I have a spreadsheet of numerous projections for the 2007 season and made a composite sheet for all of the projections. Now I want to be able to provide a point system to the players so I can rank them based on these projections.
What type of formula should I use to arrive at this point total?
I pull out all the players I expect to make a team and then take the standard deviation of each stat, then I seperate them by positions and take the standard deviation of their points based on position and add half of that to other point totals. Then I slightly modify the value of a few stats that I don't care about as much like ERA and AVG, they just aren't reliable enough to count fully.
Thats probably all too complicated and time consuming to do but I have a lot of free time at work so it works for me.
I do pretty much the same thing. I take the category and take the leader and subtract the player who's being compared. Just like adding fractions you'll have to come up with an amount in every category to multiply by to get every category as close to equal as possible. You have to do that with every position and your overall rankings. It's really simple to do with Excel and using functions. I like using the leader rather than the average because it doesn't change much and gives a better idea of how much better the elite players are from the almost elite. But, it doesn't change much at all.
I do essentially the same thing as the previous posts. But I do mine a little differently. I take each scoring category that we use and allow the top player in that category 100 points and I allow every other player so many points based on a percentage of the maximum 100 points. I can total the points for each player's ranking. Or I can keep them separate so as to target a player for a particular category.
Check the values off to the right near the bottom of the page. I plugged them in a spreadsheet with my projections and they gave me darn near what I would have expected to see. Only for 5x5. For other formats you're going to have to run std deviations.
First I convert all of the ratio stats like era, whip, ave, k/9 etc... to counting stats. Then I use a replacement level value algorithm to give each player a value. That way you take position scarcity into account.