Page 1 of 1

STARTING PITCHER RATINGS PROJECT

Posted: April 17th, 2010, 8:04 pm
by Todd Zola
I'm looking for some input on a work-in-progress type project I want to embark upon. Last season, on the behest of mlb.com, I developed a means to rank the two-start pitchers for the weekly review article I wrote for them. I am not doing the piece this season, but would like to continue tweaking the process, in an effort to make it applicable to more than just two-start pitchers.

The basic idea is to come up with a series of indexes, then weight them in a manner to end up with a number which we can quantify the start potential. I would then like to come up with a means of determining a "game score" and grade the accuracy of the ranking.

Make sense?

I am completely open to any suggestions with respect to the game score. I know there are some versions out there. If there is one that does not use proprietary analytics, perhaps it is best not to re-invent the wheel. Please feel free to suggest means of coming up with a quantifiable game score.

Here are some of the elements I incorporated into the ranking. Please feel free to comment upon these, suggest others, suggest relative weights of these, etc.

K/9 index (100 x player K/9 / league average K/9)
BB/9 index (100 x league average BB/9 / player BB/9)
HR/9 index ((100 x league average HR/9 / player HR/9)

For these, I incorporated recent and year to date numbers. I started the procedure a few months into the season so the sample was reasonable. Starting now, I am leaning to using YTD and projected. Then once the YTD sample is larger, switching.

Park index HR
Park index runs
Park Index hits

These are by definition based on an average of 100, so it is just a matter of coming up with the proper weighting.

Opponent run index (100 x runs scored per game team / average runs scored per game)
I also did this for HR, BB and K.

These were not weighted a whole lot, but enough to help differentiate good from bad match-ups as it is my contention that the most important factor is the skills of the PITCHER.

I used recent and YTD for this as well, to account for "hot" and "cold" teams.

Home/away index. Studies show a player's performance is about 10% improved at home over the road. Last season, I simply used 110 for a home index and 90 for a road index and weighted them. What I am considering doing is actually adjusting the above if the numbers according to home/road by 5%. This is akin to how we adjust our neutral projection via a park effect, since half the games are at home and half on the road. By means of example, if a pitcher's K/9 is 6, if he was at home, the index calculation would utilize 6 x 1.05 or 6.3 while if the pitcher were away, it would be .95 x 6 = 5.7. A similar adjustment would me made for all numbers. If a team averages 4.5 runs a game, for home games, the number for the index would be 4.5 x 1.05 = 4.725.

So again, I welcome any comments. The idea is we can come up with a daily ranking of starting pitchers based on their skills and matchups, incorporating both recent and expected performance. Obviously, if a pitcher has multiple starts in the week, the individual scores can be summed in an effort to rank them overall.

Something I should mention is I purposely left out anything related to hits allowed by the pitcher as in such a small sample, it is more happenstance than skill.

Maybe it would be best to separate the ranking discussion from the game score discussion. Let's use this thread for the RANKING discussion.

And on that note, discuss...

Re: STARTING PITCHER RATINGS PROJECT

Posted: April 24th, 2010, 9:16 pm
by viper
any progress?

Re: STARTING PITCHER RATINGS PROJECT

Posted: April 25th, 2010, 1:39 pm
by Todd Zola
With the lack of interest from the peanut gallery, which is fine, I have been focusing on updating that silly little resume of mine as there are a couple of recent job openings posted that I am interested in.

Re: STARTING PITCHER RATINGS PROJECT

Posted: April 25th, 2010, 6:56 pm
by AllstonRockCity
i'm interested. i think this is a fantastic idea. at first glance, i don't know that i'd have anything to add or any comments on your initial outline, but most of my ideas come at the oddest times so i may still have some input yet.

but, i'd have to say that a resume update probably takes precedence, for you anyway. i hope to never update mine again.

Re: STARTING PITCHER RATINGS PROJECT

Posted: April 25th, 2010, 7:09 pm
by viper
I agree that the silly little resume is much more important