Fans will recognise the quote used as the title of this article as something often said by Jeremy Clarkson on Top Gear when the trio of presenters are about to attempt a challenge posed by the show’s producers. Since it inevitably turns out that the correct answer is “Very!”, it is usually followed by one or both of the other commentators yelling “Don’t Say That!”

The challenges that the team undertake are always entertaining and often fascinating. But the question itself got me thinking (again) about a subject that I’ve ruminated on a number of times in the past.

Every skills system has some objective measure for how skilled a character is. The usual basis is a roll plus a modifier which has to be higher than a total determined by the GM, though there are variations. Sometimes the objective is to get less than the total, for example, and modifiers don’t increase the likelyhood of success, they reflect an increase in the difficulty due to circumstances. But either way, the goal is to achieve a target set by the GM or better (with “better” being defined differently from system to system).

I want to talk for a bit about how those targets are set. I’ll be using D&D 3.x as my foundation for the discussion, but the concepts are applicable to all skill systems.

Relative Vs Absolute Difficulties

One of the big stumbling points that many game designers seem to trip over is never telling the GMs, who are actually going to use the skill systems that they create, whether skill difficulties are assessed on a relative or fixed basis.

You see, the GM assigns the target in answer to the question, “How hard should it be?”, but this form of the question is actually incomplete.

Is the right question, “How hard should it be for an ordinary person with no training?” – since the abilities of our hypothetical ordinary person won’t change, this is an Absolute-basis difficulty.

Or, is the right question, “How hard should it be for an ordinary person with typical training?” – since the abilities of our hypotherical ordinary person still don’t change, but the concept of “typical training” is both circumstantial and subjective, this is a case of relative difficulty where the target number will change for characters of different backgrounds and from one GM to another.

Or, is it “How hard should it be for an expert?”, which is another relative value.

Or even, “How hard should it be for the character who is making the attempt?” – still another relative value.

Without knowing the basis on which difficulty levels are determined, it’s not possible to properly set a given difficulty level.

It’s even possible to have compound systems, where skill levels are categorised and problems are assessed as to what the minimum skill category is for attempting to solve the problem – it doesn’t matter how skilled you are at basic arithmetic, it doesn’t help you solve a quadratic equation or determine a second differential. It’s one thing to be a skilled journeyman carpenter, able to reliably nail two pieces of wood together to make a basic bookshelf or re-shingle a roof; it doesn’t qualify you to create a china hutch or a house. You lack the basic training in the higher level of problem, don’t have the fundamental education that gives you the right questions to ask (never mind enabling you to find the right answers). A task that might be routine for a medical specialist, like a heart bypass, can be extraordinarily difficult for a skilled surgeon with a different speciality. I don’t care how good a plastic surgeon might be, or a veterinary surgeon, or a neurosurgeon – I would have to be pretty desperate to let them try to perform a kidney transplant or triple bypass on me!

Difficulty By Example

This is the DMG’s approach in D&D 3.x, and it doesn’t work, at least not as well as it should. It describes a DC 40+ check as something that’s almost impossible, an almost miraculous result if you succeed. But a little digging into the system shows that it’s not all that hard, after all.

The designers don’t seem to have communicated the answer to the fundamental question of what basis you use for assessing how hard something should be, even amongst themselves. Thus, we have the situation where some of the examples offered make perfect sense for a first-level character and others don’t, and as characters go up in levels the examples make progressively less and less sense.


Part of the problem is that the fuzzyness of the die roll easily outweighs almost every other factor, at least until you get into higher-level characters. We’re talking about a d20 here, which means that there is just as much chance of rolling a “one” as there is of rolling a “twenty”, but those two results make a huge difference to whether or not the character succeeds or fails.

How hard is it to perform a typical task?

The DMG defines these as DC 10. Let’s say that we’re talking about a 1st-level character with 1 skill level. He might have as much as +4 from stat bonus. He might also be able to claim a +2 synergy bonus.

1 (rolled) + 1 (skill level) + 4 (stat bonus) + 2 (synergy bonus) = 8. Even on the worst possible roll, the character almost succeeds.

