Learning a Skill at Level 0 (which is where I start all new skills, that is, first you learn it at Skill 0, then improve to Skill 1, Skill 2 and so on), is 4d6 weeks +0d6 weeks for Skill 0... or simply 4d6 weeks. You don't make a skill check to learn it, you use a base attribute check... whichever the GM thinks is most relevant. That might be Int or Edu for academics skills, or as in the case of Athletics it might be any of the physical characteristics (Str, Dex or End). There is no -3 to the attribute check.
Looking back at my example, I think I should have clarified that when opening a NEW skill at Skill 0 its an attribute check. For improving a skill you already have, it would be a normal skill check. Heh... editing
Another thought I have tinkered with was that during play, any skill which gets either an Exceptional Success OR an Exceptional Failure gets a +1 DM on the next skill improvement attempt. But I'm concerned that's going to drift toward a lot of book keeping again with lots of "check marks" on the skill sheet to keep track of them.
I'm still tinkering with the rule. Mainly I wanted to get away from the official method which I thought was completely unrealistic and required too much book keeping. I'm still not sure if I'm happy with my solution but as you said, its a starting point. Figured I'd put it out here and let you lot kick it around and toy with it.
I have considered another possible method that would require weekly skill checks and a cumulative Effect to be tracked, requiring a character to accumulate X of Effect points to improve the skill by 1 level. That might be more realistic but would require more book keeping, though if I reprinted the skill sheet to include a small space in brackets to note the current "Training Progress" (total number of effect points) it wouldn't be much book keeping I don't think. That would allow several skills to be progressing at the same time as a combination of training and practical experience (combat, research, doing your job, etc.) gradually accumulated "experience" towards the improvement of various skills. That would more simulate how we learn in reality.
Such a system would probably look something like this...
Learning a New Skill (Open at Skill 0)
Make an Charcterisctic Check (Attribute determined by GM) per week of training. Accumulate a total of 16 Effect.
At GM discretion, some skills cannot be learned without either an Instructor, study materials, or a relevant skill prerequisite (still considering this, MGT doesn't have prerequisites and this might be over complicating things even if it is realistic).
Improving an Existing Skill (Skill 0+)
New Skill Level x 20 Effect Total
Make one Skill Check per week and add Effect result to Total
Hiring an qualified Instructor (possesses both Instructor skill and skill being taught at least 1 level higher than character's desired skill level) adds Instructor level as a +DM to Skill check.
Negative Effect totals from Failure can reduce the Total (study harder next time).
An Exceptional Success OR an Exceptional Failure from skill use during play gives a +1 DM on next training check (we learn as much from our failures as our successes rule).
Example... Player wants their character to Open Athletics as a new skill (Skill 0). The GM decides any physical characteristic is relevant. The characters best physical attribute is Dex at 9 (+1). The character does some exercises, tumbling and gymnastics in the cargo hold during the week spent in Jump Space and gets a Skill Improvement check. The player rolls a 9 + 1 = 10 - 9 = Effect 2. A [2] is noted next to athletics to indicate 2 Effect points have been accumulated towards opening Athletics at Skill 0. During their next jump, a passenger turns out to be a gymnist and offers to help train the character. The Gymnist has Instructor 2 and Athletics 4, which makes them a qualified Instructor for a +2 DM to the check. The player rolls an 8 + 1 +2 = 11 - 8 = 3 Effect, this is added to the running total for a [5]. During their next port of call, the character gets in a fight and tries some gymnastics to get out of a pinch... and fails spectacularly... suffering some injuries and learning a lesson in what NOT to do... add +1 to the next check. This time the character talks to the Gymnist teacher who has Instructor 4 and Athletics 4, giving an additional +4. The player gets a lucky 11 + 1 (Dex) + 1 (for learning the hard way), +4 (instructor) = 17 - 8 = 9 Effect! This raises the total to [14]. The player decides one more week spent training on his own in the cargo hold will be enough and bid's his gymnists friends fairwell. The following week in Jump Space gets a roll of 9 + 1 = 10 - 8 = Effect 2 which raises the total to 16. The character opens Athletics at 0 and erases the Effect total. To go from Athletics 0 to Athletics 1 will require a total of 20 Effect points.... maybe one of those self training courses in the adverts would be a good investment...
In the example, the character Opened Athletics at skill 0 in just 4 weeks of training. This was with the help of two different instructors, one of whom was very skilled, and a lucky die roll. Had one of those checks been a failure with -4 effect, it would have set the character back 2-3 weeks.
An average character, without instruction would take 8-16 weeks to open a new skill. With a +2 instructor that same character would take 6-8 weeks.
Going from Medic 3 to Medic 4 would require an Effect Total of 4 x 20 = 80. So assume Int 8 and Edu 8 and Medic 3 would give an average skill check each week of training of (7+3 = 10 - 8 = 2 Effect), thus requiring 40 weeks to improve the skill. However, a high Edu or Int character (12+ for a +3 bonus) would get an average of 5-6 Effect and make the same improvement in 14 to 16 weeks. This doesn't include bonuses for Instructors or practical experience. Not sure if the Effect Total should be raised or not. Thoughts?
Training doesn't have to be consecutive... you could train in one skill one week and a different skill the next week, the Effect Total doesn't go down (not entirely realistic but I don't want to add another rule mechanic for skill degradation over time, too much book keeping).
I'm thinking I like this second method better, but still pondering. Feedback welcome of course.