Player Families - Theory & Practice, 2

 ..........

Also, notice that hitters and pitchers move way up and down these lists, themselves, from year to year.  Here's what I mean...

James Shields had an 8.1 / 2.4 / 1.0 line in 2011.  Should I go make a comps list of pitchers who were 7.9-8.3 k, and 2.2-2.6 bb, and 0.9-1.1 hr?

Hold on:  Shields himself was nowhere near that in 2008-09.  Shields was 6.7 k, 1+ walks, and 1.1 homers.  Was Shields a terrible comp for himself?

When we "filter" a data table to match a player's single previous season, be aware that we're going to filter out most of that player's previous seasons.  And future seasons!  Set a filter to match Mike Carp's 2011 peripherals, and you're going to filter out Carp's own 2010 -- and his 2012.  :- )

For this reason, PECOTA takes 3-year sets, and James' Similarity Scores uses entire careers.  Only.  We can't match single seasons to produce player "families."

This is why Joe Mauer might show up on a single-season comps list for John Jaso.  Mauer might have a couple of years with .135 ISO's, but normal for him going forward will be an ISO of .200 or so.  Mauer isn't a punch-and-judy hitter.

.

THE UGLY:  It isn't feasible to produce player templates using only three criteria.

How about groundball/flyball ratio?  Is Ichiro, with a 2.5 grounder rate, comparable to Jimmy Rollins, who is a fly ball hitter?  Should we sort players who hit the ball in the air or on the ground?

How about speedy players vs. slow ones?  LH vs RH?  and so on...

.

=== Exhibit A, Dept. ===

A minor note on Carp himself, who was subjected to our invasive prostate exam all through the past season.  

Mike Carp was one family in the 1H 2011 and prior ... perhaps the Tino Martinez family, or something.  And then he morphed and joined a different family in the 2H 2011 ... maybe the Reggie Jackson family.

There was a demarcation line in Carp's evolution, a point at which he decided to approach the game differently.

.............

If you frappe-blend Carp's 2011 into a whole and then comp that, what have you got?  A turtle sundae, or something, but SSI's question is simpler.  To whom does the post-demotion Carp compare?

In the second half, Carp's ISO was a whopping .208 - in Safeco.  That's .230, .250 elsewhere.  His EYE was a scary 0.19.  You're talking about the ten worst EYE ratios in the game.  Them's the guys you comp him to.

...........

Who is in this category?  Mark Trumbo, Alfonso Soriano, Adam Jones, Nelson Cruz, and Mike Morse.  Of those players, Trumbo's the lefty, of course, and also the player who just transitioned into MLB, like Carp did.

We don't say that those five players are Carp's comps (that's got to be on somebody's bowling shirt somewhere).  But if you use the Seattle 2011 system, you want to isolate to what Carp's post-callup profile suggests.

And those five guys.  That's not a bad little list for starters, do y'think?

.

.

 

Comments

1

Not to quibble, but Mauer has 1(full) season in which he ISO'd over .160, in every other (full)season he's played, his ISO has been between .081 and .159, James has his next season prediction for an ISO of .140, .140 isn't drastically better than .110 for me

2

Fair point b'wana.
Yeah, you run over and check Mauer's stats, and he's slugging 'only' .471 lifetime with 15 homers per 162.
He's a big huge guy, seems to swing hard, but all of his production is inside the park ...
Hm.  The BABIP is .341 lifetime, so he must be hitting line drives and sizzling ground balls? ... yep, sure enough, he's an extreme groundball hitter.
...............
Of course this is a radically-different type of hitter from Casey Kotchman or John Jaso.  
This underlines the fact that you can't forecast a group of hitters together, based on only on their K, BB, and ISO similarities.
A family of "Great LH Hitters who Smoke the Ball On The Ground" is going to forecast differently from those hitters who have .140 ISO's because they don't swing hard.
I wonder which other huge HOF lefty batters had 2:1 groundball ratios.
 

Add comment

Filtered HTML

  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <blockquote> <code> <ul> <ol> <li> <dl> <dt> <dd><p><br>
  • Lines and paragraphs break automatically.

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.

shout_filter

  • Allowed HTML tags: <a> <em> <strong> <cite> <blockquote> <code> <ul> <ol> <li> <dl> <dt> <dd>
  • Lines and paragraphs break automatically.
  • Web page addresses and e-mail addresses turn into links automatically.