Setting Category Targets

[RandyKutcherHair]

What is the best way to set targets for scoring categories? I am playing in a 8 team AL Only roto 7x7 league where the top 3 finish in the money. Using the last two years of results should I average the top 3 finishes in each scoring category as my target to reach in '09 or should I take the average of all 8 finishes in each category? I know there is a method; but I am having a massive brain fart.

[Todd Zola]

I'd see how many points it takes to win the league, then divide by 14 (2 x 7) and see how many points per category, on average, you need. For an 8-team league, it will probably be between second and third.

Personally, I think there is another step, and that is to adjust (normalize?) this to the projection set you are using. If the number of HR in your league last year was 2500, and 300 was the target, and the projection set you are using has 3000 HR in the draft-worthy pool, I think you need to adjust the target accordingly, in this case upward.

300/2500 x 3000 = 360

These numbers are for illustrative purposes only.

[viper]

I think the question to really to determine what sort of target you are shooting for. Looking at old league results is great but is it really a worthwhile task unless you are going to some type of detailed analysis on it. I use it for regression in my SGP stuff but that's really about it. Multiple years gives me multiple sets of SGP to average or whatever mathematical manipulations I desire.

May I suggest you look at your targets differently. First, get your projections and check off the players you would expect to be drafted. Use ADP or whatever method you want but see who they should be. Assume it is a 12 team, AL-only league with 2 catchers. That makes 168 hitters and 126 pitchers. Let's look just at offense. Find those 168 top guys and sum the stats for each category using your projections. It's more work for batting average but it can be done. Making the math easy on me, assume you have 2788 HRs. That 229 per team. You would want some number over this average. See what it takes to get a '10' from prior season and see how much above that season's average it was. Don't look at your projections yet as they have nothing to do with this amount above average. If it takes +20 over average, then your target should be 229 + 20 = 249 or 250 for rounding purposes. Just do this lots of times and you have your targets. An Excel program can do a lot of the work.

Each season plays out differently but margins tend to be relatively consistent. The actual season has injuries, replacement players and hundreds of things you can't control which is why all these projections can fall apart during the actual play. Still, they are all we have so we must use them. We just need to know how to use projections coupled with league history.

One man's opinion.

[Todd Zola]

Not for nothing, but we both just said the same thing

[JP Kastner]

I think it should be worth noting that if you don't have enough information and time, looking at last year is better than nothing. I was sitting at lunch with Lawr, Todd and a bunch of guy from MLB.com just before Tout Wars. While we were there Dean Peterson of Stats LLC called to say that he was going to be late for the AL Tout Wars auction. His plane was delayed due to bad weather.

As I was the only guy who wasn't a commissioner or in AL Tout, I was handed the phone. By the time we were finished with lunch, I had a whole thirty minutes to prepare. I simply used the previous year's results.

The window of error is larger, but it is better than nothing.