The Promised Shadowrun Fixes

But first, a retraction. The last post was not, in fact, my seventh. This post is. Boy, do I feel dumb.

To begin with the Shadowrun.

First thing, I'd add a couple of skills. Perception needs to be a skill. Right now, it's simply an Intelligence roll. I want characters who train their perceptions. The concept of the only moderately bright, but terribly sharp-eyed character is entirely absent from Shadowrun. The other skill I'd add is Dodge. Currently, dodging is handled by rolling dice from your combat pool, which means that the only way to get better at dodging is to get better Intelligence, Agility, and Strenght. Again, not particularly good, genre-wise. I don't think this idea skews the balance particularly, either. While some dodging can be accomplished regardless of how much combat pool you have left, you're now limited on the amount of combat pool you can spend on dodging (no more than the amount of skill dice you have), and combat characters now have another skill they have to spend stuff on.
Speaking of the combat pool: I'd stop making it universal for all characters. Instead, I'd allow each character to pick a knack. Your knack gives you a dice pool that you can use for a set of related tasks. The pool is determined by taking the averages of two or three attributes which are used for the skills involved. The Fighter knack gives you a combat pool, which I would also extend to Athletics skill rolls made in combat, allowing a fighting type to run faster, climb better, etc, while "in the groove". Basically, that's what a knack represents: a set of skills for which you can get "in the groove", doing better than your pure skill would represent. Other knacks would include Magic (which would expand to cover summoning and banishing spirits, rather than just casting spells), Face (charisma and social skills, basically), Pilot (driving/piloting skills, maybe split this into Driver and Pilot), Hacker (computer skills), Gageteer (build/repair skills), and Ninja (sneaking and perception skills). You could also add other knacks, like Athlete, Craftsman, etc. A character could buy a second knack for 10 points, or recieve 10 points by choosing to have no knack. Basically, this would tend to secure niches, and let each character to particularly excell at something. Dice pools would refresh at different rates, depending on the type of knack. Combat pool refreshes every round, as normal. Social pool would probably refresh every encounter, as would Sneaking pool. Hacker and Magic and Driver, probably every round as well. Another change to dice pool mechanics: You can expend one die from the pool for the rest of the session, in order to get an automatic success. So, if you want, you can insure that the one hit you need to make does hit, but you'll be handicapped in combat for the rest of the session.
Karma pool would be given to everyone, and probably work the same as normal. I think I'd give out 2 karma pool dice to start with, instead of one, but that's just because I prefer a more cinematic game where heros are lucky.

Skill costs: I'd completely change the way skills were costed. I'd make them cost the skill level squared in karma, up to double the related attribute, whereapon there'd be a further 50% surcharge. However, the cost of whatever level you're currently at in the skill would be subtracted from the final cost. For example, Bob has a skill of 1 in Stealth. Having been shot one too many times, he decides to boost this to 4. A skill of 4 would normally cost 4^2, or 16. However, it would only cost him 15, since he already has skill level 1 in Stealth. If his Agility (Stealth's base attribute) were 4, then Bob could buy up to level 8 in Stealth at this normal cost. However, level 9 would cost him 121 karma, rather than 81, due to the 50% surcharge for it being above double his Agility.
Skill learning time would be equal to the skill level squared times 2 in months (assuming 4 hours a day studying, with no days off). By doubling your karma cost, you can reduce to level squared weeks, and by quadrupling the cost, to level squared days. In addition, you can make Will rolls to study more per day. This is TN 4, and each success adds one hour to the time spent studying per day, up to a maximum total of 8 hours. Like karma costs, your current skill level is subtracted from the total training time.
This gives a result of 18 months for a skill at level 3, generally considered professional level. That's a pretty hard course of study, too, and of course a professional needs several skills at 3 to be really good. Thus, the training time seems reasonable.
Attribute costs would be similar, except that karma costs would be level squared times 2. So, level one is 2 karma, 2 is 8, 3 is 18, 4 is 32, and so on.

Defaulting needs to change. Currently, if you're defaulting, you roll your relevant attribute dice, but with a -4 penalty, and a maximum result of 8. However, if you have one level of skill, you have to use that skill. So, for example, Bob has no Stealth skill, but an Agility of 6. He rolls 6 dice for Stealth tests, but at a -4 penalty to each. However, if he gains a level of Stealth, he can only roll one die, albeit at no penalty. At certain target numbers, his chance of successfully making his roll has actually decreased. This is bad. I think an Attribute of 4 should reduce TN of skills based on it by 1, and increase the default cap by 1, while 6 would decrease TN by 2, and increase the cap by 2, and so on. Also, I think Attribute dice should "convert" to skill dice as your skill level goes up. That is, if you have Agility 6 and Stealth 1, when rolling a Stealth roll, you would roll one die normally, and 5 dice with the normal penalty and cap. When you reach Stealth level 6, your skill would completely eclipse your Attribute, and it would clear from then on.

That's all I can think of for now. I'm going to bed.