OVERALL COMMENTS
To start out, I built a rough value system for each team. From there, I made incremental refinements to try and be more precise. This is an interesting setup by juice and the more I dug in, the more refinements were necessary. Most of these, of course, stem from the free agents, which add an interesting wrinkle (or really, wrinkles, I should say). My initial valuation was just based on each team's top 8 (or 9, for League 4) hitters, partial value for the top few bench hitters, and the top few starters and relievers.
My first refinement was adding in free agents. I did this simply by having a "max" and "min" for each team, where max was if you got 1st pick and min is if you got the last pick among all the teams for those years. Obviously, max = min for years with just one option, but for years with multiple options it could be a big swing whether you get 1st pick or not.
Next, I had to deal with the pitching valuation since just using the top X pitchers is very different in 1920 vs 2020. I was able to get a reasonable expectation of the "best" 1450 innings for each team. This helped a lot, especially for the early-era teams. This did, however, introduce a new issue which was that for modern teams, the best 1450 innings often included about 20 pitchers! More on this later.
Next up I took a stab at slotting "expected" FAs for each team. To do this, I had to take a guess at which 24 teams would be picked in each league. Then, I ranked each team within each year-group. The best team got the last pick *of the projected top-24 teams that would be selected*. Obviously, the last selected team got 1st pick. This was key for a handful of teams, but even more important was how variable this is. This will require constant management throughout each draft as well as a willingness to gamble as you can't know which teams will be taken later after you pick.
At this point, I am just adding the free agent values to the team values, but I know that means there's extra PA/IP being counted. I decide to go in an manually build each (relevant) team. I didn't do this in a super detailed way, but enough to get the valuation mostly correct. This solved two issues at once. First, each team got the same number of PA (with League 4 getting more with the DH) and the same number of innings. So, if I projected you'd get a 300-inning FA pitcher, the rest of your team only had to contribute 1150 innings. Same for hitters. The second benefit was on the hitting side to remove "fluff" for each team. What I mean by that is the teams that had 4 good outfielders but could only play 3. Or teams where the best free agents didn't fit well. Or the other one that came up was the "complete" teams were hurt--for instance the 1931 Yankees have 8 very good starters so while a free agent upgrades them, it's not as much as other teams that had a hole in their lineup.
Building the League 4 teams was interesting in two ways. For one, the presence of the DH both helped and hurt. On the plus side, there was room for deeper teams to put all their players and it was also easier to slot in the best free agent, as you could always slide them (or someone else) to DH. On the negative side, some teams just couldn't fill all the at-bats and you end up just putting fillers in there. The other is the usual roster spot issue. With both PA and IP being spread around to more players, roster spots become more precious. If you used all 13 pitching spots, that's only 12 hitting spots AND you need a DH. Backup catcher is almost for sure your 10th spot and so now you have just two bench spots to backup your other 8 hitters.
The free agents made this a lot of fun. In addition to the guessing of how many teams would be picked and thus what order you'd pick your FAs in, there was also the interesting trap where whichever team was picked first in a year-group would be a worse pick than the next team, but obviously there's no way around that! 1911-12 was a good example of this. I had all 3 teams (Giants, A's, and Cubs) as draftable and relatively close in value. However, the top FA tandem was a monster Cobb (his $21m 1911) and Walter Johnson ($19m 1912). Runner-up wasn't too bad but still quite a drop off, with a $11.8m Tris Speaker and $15m Joe Wood. And whoever had to go first got a $10m Cobb and $10.9m Joe Wood. That was enough to where whoever was picked 1st, I had in the bottom half of the top-24 League 1 teams, whoever was picked 2nd would be right in the middle, but whoever was picked 3rd would be a top 5 pick!
to be continued...