Skill Check DCs

Among other things, skill-related talents replace skill ranks and the like.  I was starting to copy the RSRD skill descriptions to this site when I realized the change in calculating skill check modifiers would greatly invalidate the DCs presented in the RSRD.  I need to think about this.

Thankfully, it looks like it should be pretty easy.

Skill Check DCs in Existing Systems

Skill check DCs are directly proportional to character level, though the exact relationship is different between D&D 3.x and D&D 4e.

D&D 3.x Skill Check DCs

Skill check DCs in D&D 3.x are pretty gonzo, I think, because of the potentially huge separation in check modifiers.  It looks like most skill checks are expected to be attempted only by those who are trained and have maximum ranks in the skill, and a good ability score.  At any given level ‘standard check DCs’ are probably in the range of 11+(level+3)+(5)… [d20 + ranks + ability score modifier for a ‘decent’ ability score] though they aren’t really that consistent.

Basically, if you’re trained, you probably have around a 50% chance of success, and if you’re untrained you probably can’t do it unless it’s really easy.  Probably, because it’s not necessarily consistent.  Some things can be expected to be successful on DC 10 or DC 15, others don’t start until DC 20… and it’s not always clear why.

D&D 4e Skill Check DCs

In many ways D&D 4e skill check DCs are much simpler overall.  I don’t have exact DC numbers, but I understand that there is a “70% success target”.  That is, easy checks at any particular level are expected to be 70% successful when attempted by ‘off-stat’ and untrained characters, medium checks at any particular level are expected to be 70% successful when attempted by an ‘average’ character, and hard checks at any particular level are expected to be 70% successful when attempted by an ‘on-stat’ and trained characters.

I think this is, overall, a pretty reasonable and easy to work with.  I’m not entirely certain about the planned 70% success rate, but that’s a fairly minor detail.

I will address an interpretation of this design guideline, though.  Given the desired “70% success target”, during design skill check DCs may be calculated to generate ‘easy’, ‘medium’ and ‘hard’ values for each level.  This is handy, but has led people to complain “so the wall gets harder because I’m higher level?”.  This is an incorrect interpretation because the DC is determined at design time based on the desired relative difficulty for the expected character level and does not change regardless of the character actually facing the challenge.

Now, just what the twentieth-level warlord is doing with a DC 30 lock on his bathroom, I can’t explain.

Skill Check DCs in Echelon

The skill model in Echelon is closer to D&D 4e than D&D 3.x, so let’s follow a similar model.

When making a skill check, characters roll d20 and add their ability score, Level Bonus, and Training Bonus (if any).

  • Average ability score is 5, highest ability score is 9, lowest ability score is 1.  I have no plans at this point for characters to be able to change their ability scores.
  • Skill training gives a +4 competence bonus.
  • Level Bonus is one-half the character level, rounded down — each tier provides a +2 Level Bonus over the same level in the next-lower tier, and thus +4 difference over two tiers.

Note that additional training — higher tier slots spent on a skill — gives the character more options and potential results that are impossible to lesser-trained characters.  That is, that a low-level character and a high-level character both beat DC 24 does not necessarily mean they end up with the same effective result.

Assuming a roll of 11 on a d20 means 50% success rate, at any given level

  • Easy: should be achievable 50% of the time by Ability Score 1, untrained.
  • Routine: should be achievable 50% of the time by Ability 5 untrained or Ability score 1 and trained.
  • Hard: should be achievable 50% of the time by Ability score 5 trained or Ability score 9.
  • Very Hard: should be achievable 50% of the time by Ability score 9 and trained.
Level Easy Routine Hard Very Hard
1 12 16 20 24
5 14 18 22 26
9 16 20 24 28
13 18 22 26 30
17 20 24 28 32
21 22 26 30 34
25 24 28 32 36

I could do much as D&D 4e did and design around ‘difficulty-for-level’, but I’d like to simplify things a bit.  Instead, I will use an absolute scale of difficulty and expect that characters will generally face challenges suitable to their level.  That is, there is no longer a “level 20 easy lock” on the bathroom door, it’s just an “easy lock”… but what might have been described in D&D 4e as a “level 20 hard lock” would now be described as just a “very hard lock”.

