jdavidbakr wrote:
Right now the requests are being built based on a linear regression - so the outliers don't have as much strength until there are several pulling the line up. There is also a curve toward the top of the line for players as they approach a 100 rating to take them up above the average.
There are a couple of reasons I haven't added logic to make the top players request for themselves to be the top paid players - first, there is the question of what classifies a player as a top player - should it be the top 32 players in the league at each position always vying for the best contract? The top 20? 10? The linear regression formula does cause the top players to compete for better contracts but you are right in that they don't immediately try to get the best contract in the league. The other reason I'm hesitant is because in the NFL you have a very strong understanding of the impact of your player contracts ... if someone breaks the bank on their QB and every other team is required to match that offer for their QBs, that one player who might have ruined their team for the next 2-3 seasons has by extension ruined every other team for the same time period, unless you happen to have locked up your QB for longer than his QB will hold that insane contract. Expand that to several teams who have locked up different positions at enormous contracts and no one will be able to sign more than a few top-tier players, requiring the rest to go to free agency. The linear regression model prevents that kind of activity from ruining everyone else's team.
I thought about that too, but the scenario I see is this......
Say a team ruins their finances by overpaying a QB...... All the free agent QBs who are in the last year of their contract will ask for a larger sum. If the current owner can't, or is unwilling to match it, that player goes to free agency. If, in free agency, no other owner is willing to meet that players demand, then he just signs, say on week 5, with the highest bidder. No one is screwed, per se. We just will be paying what the "fair" price is for a QB at whatever caliber. The free agent market will decide. You can use the same formula or algorithm that is currently used for the internal evaluation of player value.
I don't know how you have it set up now JDB, but it sounds like the system uses an average salary as a baseline, then adjusts what the player requests based on an internal evaluation of value. Just my opinion, but I think a better system would be to run the regression on the top salaries per value point. Get your top salaries, per whatever the internal value formula says the players rating is, per whatever position that player plays, smooth that line out so it isn't a saw tooth, then go from there.
Not trying to come off as overly critical or anything, just giving suggestions.
