I am looking to create a different Ranking System for next year. I already have an excellent "Draft Sheet System" That I would even be william to share. I am more than willing to share what I have used in the past but I don't think I can get it off that hard drive. (dead) What I used in the past has always served me well, but I am interested in opening up a few "formulas" to see if "We" all can come up with something better.

Hootie's posting of the ballparks rankings got me thinking. They are Ranking Ballparks on a 100 point scale using the same categories that we all play with. Why not encorporate the same 100 point scale to players? 100 being exactly league average. But then where do you top out at 150 points/ 50 points .. just some thoughts ... any other numbers guys here in the forum??

Long ago I had a spreadsheet that rated players against a league average. For each category, the average would count as 1.00. Then I'd rate each player as a ratio against that. (You could use 100 instead - it's probably preferable as you can avoid decimal points). (I'd happily send this to you, except that I have no idea where it is. It's probably on some floppy disk that's sitting in the bottom of a dusty box marked 'college stuff.')

I used to rate each player by position. So let's say the average 1st basemen hit 30 HRs this year (I have no idea how accurate that is). Then someone like Derek Lee, who hit 31 HRs this year, would score 103 points in the HR category. (31/30 * 100)

If you are playing 5x5, then the average player at each position would score 500 points.

Alternatively, you can also set it up so that the league leader in each category for that position scores 100. Then you can see how each player at a position compares to the overall number leader at that position in each category. In that case, the ideal player at each position scores 500 points.

In that case, the league leader in 1B hrs this year was Thome with 47. So Lee would score 66 points in the HR category. (31/47 * 100)

Advantages to setting up a spreadsheet like this is that you can quickly sort within each position for a particular category. So let's say that you already had a ton of HRs drafted, but needed steals, you can simply click on the SB category and sort accordingly. Because you are using a ratio measure instead of hard numbers, you get some sense of how far above (or below) the mean a player is. Another advantage is that you can see where the major dropoffs in talent is at each position. This can be a huge help in draft day strategy, as you will have a good sense of whether there are still quality guys left in each position.

Disadvantages are that it takes a long time to enter the data, and it's not clear it's completely worthwhile. It's not that hard to find player statistics online that can be downloaded, and I'm not sure how much of an advantage there is to looking at ratios as opposed to looking at the actual number of, say, stolen bases. Another disadvantage is that it is tough to compare the value of players across positions. Is an OF who scores 600 total points worth more than a SS who scores 550?

If you wanted to post some of the formulas you use, I'd be happy to take a look and comment on them. I'd love to develop a new draft day formula. I've started studying a bit of probability theory over the last few years, and have been tempted to apply it to fantasy baseball. I envision some Bayesian program to help me predict player performance. Unfortunately, I simply haven't really had the time (um, in part because I'm not all that great at math and it takes a lot of time to learn something as hard as probability theory). Maybe I can get an internship with some sabrmatic folks!

Well I have used many, many rating systems. This is what I am thinking with this..

100= League Average
150= Best
50 = League Worst

This system can be applied to as many categories as possible.
Runs, Obp, k/9, you name it. Just ensure that what ever categories you run, you divide the total by the same number of categories. So Instaed of having 500 total points for a league average, 100 would be the "average" if he was a League average 5 category guy.

I am thinking of this system becuase it would allow me to use strength of ballpark indeices and I wcould simply incorporate strength of schedule by Month. I personally try to look at Monthly breakdowns becuase I spend alot more my time trying to project what players "will" do vs. what they have already done.

As far as typing formulas, well once the formula is figured out, Then that is where Excell comes into play. I usually use about three or four different "Projection" services for raw numbers. Again I just run an average of the three. Lasty I always run a "Real team" test. Line up the players from the same team and make sure that the numbers appear real. Nothing like a team scoring 1000 Runs but having 1500 Rbis

I guess where the real problem comes in is trying to figure what value to assign to Ballpark and Strength of schedule.

I agree that running numbers in this way, always shows me glaring weekness in position scarcity and stat scarcity (ie SB's and SV's) and enable me to adjust my rankings and figure in infation where necassary.

also, the impossible thing is that you have to count on the age and injury factors, which are impossible to measure (except, of course, for J.D Drew, who stubs his toe and sits the season.)

