The more I think about attribute increases, the more I think I may use 1E's limit on how often you can boost an attribute. Effectively, when you make your character, you've set the 'average' for your character's attributes. Considering your character's life, health, and so forth, your attributes are more or less fixed at what they are. I'll allow each attribute to be increased once - and only once.

Sure, working out and such could in theory do something, but that's more what your skills are for - if you jogged more often, that wouldn't suddenly make you better at all skills related to your Body Attribute, or your Agility. If you hit the books, you'd get better at one or two skills, not every Logic skill. So, yeah, I think attributes should be set more or less at chargen.

We'll see.

You're proceeding from a false notion. Jogging won't make you better at "agility skills", but the right kind of training could. Hitting the books on a subject or two wouldn't make you distinctly smarter, but working on the actual way that you think COULD. It would generally take longer as you're not dealing with the same sort of focused and directed training, but the point certainly stands.

Perhaps, but then if you slack, shouldn't your attribute degrade, then? The thing is, I figure you're character's general daily routine is what gets you to that 'point', and doing things 'differently' to get your stat boost is probably out of character for your character. And then it would need to be kept up over time. This is, I think, what would warrant the +1 to your attribute.

One thing you could do instead would be to make a simple house-ruled system for it, to slow it down and make it more of an investment that doesn't pay off for awhile. Consider the following (off the top of my head):

Assuming 4th or 5th edition.

Training to the next level of an attribute is a considerable investment in time and resources. When a character decides to improve himself in this way, he spends the necessary Karma (new level x5, if I'm not mistaken) immediately, but doesn't gain the benefit right away.

Instead, the character makes an extended Attribute test with a threshold equal to the new rating, and an interval equal to the new rating in weeks.

For example, if a character wants to improve Body from 4 to 5, he spends 25 Karma, then rolls 4 dice. Five weeks later, he makes his next roll (in diminishing returns for extend tests, this would be 3 dice). Five weeks after that, he makes his next roll, and so on.

While a character is training an attribute, he cannot be training another attribute.

If you're using diminishing returns, once a character is reduced to 1 die, he is allowed to continue rolling that 1 die until he reaches the threshold (which would be the only way for a character to improve from 1 to 2, by the way).

So if a character wants to improve Agility from 2 to 3, he might roll 1 hit on his first roll, no hits on his second or third rolls, one hit on the fourth roll, no hit on the fifth, and one hit on the sixth. The entire process in this case would take 18 weeks of training.

If he wants to improve from 5 to 6, he might roll 2 hits on his first roll, 1 hit on his second, 1 hit on his third, 1 hit on his fourth, none on his fifth or sixth, and 1 on his seventh. In this case it would take 42 weeks, or most of a full year.

I'd recommend reducing the threshold and interval by the racial modifier for an attribute, so that trolls don't take 2 years to improve their Body.

Except that there's a difference between training to get to a point versus training to maintain - specifically, the need to push yourself beyond what you might normally consider to be your limits. And given that it's an expenditure of karma involved, it should be assumed that it IS out of the norm - and a break from normal routine for a specific purpose is not, by it's nature, out of character, only in very specific cases could that ever come up as an issue.

Epicedion: That system seems much too onerous from a gamist perspective.

This seems both of little practical value and needlessly complex.

~Umi

I'm not sure this is entirely related, but this reminded me of a TED talk / interview with the person who presented it.
Took me a little while to locate a version with a transcript in case you didn't want to listen/watch. This guy (Sugata Mitra) did some experiments where he put a computer in a hole in the wall in rural India, programmed in English, and walked away, just to see what the kids would do. In as little as a few months they'd taught themselves English with no instruction.

http://tpr.org/post/how-much-can-children-teach-themselves

Search for "I decided I would destroy my own argument by creating an absurd proposition" and read to the end (its only a couple of paragraphs).