GURPSよろず

　The Basic Set limits heroes to four levels of any given Talent. This cutoff was arrived at by viewing the reaction bonus as similar to that from a Reputation, and subjecting it to the same upper bound: +4. However, the foremost game-mechanical role of Talents is to aid skills, and in that sense they aren’t unlike limited attributes. Especially in a campaign where the GM is using Alternative Benefits (pp. 18-19) and not reaction bonuses, if attributes are capped at 20 (see p. B14), then “Why not 10 levels of Talent?” is a fair challenge.

　The short answer is, “Go ahead!” Published GURPS supplements allow PCs to have more than four levels of Talent; e.g., Dungeon Fantasy 11 suggests six. To best match the upper limit to a given campaign, though, it’s useful to examine each benefit of Talent on its own terms:

　Learning-Time Reductions: Even Magery – which is openended and supernatural – posts a limit of “60% of the usual time” (p. B235), or -40%. Going down to -50% would double learning speed; that’s comparable to Intensive Training (p. B293), which is tolerable enough if the GM keeps a close eye on it. However, if a PC can acquire a point in 80 hours or less (-60% or lower), then rules-lawyers will exploit this mercilessly to improve skills “for free” almost weekly. It’s fairest to say that teaching, reading, practice, etc. take some basic, irreducible time, and stop at -40% or -50%.

　Reaction Bonuses: Most traits that grant +1/level to reactions stop at four levels; these include Appearance (mortals are restricted to Very Handsome/Beautiful, averaging the +4 described for Androgynous), Reputation, and Social Regard.
This is because reaction bonuses are cumulative and reaction results higher than 19 aren’t defined (see p. B561). Capping bonuses at +4 from a given source is one of the game’s simple controls on this. Still, Very Handsome/Beautiful gives +6 to reaction rolls made by those attracted to the good-looking person’s sex, so such a bonus might be fine if the group affected is at most half the population.

　Skill Bonuses: Skills face arbitrarily high penalties – unfavorable circumstances can pile up -5, -10, -15, or worse on a success roll. Bonuses to offset these penalties are therefore intrinsically balanced. If the GM caps DX, IQ, HT, Will, and Per, though, then Talent should face comparable bounds. The standard human limit for such scores is 20; that is, +10 relative to the average of 10.

　When setting the campaign’s maximum Talent level, the GM must decide what limit he wants to place on each benefit, which of the three is his highest priority, and whether breaking the symmetry between them is compatible with his sense of fair play and/or aesthetic preferences. This leads to several options, summarized on the Talent Limits Table (p. 22):

　If the GM believes that all three effects must advance together for Talent to justify its point cost, then he should use the original limit of four levels – or five, if he can live with +5 to reaction rolls and -50% to learning time.

　If the GM is comfortable with capping reaction bonuses – or just isn’t using them (most likely because of Alternative Benefits, pp. 18-19) – but insists on all skill-related effects tracking each other, then five levels of Talent work best.
This gives +5 to skills and -50% to learning time, but merely +4 or +5 to reactions (if used).

　If the GM is using reaction bonuses and regards those as the critical limit on Talents, but can live with +6 from “people impressed by this Talent” because that’s generally smaller than “those attracted to your sex,” then six levels of Talent can work. This gives +6 to skills and reactions, but just -40% or -50% to learning time, depending on what the GM feels comfortable with.

　If the GM is willing to cap everything but skill bonuses, then he can safely permit up to 10 levels of Talent, for +10 to skills – but still just -40% or -50% to learning time and +4 to +6 to reactions. Pushing past even +10 to skills in a narrow area might be fine; e.g., if the GM limits Per to 20 but also allows Acute Hearing 4, for +14 to Hearing relative to the human norm of 10, then he may accept Musical Ability 14!

　The GM should avoid the temptation to discount the point cost of Talent levels past the fourth when these grant only some of the customary benefits (e.g., Talent 5 that gives +5 to skills but just +4 to reactions and -40% to learning time).
Such markdowns lead to players giving PCs unsuitable Talents simply because the bargain is too good to pass up. If the GM wants to weight lower levels of Talent less heavily, so that each level buys roughly equal benefits, then the fairest way to do so is to pick maximum skill, reaction, and learning effects that divide evenly by the levels available, and then spread them out over the entire range.
Doing this with eight or 10 levels of Talent mutes reaction effects for all but extremely talented individuals, which is yet another way of addressing the objection that Alternative Benefits (pp. 18-19) exists to combat. The “Staggered” scheme on the Talent Limits Table illustrates the most straightforward example.

　Most of the substitute abilities under Alternative Benefits (pp. 18-19) don’t scream out for special caps. The GM should allow them to track skill bonuses up to whatever maximum Talent level he allows. If they seem unusually powerful, though, they might be subject to the limit on reaction bonuses instead – or kept to modest levels by invoking Alternating Alternatives (p. 19).