But wait – a roll of “1” is usually an automatic failure. And a “20” is usually an automatic success. And characters can have more than one skill point in a skill – the maximum is either character level +3 (for “class skills”) or half of this (for “cross-class” skills).

2 (rolled) + 4 (skill level) + 4 (stat bonus) +2 (synergy bonus) = 12. Success!

How likely is this result?

Critics will immediatly point out that I am employing the worse possible case, and that this is very unlikely to occur. So let’s pause for a moment and think of how unlikely it is.

Characters will rarely attempt something they don’t know how to do, especially at low levels, if they have a choice – they would hand the task over to someone else who does know how to do it. That means that most of the time skill use will be by a character for whom this is a class skill.

Character abilities in D&D are usually themed around the class concept. So, if you are good at something, there is a reasonable likelyhood that you will also be good at related tasks; and that is a situation in which a synergy bonus is likely to result, unless the GM has a specific reason for disallowing it. However, the character does need 5 skill ranks before they achieve a synergy bonus, so that means that at first and second levels, they can’t normally do so.

Characters who are good at something are also likely to have stats that are higher in those areas that assist them in those endeavours. That means that the character is more likely to have a high stat bonus for any check that they make than not, if the character is supposed to be good at this sort of task – and if they aren’t, as already stated, they are more likely to hand the task over to someone else who is. A bonus of +4 is still extreme – but the example character got a total that was two more than needed, a stat bonus of +2 would have been enough.

Another possibility that was not taken into account is “aid another” which simply requires another character to succeed in a roll by 10 or more. But there’s some vagueness about this rule as well, which I’ll discuss seperately below.

That all adds up to the “worst case” being the most likely case – once characters are third level or better!

Do skill synergy bonuses stack?

For my money, it’s a named bonus, so I would rule “no” according to the strictest interpretation of the rules.

My players would like to suggest “yes”, arguing that the bonus reflects ability with tasks related to the actual objective, and that because each skill is different, each related skill is reflecting ability at a different aspect of the objective task, and therefore synergy bonuses should stack, and sometimes I weaken and permit this line of arguement to sway me.

I also sometimes adopt a middle ground, and state that while the first skill synergy gives +2, each one thereafter only adds 1 to the score, because the source skills also have some relatedness. This, in effect, uses the players own arguements against them.

And yet, page 21 of the DMG does not list “Synergy Bonus” amongst its named bonus types. Neither does the Rules Compendium. And so the debate continues.

Aid Another

Every skill system needs some way for multiple characters to cooperate on a problem. In D&D, that subsystem is covered by “Aid Another”. According to the PHB, if you roll a 10 or higher on your check, you give the character actually trying to perform the task a +2 to succeed.

It doesn’t matter how difficult the task is. It doesn’t matter whether or not the character trying to help knows what they are doing (or does it? more on that in a moment). It doesn’t matter how naturally gifted they are or aren’t, ie what their stat modifier is (same comment). It doesn’t matter whether or not multiple characters are actually required to achieve the result. But the big problem is that the rules don’t say what kind of check is required. According to what the PHB says, it could be a straight die roll – if you get 10 or better, you succeed in conferring a +2 to the character making the attempt. And the DMG doesn’t mention Aid Another at all.

But an equally acceptable interpretation – and one that is supported by the use of the word “Check” rather than “Roll” in the PHB – is that the character has to make a check against the same skill as that being used to achieve the primary task, or a skill that is somehow related to that task – and that the latter, if successful, confers a synergy bonus that DOES NOT STACK with any synergy bonus the character is already getting for that skill.

In fact, consulting the Rules Compendium reveals that this is in fact the correct basic interpretation, though it doesn’t mention the Synergy Bonus refinement, that’s all me!

The other rule (that is somewhat reminiscent of the Synergy Bonus rules) states that characters with 5 or more ranks in the skill in question can add more than the standard +2, they add +1 for every 10 points by which their skill total exceeds the initial 10.

Does Aid Another stack?

