Add new comment

Dependent and Interdependent Variables

 ..................

Remember that Pixar short in which the little alien had a bank of 9,000 toggle switches and was trying to use them to lift a human into their ship?  Things got worse and worse, and at the last possible second, the big mentor alien saved the day (night) by poker-facedly luuuuunnnnnngging wayyyyyyyy over to toggle switch #3,217 ....

Dr. Grumpy hit a random toggle switch (or what do you call those switches?!) that provoked Dr. D into a post.  :- )  Whether this result makes Dr. G the idiot alien or the mentor, you be the judge....

..........

In the Hultzen thread, there was a ton of think-tank chat on the effectiveness of free agent bats.  Eventually we got a bit off the rails, posting mini-lists of teams that did well with and without free agent bats, which was fine, and then arguing that these mini-lists were persuasive, which maybe was not as fine.

Dr. G finally reacted:

There are just so many variables. It would be a tall order to create any kind of study. You might have to look at subgroups. For example, it might be that adding a superstar hitter to a team that already has some serious pitching talent (eg current Ms) is much more beneficial than it would be for a team in another situation (eg Arod to the Rangers).

It is hard enough to isolate even one variable, much less to identify which other variables affect the independent one.

Sabermetricians prove that all plants in America, as a group, grew faster when composted.  Voila!  We understand horticulture.

Then you ask, should I compost my cucumbers - will they taste different?  What if my soil already had a lot of salt in it; could my particular garden get over-salted if I add the sodium in the compost?   Does compost work well if I get very little, or a lot of, sunshine?  How about if I have too little salt AND too much sun AND I then compost?  ... 

It is at about this point that Dr. D usually gives up.  These variables spiral out of control so fast that, if we are going to tackle the problem at all, we have to retreat to "fuzzy logic" and (data-nourished) intuition.

SSI does not minimize this data nourishment.  Matthew, at Lookout Landing, noticed that Jaso's "kill rate" on stolen bases has been terrible.  But he also noticed that the league runs on Jaso less than average ... 

Dr. D's intuition is that, if you keep Jaso away from the speediest teams, the throwing arm won't be a big issue.  Thanks to Matthew, the intuition is data-nourished.

The arm probably isn't an issue.  The handling of pitchers might be.  That topic, and that topic precisely, stands to make-or-break John Jaso as a PYCWYNPW.

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

:- ) My #1 complaint with sabermetrics is:  too little respect for the complexity of the problem.  Guys represent themselves as scientists and obviously haven't begun to think about the concept of dependent variables, much less to capture them before we claim that the Truth is published, and our best friend who edits the JAMA has peer-reviewed it and signed off ...

No offense meant.  An offseason plan that is based on piling up X amount of last year's OPS, or WAR, or RC,  is a perfect example of this.  A little table with three columns -- player, WAR, and salary -- is helpful as we begin to organize our thinking.  It is only the point at which we represent this as a Periodic Table, based on Laws of Science, rather than a first hypothesis, that we would do ourselves more harm than good.

.

=== John Jaso ===

The main question on my mind is:  could there be some fatal flaw in Jaso's catching that creates a Chris Gimenez situation.  

The 2011 Mariners lost the first seven (7) games that Gimenez started at catcher.  For the year, they won 4 games that he started and lost 16, which equates to a 131-loss baseball season.  The 1899 Cleveland Spyders, who intended to be a sideshow, only lost 134.  ;- )

No sabermetric method whatever was going to capture Gimenez' effect on a pitching staff.  The variables were far too complex to make a forecast based on math.  In May 2011, SSI attempted to capture it intuitively, in this article and in this one.

If Jaso is not cycling in a Chris Gimenez syndrome, and if the bat rebounds, then POOF you grab a championship-level platoon player out of thin air.  Is he, though, and will it?  Jack Zduriencik thinks not.  On this particular subject, you are not going to get closer to the Truth than by asking Jack his opinion.

Which won't keep us from a blog hemorrhage on it,

Jeff

Interest categories: 
Interest locations: 

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.