In explaining why I am considering a reconfiguration of D&D and the d20 system, I mentioned the key goal of making a single universal skill system the driver of the entire gaming process, so that it would cover skill resolution, combat, magic and saving throws. Under such a model there would be no process which could not be resolved by application of a suitable skill to a task. This begs the obvious question, since I am reconfiguring D&D: why the d20 skill system?

I used to be quite enamoured of the idea of a skill system with grades of success, as represented by the over-complexity of Rolemaster. Under this model, a skill check gives a series of possible outcomes – complete or partial failure, partial or complete success. However, continuing attempts to apply this skill system in actual gaming situations over many years (using Rolemaster, derivatives of Rolemaster, and various homebrewed systems) all led to the same problem: many tasks aren’t amenable to a graded system of categories of outcome, and these categories often bear little relationship either to reality or to the relative powers of the characters as represented in their attributes. For example, should a Rolemaster thief with a pick locks bonus of 36 be the sort of character who routinely gets partial or complete success against a lock which is, say, “difficult” to open? The whole process of assigning difficulties and results is too abstract. And in the case of opening a lock, what is “partial success” anyway? Either it opened or it didn’t. All too frequently, “partial success” meant “try again”, which really just slows gameplay down considerably. In situations where partial success might be relevant – say, trying to sneak up on someone – then it was the skill of the opponent which ultimately became  relevant to success, when they got a chance to do a search check as a consequence of their partial awareness of their stalker. And then how does one resolve a result of “partial success” on both sides of this challenged skill check?

In such instances the character is essentially making a challenged skill check against their target’s skill, but mediated through the assignment of difficulty levels dependent upon the environment. To follow our previous example, first one assigns a difficulty to the thief’s stealth attempt dependent upon lighting, cover, perhaps also the target’s degree of alertness; and then, one assigns a difficulty to the target’s perception check dependent upon the result of the player’s roll. It would be much easier just to have a direct challenged roll – one skill against the other – and apply a net adjustment to the player’s roll dependent upon the environment.

This is exactly how the d20 system works. The difficulty is set by the environment or by the target’s skill check, and the result is either success or failure. The success contains within it a certain level of magnitude, given by the difference between the roll and the target number. This skill check system significantly simplifies the process of skill resolution, and speeds it up. Even combat rolls as currently envisaged contain an element of this. The attacker makes a skill check, and the target is the enemy’s AC, which behaves very much like a challenging skill. If one were to replace AC by, for example, the result of a reflexes save, one has essentially a challenged skill check. One could then go further, and rule that all combatants are assumed to be “taking 10” if they also attack … which is what the current rule system pretty much assumes with AC. 

This can be taken even further if one takes the difference between the target number and the roll as damage done, with some appropriate weapon-related modifiers. So if the opponent’s AC is 15, and the attacker rolls 18, 3 hps damage are done (plus strength and a weapon-related modifier).  In such a case we have wrapped the skill roll and its consequences into a single determination, removing the need for further dice rolls and decisions, and relating damage done in combat directly to level of skill. 

Of course, the same considerations could be applied to magic. The target for the spell is set by some kind of saving throw, and the difference between the spellcraft roll and the target number, multiplied by spell level, is the damage. From this determination system it is a simple step to eliminating spells altogether. Effects could be given a level, so that for example daze is level 0, stun is level 1, fear is level 2, etc., and  the duration of the effect is determined by the spell roll… 

This covers all the major aspects of the role-playing process. The best part, however, is that under the d20 system the scale on which difficulty is determined for an action (usually between 9 and 20, for example, or a d20 roll plus a skill) is very similar to the scale  on which skill bonusses operate (usually between 1 and 20). So there is a natural symmetry in using skills to determine the difficulty of all tasks, and using skills to determine the success of those tasks. We can complete this symmetry by finding a way to relate hps to a skill in some way, and also relate the availability of spells to a skill in the same way. Having done this we have reduced the entire process of character development to the process of generating attributes, and generating skills. There will be no saving throws and no armour class; just skills, and skill modifiers.

This completes the other part of the challenge which I have always seen in role-playing systems – making a single, internally consistent system for resolving all tasks in an imagined world. There is no particular reason that one would have to do this except a sense of completeness, and it is this completeness I want to achieve by reconfiguring the D&D skill system.