I project players, assign SGPs based on historical averages, adjust for position, and then assign a dollar value.
The one issue is that this approach doesn’t exactly incorporate the playing time constraint of Ottoneu (i.e., 162 game limit per position). For example, I have both Jansen and Molina at catcher on one of my 5x5 teams. If I project both for 450 PA, then I can’t get the full surplus value of their expected production minus salary since I’ll have to play at least one of them less than 450 PA.
Now as it happens, I project both at similar levels of production, which if that’s the case then it really doesn’t matter which one I bench… but for sake of illustration, let’s say I think Molina is going to be a little better. So Molina’s value to me is the full projection in 450 PA, but Jansen’s projection has to be discounted to 250 PA (assume 700-450=250). There are some nonlinearities in valuation, but to keep it simple assume that it’s 44%. Now there’s a chance that Molina won’t get to the full 450, in which case I’m going to need Jansen for more than 250 PA, so I need to adjust the “useful PA” projection for Jansen up a bit. But the point is that I need to discount the value of my bench players.
So that’s why I’m willing to be bid higher on expected starters, but need to get my bench players for real cheap. That is, the expected value of the bench players is lower than a linear valuation based on a constant $/SGP because playing time needs to be adjusted. This is sometimes referred to as “stars and scrubs” valuation (i.e., paying a premium for stars/starters at expense of your bench).
I don’t know if that exactly answers your question, but for me the discount for bench players and how large a premium to bid for starters is the biggest challenge to valuing players for Ottonue (whereas it’s much more straight-forward in conventional fantasy games).