DK wrote:also, the impossible thing is that you have to count on the age and injury factors, which are impossible to measure (except, of course, for J.D Drew, who stubs his toe and sits the season.)

Actually, Usuing this scale. you could factor in injury "History". Generally its is covered by AB's. but useing a career average or a three year everage for time missed due to injury could be factored in too. Of course 60% of injuries are impossible to measure, so very little weight should be given overall. that is why AB's are the best.

wrveres wrote:Well I have used many, many rating systems. This is what I am thinking with this..

100= League Average 150= Best 50 = League Worst

This system can be applied to as many categories as possible. Runs, Obp, k/9, you name it. Just ensure that what ever categories you run, you divide the total by the same number of categories. So Instaed of having 500 total points for a league average, 100 would be the "average" if he was a League average 5 category guy.

The trouble with this form of scaling is that not all categories are equal. In stolen bases, for instance, the gap between 'best' and 'average' is a lot greater than it is in, say, runs scored. By assigning a value of 150 to all category leaders, you'd be undervaluing speedsters like Juan Pierre.

DK wrote:also, the impossible thing is that you have to count on the age and injury factors, which are impossible to measure (except, of course, for J.D Drew, who stubs his toe and sits the season.)

It actually wouldn't be all that hard to factor this in. I think wrveres is right that ABs is usually a good way to factor in injuries (especially if you use 3-year averages or totals). There are a number of ways to factor in these numbers.

As for age, I've been doing some thinking about this recently. At least according to some, batters tend to have their best years at age 27, whereas pitchers have their best year at age 31 (I'm a bit skeptical of this last one, but it doesn't matter too much). One way to factor in age is to factor in an expected improvement or decline in numbers based on age. Have an age category where for batters ages <28 factors in a 10% improvement, 28-32 no change, 32-37 a 5% decline, and >37 a 10% decline. Something like that.

Arlo, I think you may be right about the downplaying of big differences, but I'd have to actually work through the numbers. It might turn out that Pierre's value still comes through.

ramble2 wrote:At least according to some, batters tend to have their best years at age 27, whereas pitchers have their best year at age 31

i gotta say, i disagree-sort of. i think that although batters are in their physical prime at 26-28 years old, the prime years are really 28-33 because that's when they have the highest combination of skill and experience.

for that 5%, i suppose it could work for some, but injuries is too unpredictable. you can use history for it, but remember everyone has to get injured a first time.

ramble2 wrote:At least according to some, batters tend to have their best years at age 27, whereas pitchers have their best year at age 31

i gotta say, i disagree-sort of. i think that although batters are in their physical prime at 26-28 years old, the prime years are really 28-33 because that's when they have the highest combination of skill and experience.

for that 5%, i suppose it could work for some, but injuries is too unpredictable. you can use history for it, but remember everyone has to get injured a first time.

Good points. I was actually just citing the age ranges I did for examples sake. I'm not sure what kind of age ranges to use, or what kind of expected decline/improvement to factor in for a given age range. To figure that out correctly, someone would have to comb through the data to see what happens to the average player as they age.

But even using SOME kind of age related adjustment - even if it is simply a rough guesstimate - might still prove to be useful in predicting player performance.

ramble2 wrote:At least according to some, batters tend to have their best years at age 27, whereas pitchers have their best year at age 31

i gotta say, i disagree-sort of. i think that although batters are in their physical prime at 26-28 years old, the prime years are really 28-33 because that's when they have the highest combination of skill and experience.

for that 5%, i suppose it could work for some, but injuries is too unpredictable. you can use history for it, but remember everyone has to get injured a first time.

Good points. I was actually just citing the age ranges I did for examples sake. I'm not sure what kind of age ranges to use, or what kind of expected decline/improvement to factor in for a given age range. To figure that out correctly, someone would have to comb through the data to see what happens to the average player as they age.

But even using SOME kind of age related adjustment - even if it is simply a rough guesstimate - might still prove to be useful in predicting player performance.

yeah, it could certainly be useful, it's just i wouldn't make it my first research when making picks