Here’s another grey area, and this time the Rules Compendium isn’t much help. And, as usually seems to be the case in such matters, there are multiple possible answers.

A “Yes” can be justified simply by pointing out that there is no reason why more than one person can’t assist someone else.

A “No” can be justified simply by pointing out that there is a practical limit to how many people can work on a single task, and rather than forcing difficult and dubious decisions on the GM, it’s better to simply cut your losses.

A middle-ground approach can be implemented by ignoring the judgements suggested by a ‘no’ case and specifying that the number of characters aiding another has to double each time an additional bonus is awarded. So you can get +2 from one assistant, and two would give +4, then four for +6, and eight to get a +8, and 16 to get +10, and 32 to get +12, and 64 to get +14, and so on.

A sightly more conservative approach is to say that each additional success only confers +1 and not +2. That would mean that instead of 32 people giving +12, they would only give +7; to reach that +12, you would need 2048 assistants.

More conservative still is to state that after a certain limit, the base benefit of aid another is +0 – and therefore only results that come from greater expertise can further improve the total.

My personal inclination (and I have not discussed this with my players) is to set a limit even to the +1’s of +4. So, what that gives is:

  • 1 assistant = +2
  • 2 assistants = +3
  • 4 assistants = +4
  • 8 assistants = +4, +1 if 4 of the assistants get a total of 20+ (results of 30+ count for two);
  • 16 assistants = +4, +1 if 4 of the assistants get a total of 20+ (results of 30+ count for two), +2 if 8 get a total of 20+ (results of 30+ count for two);

and so on.

A total of 20 is not all that hard to achieve once a character hits middle levels, but the weight of numbers means that it will add up.

I would also rule that the DC required to aid another goes up from the base of 10 with each increase in bonus.

  • So to get the +3 with 2 assistants, the first assistant would need 10+, and the second would need 11+.
  • To get the +4 with 4 assistants, the first needs 10+, the second needs 11+, and the third and fourth need 12+.
  • To get +5 with 8 assistants, the first needs 10+, the second needs 11+, the third and fourth need 12+, and the last four must get 23+.
  • To get +6 with 16 assistants, the first needs 10+, the second needs 11+, the third and fourth need 12+, four need 23+, and the last eight need 24+.

This combination means that there is an upper limit to how many assistants will be helpful; eventually, the required total will reach or even exceed the initial DC 40; and sooner or later, there won’t be enough assistants who succeed to push the bonus any higher. In fact, to have any confidence in achieving even the +6, you might need 20 or more assistants, just to allow for the fact that some of them will fail. And the higher the target bonus, the higher the risk of failure, and the more non-contributing overhead is needed in assistants.

None of this is official, of course.

To Dream The Impossible Dream

So, if a character is good at something, they will succeed in a “typically difficult” task 95% of the time. How hard is it going to be for a character to get a DC 40 total?

Let’s start by rephrasing the question: “how high a level does a character have to have before they have a chance of success at a DC 40 task for which they are suited?”

I’ll ignore the “twenty always succeeds” and say that the character has rolled a 19.

That leaves 21 more to come from other sources. If there are no other bonuses, that has to come from the character’s skill ranks and stat bonuses. So let’s start by figuring it with no stat bonuses either: Skill = 21 = level + 3. So with no other assistance, level 18 characters will succeed at a DC 40 check on a roll of 19. So far, that seems reasonable.

For every +1 that comes from any other source, the level will drop by 1. So let’s start by adding in a stat bonus. +2 is a fair enough bonus for a 1st level character, but it could be +3 or even possibly +4. I’ll use the middle value – a plus 3. So that reduces the character level from 18 to 15.

By the time a character reaches level 15, they will have had three stat increases. Let’s assume that two of those were on the stat whose bonus we’re considering here, because we’ve specified that this is a character who is going to be skilled at this task. That’s worth an extra +1 stat bonus, so now we’re down to level 14 (and we lose the third stat bonus anyway).

Next, let’s throw in a +2 synergy bonus. It could be higher, as discussed earlier, but let’s be strict about this. That drops the level required down to 12.

