Reconciliation of adventure rewards and the GURPS skill learning system

I've posted a few times on the topic of learning and skill acquisition in GURPS. There seems to me that there is a drastic disconnect between getting a point at the end of a session and placing it in a skill, and having to learn for 200 hours in order to accomplish the same. Learning is learning, I propose here a records -keeping free way to handle abstracted learning which is compatible with the possible fast-pace of adventuring-based learning.

My motivation here is for the campaign Reclaming Khazad-Dum, where a large number of NPCs will have to learn vital skills, and time is sensitive. It can be applied in all other situation where PCs want to improve an abstracted manner.

Punchline : Quick table for acquiring skill points through learning/doing things

Target on 3d6	Hours of activity
3	2
4	8
5	20
6	40
7	64
8	100
9	150
10	200
11	250
12	300
13	330
14	360
15	380

Modifiers: Learning on the Job: -1, Self-teaching: +0, Education with a teacher: +1, Intensive training: +2,
Learning beyond 16 character points:-1, beyond 20 points:-2, important use in the narrative: +4

• This table should be used to determine whether abstracted learning yields a character point in a given skill. 
• It is better than the RAW because: 
• it requires no records-keeping of time. 
• Less predictable.
• More compatible with the faster point acquisition rate when adventuring.
• Single 3d6 resolution : Approximate up the cumulative learning time, apply modifier, roll, apply results. 

---

Geek zone : Deriving this table

First, let's get rid of keeping track of hours. This makes the whole thing very predictible, and I really dislike records-keeping in the first place. The base learning rate is 400 hours of learning to acquire one point (self-learning). Let's turn this concept on its head and assume that it take on average 400 hours, and that the likelihood of acquiring a point in a skill is 1/400 per hour. This probability leaves us with an expected learning rate unchanged, but less predictable.

P = 1/400 is exactly the probability to get a 3 with 3d6. How convenient. Also, one of the coolest property of 3d6 is the fact that the probabilities are roughly doubling for each increase in target score on the dice roll (except for the 9 to 11 range, where it is flatter). This was the whole basis of my own resolution system that I created before I discovered GURPS. This properties thus implies that the probability of acquiring a point roughly increase by 1 for each power of 2 hours spent learning.  Here is a quick table based on self-learning where it takes 400 hours to acquire point:

Target on 3d6	Hours of activity
3	2
4	8
5	20
6	40
7	64
8	100
9	150
10	200
11	250
12	300
13	330
14	360
15	380

The probability of acquiring 1 point rises to 50% after 256 hours. I first modeled using a power of two progression, but the actual values based on the probability distributions make more sense, even though the table is less intuitive.

Self-Teaching, Learning on the Job, Teachers

According to Basic Set p. 292, the rate of learning is affected by the condition of learning, conveniently in powers of two!

Learning mode	Check Modifier	Effect on Rate
On the job	-1	X 1/2
Self-teaching	0	X 1
Education	+1	X 2
Intensive training	+2	X 4

This is pretty straightforward and matches the RAW well enough. The only difference is that I used the self-learning rate as X1 instead of the Education rate: people usually read on their own.

 Discussion and parting shot(s)

This translation of learning time into probabilities slightly increase the learning rates if 3d6 are rolled more often rather than wait for longer period before rolling. This is acceptable if we assume that the best bang in learning happens earlier, and refrain from exploiting the mechanics by breaking longer abstracted learning times into smaller chunk just to roll multiple times. 

My favorite aspect of this alternative mechanics, beyond the jettisoning of records-keeping, is that it simulates somehow adventuring-based learning, where it is possible to acquire a point into a skill in much less time that 200hrs of dedicated education, or 800hrs of on the job training! 

Acknowledgements

Thanks to +Chris Bradshaw for his suggestion about slowing the rate for high skill levels.