This is purely a difference in interpretation, and I may even have a table mapping “champion hard” to one of the more absolute ratings, but it will hopefully make it more clear that yes, high-level characters do harder stuff, and make things that should not be proportional to level be not proportional to level.

Also, while I could reasonably make the target numbers differ by two per step, I think stepping by four will be simpler and work as well.  This leads to the following target numbers (assuming about 50% success), and indicating the tier (and level) that you should expect to see that 50% success rate.

Difficulty Target Number Check Modifier +1 Check Modifier +5 Check Modifier +9 Check Modifier +13
Very Very Easy 12 Basic (L1)
Very Easy 16 Heroic (L9) Basic (L1)
Easy 20 Champion (L17) Heroic (L9) Basic (L1)
Routine 24 Epic (L25) Champion (L17) Heroic (L9) Basic (L1)
Hard 28 Epic (L25) Champion (L17) Heroic (L9)
Very Hard 32 Epic (L25) Champion (L17)
Very Very Hard 36 Epic (L25)

I was tempted to use the tier names for these difficulties (almost the right number of them, and an average Basic character should expect to make DC 16 half the time on a check, while someone trained — an ‘expert’ — should expect to make DC 20 half the time) but it doesn’t work at the high level.

I am open to new names for the difficulties, I’m not really happy with the ones I have here… though they are simple.

If I want a challenge that most ninth-level PCs can expect to succeed at in a handful of tries, I’d look for an ‘Easy’ challenge (DC 20 — Dex 5 + Level Bonus 4 gives a total modifier of +9, for 50% success).  Someone trained (+4 competence bonus) or gifted (Dex 9) will have about a 70% success rate, while someone gifted and trained will have a 90% success rate.  On the other hand, a ‘Hard’ challenge (DC 28) is only going to be successful on the first try for about 10% of average ninth-level characters (and 50% of all gifted and trained characters).

This sounds pretty reasonable so far.  The table above is arranged incorrectly for the direction I’d actually be working.  It identifies the tier at which a character with a particular ability score and training should expect 50% success, not the success for the various tiers and ability score and training.

Hmm.  The obvious transformation leads to either a big, ugly table, or several tables (for +1/+5/+9/+13 total)

Closing Comments

It does seem a little odd that DC 20 is considered an “easy check” for a trained first-level character, but it does appear to work out that way for the 50% success rate.

If I wanted to aim for a 70% success rate, much as D&D 4e, I just need to shift the DCs down by four (so ‘Very Easy’ becomes DC 12 rather than DC 16).  I may do this anyway, since it makes ‘Routine’ DC 20 and thus 50% success rate for average trained characters.  This is probably minor detail work at this point, I’m pretty happy with how the numbers shake out.

I can still have intermediate check DCs (14, 18, etc.) but I don’t know that I’ll bother for the fixed-DC checks.  They may still come up for calculated DCs, but I’m not sure how many of those I’ll have.


  1. Pingback: KJD-IMC » Links of the Week, September 12, 2011

  2. I would say go for tier names for the named difficulties! Call 16 Basic, and make up separate names (“Wimpy”) for the less-than-Basic difficulties. This makes it really easy to see what sort of challenges the PCs can overcome routinely; it’s exactly the ones with the same tier as them.

    As far as the design process goes, I would address it more like “who is this aimed at?”. To take the classic example of a lock, who is it designed to keep out? It should then be a lock of quality above those people’s tier (to keep out Expert thieves, use a Master lock or better). Now the PCs may be Master themselves, so it’s routine for them! On the other hand, a lock designed to keep out the Master PCs would be Legendary or Epic — provided you can get a locksmith good enough to make it, of course. Likewise, the baron’s wall was designed to be unscalable for enemy soldiers, but he’d only considered Basic infantry so the wall is a Heroic challenge; the Master PCs can waltz over it.

    And the baron’s bathroom lock is, of course, only designed to protect his privacy, so it is Wimpy. Actually I mightn’t even bother with a check for modern bathroom locks, which are designed to be trivial to “pick” if you have so much as a butterknife to hand!

  3. Pingback: Reconsidering Level Bonus | Echelon d20 - An RPG framework based on the d20 system.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Back to Top