Now, let’s have “aid another” from someone else. That’s another +2, so now we’re down to level 10.

By the time they are level 10, the character will have received at least 4 feats. Let’s assume that 1 of those feats gives a +2 at doing this particular task. That drops us down to level 8.

By the time a character reaches level 8, he will have a number of magic items, some of which confer bonuses to performing certain tasks. So let’s give another +1 from that source – even though it could be much higher.

That brings us back to level 7, but it goes back up to level 8 because we’ve lost the extra +1 stat bonus.

Some spells confer additional bonuses at performing certain tasks. Let’s also add in +1 from that source. That gets us back down to level 7.

Favourable conditions can give +2 or more bonus. So let’s add that in, as well. That brings us down to level 5.

A fifth-level character can succeed at a DC 40 task. Admittedly, there is a low chance of success – a 19 or a 20 – but that’s still one time in ten that the character can do the almost-impossible. And that’s being reasonably conservative with the modifiers; while some of them might not be available, others can be much higher.

A 50/50 chance of success would mean that the character needs to roll 11 or better. Since 19 minus 11 is 8, we can determine the level at which a character has a 50/50 chance of success by adding 8 to the level. Let’s go one better, to give the character a better than 50-50 chance.

A fourteenth level character can succeed at a DC 40 task more than 50% of the time. Under reasonable circumstances, he is more likely to succeed than to fail.

That’s a sobering thought.

But if we allow Aid Another from more than one assistant, and a better stat bonus, a first-level character can easily achieve a DC of 40. Depending on the house rules one adopts to restrict Aid Another, a first-level character can even have a better than 50-50 chance of achieving that success!

So much for DC 40 being almost impossible, a near miracle.

Even just using the more conservative approach to aid another that I outlined above, an eighth-level character can easily have a greater chance of succeding than he has of failing a DC-40 test.

At least one character in a campaign that I run is in his early teens in terms of character levels and has a bonus to selected skills of well over +40, achieved absolutely honestly, using all of the above techniques – except aid another! – AND using the most extreme interpretation of the synergy bonus rules. I mention this purely because it shows that the preceeding analysis is not some theoretical or hypothetical discussion – it really happens.

So the answer to the question, “Just how hard can it be to achieve DC 40” is “not very.”

By the time the character reaches his upper teens in level, it will take a DC of 55 before the character has even a 50-50 chance of failure in those skills!

Which brings us back to the question posed at the start of this article: “How hard is it” For Whom?.

Difficulty Options, Redux

With everything that’s been learned in all this discussion, have we made any actual progress in resolving the issue?

The answer is both yes, and no.

An answer to the central question remains elusive, but at least we’re in a better position to understand the implications of the choices; we’ve gained perspective. Those implications may well be the final element needed to make an intelligent choice amongst the options.

Absolute Difficulty

It’s now clear that this is the standard by which the DCs in the DMG are measured, and that it is a flawed standard. Choosing task DCs according to this standard means that there is virtually no chance of even low-level characters failing at a task that is within the province of their expertise unless the task is superhumanly difficult. There are so many flaws in this arrangement that have come to light in the course of this analysis that the Absolute Difficulty model must be considered hopelessly broken. It’s too easy for even low level characters with appropriate support to work miracles.

Yet, it’s not all that difficult a fix. Increasing the range of all DCs greater than ten proportionately is all that would be required; so that a formerly DC 40 task is now a DC 55 or 60. The latter would be easier, mathematically. The adjustment would be

New = 10 + 1.2 x (Old – 10).

But I don’t recommend this solution. It’s a kludge, and it feels like a kludge. We need a better solution than Bailing Wire and Duct Tape.

Relative Difficulty: The Background Standard

This has a lot going for it as a concept; the notion that some races inherantly find certain tasks easier than others, that certain ethnic backgrounds might inherantly be better or less capable in certain fields of endeavour, whether that be by virtue of personality or general disinterest or whatever.

