment.
- Beginner: Skill 5, so needs to roll 8 or better – 60% chance of success. With equipment, needs to roll 5 or better, so 80% chance of success.
- Competent Climber: Skill 12, so needs to roll 1 or better but 1 is always a failure. 95% chance of success without equipment.
- Expert Climber: Skill 18, so needs to roll -5 or better – but a ‘1’ is always a failure. 95% chance of success without equipment.
- The Dynamic Model:
- Beginner: Skill 5 and DC 18 (15 with equipment), so needs to roll 13 or better – 40% chance of success. With equipment, needs to roll 10 or better, so 55% chance of success.
- Competent Climber: Skill 12 and DC 13, so needs to roll 1 or better but 1 is always a failure. 95% chance of success without equipment.
- Expert Climber: Skill 18 and DC 5, so needs to roll -13 or better – but a ‘1’ is always a failure. 95% chance of success without equipment. But I wouldn’t usually bother getting an expert to roll such an obviously easy task.
Now lets make it a more difficult climb – a shortage of hand-holds, crumbling unstable rock, and 40m instead of 20. All told, those have to be worth +10 to the DCs, maybe more. But that will do to illustrate the effects.
- The Everyman Model: DC 28, or DC25 with equipment.
- Beginner: Skill 5, so needs to roll 23 or better – 0% chance of success, but a 20 always succeeds, so 5% chance, effectively. With equipment, needs to roll 20, so 5% chance of success.
- Competent Climber: Skill 12, so needs to roll 16 or better – 25% chance of success without equipment. With equipment, needs to roll 13 or better, so 40% chance of success.
- Expert Climber: Skill 18, so needs to roll 10 or better – a 55% chance of success. With equipment, needs only 7 or better, so 70% chance of success.
- The Competent Model: DC23, or DC 20 with equipment.
- Beginner: Skill 5, so needs to roll 18 or better – 15% chance of success. With equipment, needs to roll 15 or better, so 30% chance of success, but will still fail two times in three.
- Competent Climber: Skill 12, so needs to roll 11 or better – a perfect 50-50 chance. Adding equipment improves his chances to 70%.
- Expert Climber: Skill 18, so needs to roll 5 or better, so 80% chance of success without equipment. With equipment, needs 2 or better, so 95% chance of success.
- The Dynamic Model:
- Beginner: Skill 5, so needs to roll 23 or better – 0% chance of success, but a 20 always succeeds, so 5% chance, effectively. With equipment, needs to roll 20, so 5% chance of success.
- Competent Climber: Skill 12, so needs to roll 11 or better – a perfect 50-50 chance. Adding equipment improves his chances to 70%.
- Expert Climber: Skill 18 and DC 15, so needs to roll -3 or better – but a ‘1’ is always a failure. 95% chance of success even without equipment. Still in the category of maybe not even requiring a roll.
This shows that the Dynamic model increases the ‘spread’ of DCs for a task, emphasizing the difference not only of some skill vs little-or-no skill, but between experts and the moderately skilled.
Choosing Your Model
This isn’t the first time I’ve written on this subject – the last time, which also delves into other aspects of the questions raised, was “How Hard Can It Be?“. In that article, I talked about altering the standard scales of DCs, using the formula
New = 10 + 1.2 x (Old – 10)
…and advocated adopting what I’ve labelled the Dynamic Model for the purposes of this article. Further reflection has shown that this isn’t always the right answer, and hence this follow-up article.
GMs should choose the model that’s right for their game based on what they want to do with the game system and the campaign. If the intent is for the campaign to end by the time the PCs get to 5th level, the Dynamic System is probably overkill, and the Everyman approach is probably the easiest. If the campaign is going to be skill-based enough that skills will make a critical difference, but characters aren’t going to rise higher than 12th level or so, the Competent Standard is probably the best choice. But if skills are to be critical, or there’s even a possibility of the campaign lasting long enough for the characters to get to 15th level or better, the Dynamic Model is still your best choice.
Comments Off on Six Impossible Things Before Breakfast