Welcome. I too am very new to the game but have had some of the very same questions you have and learned quite a bit already so happy to share what I know. However, I am VERY FAR from having a strong understanding of the game so take all this as one noob to another.
1). I’m just beginning to create a plan for my integration of baseball in a replay I’m planning to begin in 1901 season. What I have learned is that if you want Josh Gibson to be JOSH GIBSON you will need to manually alter his ratings each and every year unless you allow the development engine to take over your game. But note, the development model will NOT guarantee that Gibson becomes great (more on that in the answer to your #2 question).
For any of the players in Negro Leagues you need to basically understand how the game adjusts players from minor leagues to major leagues. Currently, all NLB leagues (sans MLB’s recent decision to list NLB as a major league so this may change in say OOTP21 or 22 or 23
) are minor leagues in OOTP and so any player manually added to the majors will come in with their ratings based on MINOR LEAGUE STATS. Which of course means that Josh Gibson’s great stats are merely minor league level and so his ratings in the major will be decreased.
It is possible to “roll your own” database but that’s WAY beyond me so can’t even explain how/what is needed for that.
2). I’d suggest learning more about recalc and the development engine from the pros here, but IMO the recalc is essentially an “average smoothing” engine while the development model completely takes over control of the player’s career and may produce careers which far exceed or far reduce a player’s actual stats they had over their actual careers.
For recalc, and from my (very limited) toying with OOTP, the engine with various adjustments to its algorithm, will take someone’s career and smooth out any highs and lows by adjusting the ratings (which in turn govern what stats the player will ultimately have). So for example, lets take Reggie Jackson at the start of his career. By using the recalc with settings at 3 or 5 year recalc (as opposed to 1 year), Mr October would not see his ratings go down as much from 1969 to 1970 and then go up again in 1971. Of course, OOTP is a simulation and as such will come up with quite a bit of variability in terms of real stats and game produced stats. That being said, Reggie’s recalc’d ratings (on average) should not produce the big drop in production he had in 1970 and instead have 1969, 1970, 1971 more in line with each other (IF using the 3 or 5 year recalc). Using just 1 year recalc will lead to the highs and lows but will still have the variablility becasue again OOTP is a simulation and so player's stats can (and will) be off their actuals. I come from decades of playing APBA and while OOTP IMO is a FAR better game overall, it is also FAR away from the accuracy that APBA had.
A further caveat, there are gazillions of settings in OOTP which can effect the game produced stats. So don’t look at the recalc options as the only thing affecting player's stats produced by the game.
Back to development engine, I suggest reading more from the old timers here that have run multiple (and very long based) historical replays. But from what I’ve learned, the development engine can (and often does) create wildly different outcomes to player’s careers. How wild? Again, look at guys who post their results from their long term replays and you’ll get a sense.
3). The draft can be turned on or off and I believe can be turned off until, as you wish, a later period and then turned on. So I believe the answer is Yes and you can have players show up as FA first and then create a draft some time later in the 60's.
4). Although I haven’t tried it, I do believe the answer is YES. You can incorporate ANY player from ANY league into the majors (all the same issues from #1 & #2 of course will apply).
Hope this helps