THE BOOK--Playing The Percentages In Baseball

Shouldn’t players who spent a significant amount of time on the DL be omitted from the study to correct for a selection bias.

I can see pitchers who spent time on the DL last season coming back into 09 throwing harder only because their injury has been corrected.  This could work in the opposite way as well, a pitcher like Matsuzaka lands on the DL after a couple starts where he struggles to hit 90.

I think this would have more value with injuries removed, as then going forward you could point to a drop in FB velocity and have somewhat of an either or proposition (hurt or going to do poorly).

This could be very interesting stuff, seen as when a pitcher drops their arm angle they get more movement at the sacrifice of sheer speed.

Interesting stuff.  I’d say I can’t wait to see the end results, but that would mean the season is over.

I should also point out that I disagree with the way we are calculating an “average fastball speed”, just as much as I’d argue about “average HR distance”.

Suppose you have these HR hit by Pujols:
430 ft
420
410
390
390
380
380

That’s an average of 400 ft.

Suppose you have Inge hit these HR:
410
405
400

That’s an average of 405 HR.

Do we really want to say that Inge’s “average” HR is longer?  Suppose he then gets a squeaker at 370 feet?  Now his average is lower?

This is the same thing here.  What if a guy intentionally throws his fastball slower on occasion?  Instead of us classifying it as a “slow fastball”, it counts as fastball.

So, my proposal is to take the 25% fastest pitches each game, and classify those as “regular fastest pitches”.

Steve,

That is a very good suggestion.  I did try and take out all guys switching between the bullpen and rotation (and vice veras), but I didn’t think of omitting players coming off of DL stints.  Do you, or anyone else for that matter, know of a DL-database?

Tango,

Very good point, maybe I will rerun the 08 numbers with that definition of ‘average fastball.’

Dave,

The only way I can figure to calculate a DL stint or extended rest would be to use the dates of their appearances.  I have no idea how much work it would be to through each pitchers appearance date into a database and calculate the days between appearances.

This may be something else good to look at for before and after trips to the DL and skipped starts.

You can easily use Retrosheet to create a database of a pitchers appearance dates, whether in a relief or starting role, how many batters he faced, how many pitches he threw, how many days rest between appearances, etc.  This can be cross referenced to MLBs transaction file (found under Players) which gives all DL assignments with a brief statement of the injury on a daily basis.  They go back to 2000.  You can also check the MLB official team sites which often have news articles on a player giving more detailed information on the cause and extent of a player’s injury.  I think I remember correctly that someone has compiled all this information into a database, but I am not sure how publicly available it is.

I tried to post this on the site, but it wouldn’t let me:

Great stuff.  I like the idea of separating or least “marking” those pitchers who spent time on the DL or otherwise had some serious injury such as surgery in the off-season.

I would also REALLY like to know the average difference for all pitchers between first 100 pitches of the season and the whole season. I would guess that it is less. If it is, you want to correct for this before you do your comparisons.  It won’t affect the data plot of course, but it definitely affects our “thinking.” For example, if all pitchers are .5 mph less in the first 100 pitches than for the whole year or rest of the year, I don’t want to think of a guy who is .6 less in the first 100 pitches as potentially having a problem…

I have DL databases.  If you need one, Steve, email me…

MGL,

The average was -0.86 mph.  I should have included that.

It would be interesting to do an analysis with the injury data--seeing how pitchers perform just before going on the DL or just after they come off.

Dave send me an email and I’ll send you a DL database from last year…

I must be completely dense because I cannot find your email address anyway.  Mine is dnallen -at- baseballanalysts -dot- com.

Thanks

Dave, I once spent almost a half hour trying to find it one time, you’re not alone. Go to one of MGL’s blog posts and click on his name at the top of the post. If you do this for Tango on this post, it brings you to tangotiger.net, but if you do it on one of MGL’s posts it will show you his email.

The problem with avg HR distance is it only measures HR.

It looks like K/9 are up so far.