But the fact is that all of these things are more easily addressed within the current system in other ways than changing the DC. Giving races a specific bonus or penalty to certain tasks or with certain skills; or giving an additional allocation of skill points to be expended in certain key areas; or simply mandating that a certain number of skill points have to be spent in a more restricted way; or any one of several other alternatives, would accomplish this same goal, while creating less work for the GM.

This solution, too, is a bust.

Relative Difficulty: The Expert Standard

With only two options left, we would seem to be closing in on either a solution or a massive anticlimax! I have an almost-instinctive suspicion at this point as to what the solution is most likely to be, but I will continue to examine the alternatives dispassionately; I might be fooling myself, or miss a bet, otherwise!

So let’s think about the “Expert Standard” for a minute. This means that difficulties are assigned based on how hard they should be for someone who knows what they are doing – and that a certain level of expertise may be required before characters can even attempt certain tasks, no matter what bonuses they have from sources other than skill ranks.

Right away, this solves part of the major issue that was encountered with the Absolute Difficulty. It is no longer too easy to achieve those miraculous feats of skill. So that, at least, is encouraging.

However, this solution comes with a problem: who decides how “expert” an “expert” is? How do you keep the standard consistant? There is no easy solution. And, what’s more, there remains the same problem that was encountered when considering “The Background Standard” – the question of having multiple fixed standards to create and keep in mind. If this flaw was enough to skupper that proposal, it is also enough to cast shadows of doubt over this one; there may be fewer DC standards to bear in mind, but there are still several of them – possibly too many.

Relative Difficulty: The Character Standard

This proposal is radically different from all the others in that it abrogates all fixed standards. The concept is that each time a task is to be achieved, the GM determines hwo difficult it is for this specific character, at this specific time, and under these specific circumstances. The same task might have a very different DC tomorrow.

This pretty much eliminates the modifier for Favourable or unfavourable conditions, building that straight into the DC determination. It also permits the inclusion of the background standard as just another element for the GM to consider. So this concept includes several of the nice-to-have ideas that we’ve come across in earlier discussions.

What’s more, it solves the primary problem of high DCs being too easy to achieve. So far, then, it has incorporated every useful feature of every alternative that has been considered.

Nor is there any real downside. On the contrary: it explicitly rewards characters with high skill levels by permitting them to attempt tasks that a less-skilled character is not permitted to succeed at, no matter HOW good their skill totals are, AND it simplifies the GMs workload.

It might not be canon, but it seems to be a far better solution than any of the alternatives.


Ultimately, how you choose to answer the questions posed by this article is up to each individual. There are clear arguements in favour of a “Relative Difficulty/Character Standard” approach, but the fact that it is NOT canon, and is even contradicted by the examples given by the rules, may make this an unacceptable choice for some GMs.

What’s more, since the players are ultimately going to have to live with the decisions, and with the clearly significant impact on what their PCs can do, it may well be the case that some compromise is necessary, no matter how much the GM might favour one answer over another.

There are some serious issues raised about game balance and playability, and characters’ abilities to perform seemingly miraculous feats of skill. It might well be that different campaigns will require different answers, especially campaigns in which the GM expects the players to achieve epic levels. At very low character levels, all three systems give essentially the same answers; it’s only from lower mid-levels (around level 5 upwards) that they begin to diverge.

The same questions also arise with all other game systems as well, because ultimately all tests of skill come down to the basic principle of ability vs required skill defining a probability of success, which is then checked with a die or dice roll. This is true of the Hero System, Call of Cthulhu, Hackmaster, Rolemaster, Empire Of The Petal Throne, Space Opera, Traveller, TORG, Star Trek – in fact, of every RPG that I’ve ever played, refereed, or read. All of them require the referee to decide “How Hard Can It Be” – and all of them offer examples – and none of them do a good enough job of telling the GM how to find an answer to the central question of this article. Which leaves it up to the GM.

My answer: decide on your solution, discuss it with your players so that they know what to expect – and then stand by it until it becomes clear that the solution isn’t working. In other words, set your own policy and live by it. If this article has given you the information you need to make an informed decision, then it has done its’ job.

Related Posts with Thumbnails
Print Friendly