In my opinion that’s kind of a big IF. I bet you or I could setup reasonable starter ratings but I definitely wouldn’t expect that from my current league director. If the teams on this thread had bad starter ratings then it kind of serves as an example of my point. I’m not saying it can’t work but I do feel my statement would be a reasonable strategy.
[...]
I don't know what "if" you are referring to.
I think I need to explain my point more clearly. Let us assume you have a league and that you want to handicap the play. That's a given.
If you have well established Fargo Ratings, then all is good.
But suppose you have the opposite extreme--nobody had any games at all in FargoRate. What is the best thing to do?
I contend the best thing to do is to convert ANY knowledge you have about how people play from ANY mechanism as well as you can into Fargo-Rating Guesses. Import those, and let the system take over.
If you have no knowledge at all of how anybody plays, give everybody the same starter guess and let things evolve from there.
If you have SOME knowledge, say APA numbers or local A,B, C ratings or averages from previous league seasons, then the best thing you can do is convert those to starter guess and let the system take over.
Why is this?
YOU may think John needs two games on the wire to five from Bill and John can give one game on the wire to 5 to Mary and so forth. By converting these estimates to rating differences, the system now knows reasonable--as reasonable as your guessed input--matchups between any of the people, like between Mary and Bill. That is, the system reflects your judgment for situations you have actually considered, and it extends your judgment in a consistent way to situations you haven't considered. There is internal consistency to everything. And the system can estimate points if you use point scoring. Further, this approach is self correcting. new data moves ratings in the right direction.
A key is you lose nothing by converting your judgment to effective Fargo Ratings and you gain a lot.