In the end it boils down to what kind of scaling you want. Linear, exponential or something like log.
In a linear system every skill point is worth the same. In an exponential system the later points are more impactful than the earlier points, while in a log system it's opposite. Just look at the graphs of those functions.
Math point would just be *skill (linear), ^skill (exponential) and log(skill). You can then freely add scaling factors (multiply by a constant) to adjust the exact growth you want to have.
Another thing you should consider is your RNG. Do you want a uniform distribution? Or a normal distribution? Think rolling a d6 vs rolling 3d6 and how likely each result is in each case.
In the end it boils down to what kind of scaling you want. Linear, exponential or something like log.
In a linear system every skill point is worth the same. In an exponential system the later points are more impactful than the earlier points, while in a log system it's opposite. Just look at the graphs of those functions.
Math point would just be *skill (linear), ^skill (exponential) and log(skill). You can then freely add scaling factors (multiply by a constant) to adjust the exact growth you want to have.
Another thing you should consider is your RNG. Do you want a uniform distribution? Or a normal distribution? Think rolling a d6 vs rolling 3d6 and how likely each result is in each case.