(These rules are adapted from this blog post.)
The rules given here apply whenever a character wishes to conduct research into something—figuring out the workings of a unique magical object or phenomenon, inventing some new form of magic, creating a unique construct, etc.
A research project is an endeavor carried out with the aim of discovering or creating something new. Research projects are generally open-ended and exploratory in nature (i.e. a character who undertakes a research project doesn’t necessarily know exactly what will be necessary in order to achieve his goal).
What form the research takes depends on the goal of the project, but it’s usually a full-time job (that is, it takes a full day’s activity—8 hours—to engage in research, and can’t be done while adventuring, or while crafting magic items, etc.).
Requirements: Special equipment may be required in order to conduct research; or an appropriately stocked library; or facilities set up for the purpose; etc. Consumable resources of some sort may need to be expended, and these may be generic and easily acquired, or special and rare. The research may involve travel to particular places, or other specific activities.
Risks: Some forms of research may be dangerous, exposing the researcher to assorted perils (chance to accidentally poison yourself while developing a new concoction, random encounter with demons while perusing an extraplanar library, ability damage due to interacting with an alien device, etc.).
The rules given here govern research projects in general; as research projects may vary greatly in purpose and scope, additional rules may apply in particular circumstances.
A research project has a goal, which is determined by the character undertaking the project.
Complex Goals: A research project may have a multi-part goal, consisting of several closely related components. (The DM’s judgment determines whether several sub-goals “fit” into a single project—that is, whether a line of inquiry can advance a character toward all the sub-goals in the course of a single project—or whether the sub-goals are disparate enough to require separate projects.)
Example: Vasily has a magic staff (taken from a vanquished enemy) which has unusual powers. He wishes to determine how this strange magic works, and what is required in order to create another staff like this. These would be two research goals, which may be pursued in the course of a single project. (In this case, Vasily could reasonably expect that the first goal would be attained before the second; and the second goal may be unattainable before the first is achieved—see Progressive Success, below. This may not be the case for all research projects, however; see Modularity, below.)
After each full period of research activity (usually a day’s worth, but may be longer or shorter, depending on the project), the character makes a skill check, followed by zero or more progress rolls.
The applicable skill, and the DC, is set by the DM.
Knowledge skills are often used, of course; Spellcraft may apply to magical research; Gather Information (when pursuing some elusive piece of knowledge that may be known by someone, or some similar endeavor); Decipher Script; Craft skills; Disable Device (when attempting to figure out some strange mechanism, etc.); or other skills. A research project doesn’t necessarily use the same skill throughout; the relevant skill may change at each step along the way to the goal (i.e. any time a success is achieved).
Research check DCs are generally 10 higher than the DC for a corresponding skill check would be if the task were a “known” one, and so range from 25 to 45.
A research progress roll is usually a d20, but may be a different die—anything from d4 to d100. (The size of the die represents the learning curve of the project; smaller dice for projects that require less exploratory work and let a character get to the “meat” of the project immediately, larger dice for research that may present many frustrations and dead ends before a fruitful avenue of inquiry is stumbled upon.) The progress roll is an unmodified die roll.
A progress roll may result in a failure or a success. Track the number of successes achieved on a project.
A progress roll results in success if the die roll matches or exceeds the success threshold, and in failure otherwise. The success threshold is initially equal to the die size + 1 (i.e., 21 if the progress die is d20). Each failure reduces the success threshold for the next progress roll by 1.
When the character achieves the requisite number of successes, he completes the project (successfully creates what he was trying to make, discovers what he wanted to learn, etc.).
The number of successes required is generally not known at the outset (i.e. the DM keeps it secret). A character may have some indication of how much progress he’s made (see Progressive Success, below), or he may not.
The following additional rules apply to some research projects, or in some situations.
In some cases, each success can yield additional information or some other tangible benefit. Sub-goals may be achieved when certain milestones are reached (i.e. when some set number of successes on the way to the final goal are achieved).
After a certain number of successes, progress in the task may only be possible when some other prerequisite is met (unique resources or capabilities required to proceed, etc.). In many cases, the nature of the roadblock may not be known until the roadblock is reached.
Most (though not all) research projects can’t simply be brute-forced through, and require some degree of insight. Such projects have a limit on the number of attempts to make progress that a character can make; when that limit is reached, no more progress can be made. This limit is generally equal to the character’s skill modifier with the most relevant applicable skill (as determined by the DM).
