Increasing Skill Levels From default

[Pseudonym]

Originally Posted by Terwin

Remember: when you rearrange a skill default tree, no skill can ever go down. So if you change the tree such that a skill would have a lower value that it did previously, you *MUST* spend points to bring it back up to it's prior level.

As such I suspect that you cannot do your trick of rearranging the default tree to maximize each skill in turn, as you would then need to buy up any non-maximized skill back to its maximized value.

Ah dang, looking at it closer again, it appears to me you are right.

As an aside, even though it isn't RAW, I feel like it would be a good simulation of the realistic idea of certain skills giving each other a certain synergy. EG, in the real world, If someone were an expert level in all of Chemical Engineering, Materials Engineering, Bio-Engineering, Mechanical Engineering, Bio-Mechanical Engineering, Electrical Engineering, Electro-Mechanical Engineering, and Computer Engineering, I think that each skill would have at the least a teeny synergistic effect with the others, some a little less than others. But for example, just because someone uses their insights from Electrical Engineering to be better at Computer Engineering, doesn't mean they can't use their insights from computer engineering to simultaneously be better at electrical engineering.

BUT, that all being said! It's not how GURPS works RAW.

[Desthro]

From what I understand, you can't double default with chains anyway. So you can't have skills A, B, C, with A being the main skill, and then default B from A, and then default C from B. C would ALSO have to default from A. This is true for any given set of defaults.

This prevents default "trees" since one skill bought up from default cannot be used to buy another skill up from default. Programming it is fairly simple IMHO, since you systematically have to test each skill in the set as the primary skill. (You might even test having multiple primaries, but since the default groupings are fairly limited this might be harder than necessary.)

It vastly simplifies the problem- (Capital denotes primary, lowercase denotes defaulting)

Given:
1) All skills are Average

2)
a defaults to B-1 or C-2
b defaults to C-1, D-2, or A-3
c defaults to D-1 or A-2
d defaults to A-4 or B-2

For the given set, the following are possible with one primary:
A, b:A-3, c:A-2, d:A-4
B, a:B-1, c:N/A, d:B-2
C, a:C-2, b:C-1, d:N/A
D, a:N/A, b:D-2, c:D-1

In terms of efficiency, ANY skill that can have ALL other skills default to it, is going to be the most efficient with an arbitrarily large number of points. Total Skills = 4, so any primary skill that has 3 skills defaulting to it, is likely to be the most efficient. In this case, only A has 3 defaulting skills (b,c,d) so it is going to be the optimal skill for total bonuses for an arbitrarily high number of points. B, C and D only have two defaulting skills so they tie. In this case, the total number of penalties is probably going to be a strong indicator of efficiency. B has total penalties of -3, C has total penalties of -3, and D has total penalties of -3. In this instance, for this particular set of skills, A is the most point efficient, with B, C and D tying. A tie-breaker would likely be based on which skill is the LEAST important, so if skill C is the least important, skill B would be second most efficient.

This is true for any set of skills with a single primary skill. If you add multiple primary skills for any given skill set, it changes a little bit, but is still largely the same.

[David Johnston2]

Originally Posted by Pseudonym

But the second part of the problem... the shuffling of the points is what gives me the biggest problem. Once the optimum path of default learning has been established, what is the smart way to optimize points?

...Raise either your IQ or your DX (or both). Then dump your remaining points into a skill that particularly expresses the character.

[Desthro]

Originally Posted by Pseudonym

Ah dang, looking at it closer again, it appears to me you are right.

As an aside, even though it isn't RAW, I feel like it would be a good simulation of the realistic idea of certain skills giving each other a certain synergy. EG, in the real world, If someone were an expert level in all of Chemical Engineering, Materials Engineering, Bio-Engineering, Mechanical Engineering, Bio-Mechanical Engineering, Electrical Engineering, Electro-Mechanical Engineering, and Computer Engineering, I think that each skill would have at the least a teeny synergistic effect with the others, some a little less than others. But for example, just because someone uses their insights from Electrical Engineering to be better at Computer Engineering, doesn't mean they can't use their insights from computer engineering to simultaneously be better at electrical engineering.

BUT, that all being said! It's not how GURPS works RAW.

Sounds like a GREAT explanation for the TALENT ADVANTAGE.

[Pseudonym]

Originally Posted by Desthro

A buncha good math

I think I get what you are saying, and it was another thought I had because I was thinking that maybe the rule about using a skill at double default not being allowed (eg a is b-5; b is c-5; a is NOT c-10) would also extrapolate from this case.

So, we try reversing the relationship and instead think about what skills default to skill A, not what skills A defaults to.

As long as any skill is improved from default of Skill A, skill A may not improve from default of any other skill.

It does make the math easier in a way because we don't need to recursively see if a default of a default is better than a default of a default of a default.

So this doesn't seem like a process best served by automation because it requires a human that understands the context of his campaign to say "a and b to c earns me more value than b and c to a" But maybe some kinda hint that someone could see that would say when looking at skill c that says, "you could get 1 extra level on skill c if you developed from the default to a" or "you can't change skill d to develop from IQ unless you also spend 3 points on skill d"

And yeah David Johnston 2, the obvious obvious solution for really optimizing a character would be investing in talents or base attributes, I get that, but in a hypothetical situation if a character found himself in a place where he already had a bunch of points in skills that default to each other, you are, by the rule of developing from default, allowed to rebalance the points between skills defined as having a dependency relationship....
Now if we want to convolute the original problem even further, I think the social engineering supplement on learning says that if you happen to have a lot of extra points in skills based off of the same attribute, you can take them out of the skills and reinvest them in the base attribute at a slightly inefficient exchange rate... Except the go/no go call on that one is pretty obvious so I don't care to work it out (If I can subtract points from x skills, so that k skills go down a level, it is probably worth it if it makes at least k+1 skills improve by 1.)

[simply Nathan]

Me, I could never understand this rule from the examples in the book. If anything, I understand it even less when I see explanations of the math at play (as in this thread). Personally I had just been hacking the Technique rules, allowing purchase of a Hard technique "I use (X defaulted skill) at reduced penalty" up to the point where the default equals the main skill. Then the Skill Adaptation perk came out and I've never looked back since it was an even cheaper and even easier way to do the same thing.