I like the idea of automating arbitration. However, I’m not eager to have to pay attention to contract status from year to year, and the whole strategic framework would shift quite a bit under the escalating salary concept. If we want to automate the process, let’s keep the $ amount to $300, like it is now, but simply distribute it according to a predictable algorithm. To mimic the split of $ added to great players vs. high-surplus players, we could distribute $150 based purely on points, and $150 to fill in low salaries among high performers.
For example, add $6 to the top 6 finishers in points, $5 each for #7-12, $4 for 13-20, $3 for 21-28, $2 for 29-36, $1 for 37-48. Then add the number of $ required to bring all top 12 finishers up to $30, #13-24 up to $25, #25-36 up to $20, #37 or greater up to $15 (until the $150 run out.) Depending on league salary details, you probably run out of dollars around #50 or so.
This approach would push most of the dollars to the best players without introducing a lot of complexity and unpredictability to where the dollars are going. It would also keep valuations much more in line with the current model.
I chose top 50-ish because that was 900 points this year, and it seems about where the high-level players sit. There might be a better way to formulate an automated arbitration scheme, but this seems like a decent place to start.