The only things that I consider absolute fixed values are training-20 and ADV-0. Training is obviously super good, universally respected, ADV is almost universally dismissed by top vets. Allegedly WIS "changed" Advanced scout last year to make it important, but I can't even tell what they changed, it still has absolutely no importance or value. Even if you are looking to rebuild and trade some vets for prospects, post a guy in WC to see which teams are interested, and if you've played the game for 10+ or even 5+ seasons, you will have a pretty good feel for which guys to pluck from their depth chart and which guys to leave alone. You don't need ADV to clarify whether a guy is exactly 78 potential or 75 potential, it doesn't matter. [I could imagine an environment where ADV would have value, but it would require at least 5x the amount of DITRs, it would require quad-A boosts, etc etc etc so that teams' depth charts were overflowing with ML-quality guys beyond their 40-man roster limitations for a more comprehensive Rule 5 draft ... but those changes are simply not anywhere near the horizon]
I personally prefer 20 medical because I like roiding my guys up, but it's definitely not for everybody and you also don't need it right away when you're rebuilding (because your ML guys suck and it doesn't matter if they get hurt, right?). Medical is a huge long-term commitment and it's not necessarily productive even when you are in the 15-18 range. I would recommend to either commit to maxxing to 20 or commit to flooring to 0, but anything in the middle is kind of a waste. A lot of good owners do 0 medical and there's merit to the savings, that's literally an entire extra max-contract free agent in the long-run
At the very beginning after rollover, the first thing I calculate is Exactly how much I need for coaches. It's important to forecast. Most importantly, check if your FI is re-signing, that guy is super important and we always want him to re-sign for peanuts. Also, see if your AAA bench coach is worth promoting to the majors because he will only make 600k instead of 2-3 million, that can be a good way to save money. On the other hand, Top-level ML PC and ML HC seem to be worthwhile things to invest cash in. Those are different than bench coach because if you promote your AAA HC/PC those guys still want 1.5 or 2 millionish to start, so you're not saving as much money on those guys.
For payroll budgets and scouting, I have no set strategy besides being as flexible as possible. It's fun to do Moneyball and pay everybody nothing and trade them as soon as they want money, but if I'm contending, I accept that veteran players cost money. After 4 or 5 ML seasons, top draft picks cost big money on their arb deals / long-term deals as well. On the other hand, rebuilding projects require patience, discipline, and restraint when it comes to staying away from stupid long-term ML player contracts, etc. You have to be realistic about which type of team you are and spend or save depending on the circumstances of your roster and upcoming changes.
When it comes to draft budgets, it sounds like I am in the minority here but I prefer spending 20 budget units as 10-10 rather than 20-0 or 0-20. My primary reasoning relates to my preference of letting go of 100% of my type B free agents (you can just go to the list and sign one of several equivalently-rated players for the exact same money, so why not take the free prospect ???), and most of the fringe type-As also. On my SF team this season, None of the guys would re-sign so we picked something like 11 times in the top 100. Those comp picks are super useful for filling out minimum-salary bullpens and benches in the future, as well as grabbing an occasional starting position-player or even an all-star here or there. Additionally and arguably moreso, those comp pick prospects are super-useful for trading to rebuilding teams for solid veterans. It's so much easier to acquire players for prospects rather than prospects for players. It's usually a buyer's market... there's always going to be 8 last-place teams, but there's rarely more than a handful of buyers at any one time, right? Dial up the last place teams, offer a little 2-for-1 action, maybe some 3-for-1 if they're trading me cash or if I happen to like them
Also, you give yourself more budget flexibility across seasons if you split up your draft money. Let's say one season I have no picks (losing no free agents, signing a bunch) and in the next 1-2 seasons I know that I will be losing several, I can go +/- 8 units overall per season instead of +/- 4. If you go 20-0 then you can only go downwards to 16-0 or upwards to 20-4, but if I go 10-10 I can go upwards to 14-14 or downwards to 6-6, depending on my needs. Plus, on the Budget Analysis screen you can see the league averages but you can't see who spends what anymore, so you don't know exactly how the teams picking around you budgeted. If you're pulling from both lists instead of just one, there's a chance that good players are slipping to you from one of the piles whenever all the teams in front of you happened to have budgeted for the same category.
Lastly, IFA is either all-in or all-out. They let you budget up to 20 on prospect but you will typically only need to spend about $5M-$8M on the draft. However, the $12M-15M leftover cash alone is almost never going to net you one of the elite IFAs, so therefore you either need to either hoarde all of your extra player/coach payroll to halve into $5M-$10M extra IFA, or just stay at 5-8 and spend the would-be prospect money for players. For IFA itself, I tend to hover around 10 because that gives me flexibility to polarize my prospect range depending on contending/rebuilding. If I know I am going to have a budget surplus because ML players are leaving, I can go IFA 14 (plenty enough) with prospect of 20 plus whatever extra payroll I chop. If I know I need to budget extra cash for a new fielding coach or ML contract renewals or big free agent signings, I will go IFA 6 with a prospect of 6. IFA is a huge investment without guarantee of success, it can be rewarding but it is the most risky. [In my opinion, IFA is not as good as it used to be ever since they patched up the HTML glitch where ppl could review the entire batch of prospects one-by-one after rollover.] For me, IFA is the "there's nothing else to spend it on" category and is completely impractical when I'm contending