If the character reaches the limit of his insight on a project, he can attempt to restart the project if he gains a rank in the appropriate skill, or gains access to some critical information or resources which he didn’t have the last time. (When restarting a project, the success threshold returns to its initial value, but previously achieved successes remain.)
Some research projects are modular—that is, some or all of the project’s sub-goals may be independent of each other, and may be pursued in any order. In such a case, the researcher chooses which sub-goal to pursue when attempting progress, and any successes that result from the attempt count toward the chosen sub-goal.
A project may be wholly modular (i.e. any of its sub-goals may be pursued and achieved in any order), or only partly modular (such as, e.g., a project consisting of four sub-goals, where the first three may be pursued in any order, but all three are jointly prerequisites for the fourth).
Multiple characters can work on the same project, in concert or in parallel.
Multiple Characters Working in Parallel: If multiple characters work in parallel, treat each as making progress individually (and paying all requisite costs—of time, resources, etc.—separately). If the project admits of progressive success, and one of the characters achieves a milestone or sub-goal before the others, he may share his discovery or accomplishment with other characters working on the project, who may then count that success as their own also.
(Cooperating in this way is often of use on a modular project; multiple characters may each pursue different sub-goals of the same project, and may share what they’ve learned or accomplished, collectively completing the project as a result.)
Multiple Characters Working in Concert: If multiple characters work in concert on the same project, treat them as a single researcher, but have each character make research skill checks for each attempt individually, and use the best result. (Each cooperating character still has his own limit on number of attempts, if applicable.)
(Cooperating in this way is often of use when a project requires multiple skills, of which each researcher on the team possesses only a subset.)
Researching a new spell is a research project. A research skill check may be attempted after each full day of spell research. The research requires the expenditure of 1,000 gp of special materials per day. (In the case of some spells, other requirements may apply.)
10 successes are required to complete the new spell. Research skill checks are mostly made with the Spellcraft skill (though other skills may be required also for parts of the project, in the case of some spells); the DC is 20 + 3 × the spell’s level.
The learning curve die (the die used for progress rolls) for spell research varies according to how unusual the spell to be researched is: spells that are simply direct extrapolations of existing spells (such as a hypothetical mass harm spell) use d4, whereas spells that are to have a totally novel effect that is unlike any existing spell may use d20 or higher.
Creating a magic item usually requires that the creator imbue his own power into the item, a process that involves synthesizing special magical substances, reagents, etc., which can hold that power; these are then used to construct or enchant the desired item itself. However, it is often possible to use, in a magic item’s creation, additional, special components, which take the place of some, most, or even all of the creator’s own power; doing this reduces the cost (gold and XP) of creating the item, and the time required to craft it.
Such special components may be physical (rare substances, parts of creatures, etc.), or they may be special locations, events or phenomena, etc. Not all magic items may admit of having special components used in their construction, and where this is possible, the required components are not commonly known, and require research to reveal.
Determining what special components may be used in an item’s creation is a research project. A research skill check may be attempted after each full day of this research. No cost is incurred, but a well-stocked library (containing material bearing on general arcana and topics relevant to the particular item) is required; in the absence of such, research skill checks are made at a −5 penalty.
From 1 to 5 successes are required to determine the special components that may be used in an item’s creation, depending on the item’s complexity; a straightforward or commonplace item requires but 1 success, while complex items, or those with novel or unique capabilities, require more successes. Research skill checks are mostly made with the Knowledge (arcana) skill (though other skills may be required in some cases). The DC is set in secret by the DM, but is generally 20 + the item’s level.
The learning curve die (the die used for progress rolls) for this sort of research is always d20.
Research to determine applicable special crafting components must be done separately for every sort of magic item (or closely related family of items). Once done, the knowledge thus gained may be applied whenever the crafter wishes, shared with other characters, etc. (For instance, suppose a wizard discovers that the intact heart of a bulette, harvested after the creature has consumed a live goblin but before having fully digested its meal, may be used as a special component in the creation of gauntlets of ogre power that cuts the crafting costs and time in half. He may then feed goblins to bulettes and harvest the hearts of the latter as often as he likes, cornering the market on gauntlets of ogre power and striking it rich—limited, presumably, by the supply of the said creatures. He may also record his discovery in his memoirs, relate it to his grandchildren, etc.)