Strikeouts and Walks

Strikeouts and Walks

Some time ago, actually a long time ago, I posted a
brilliant pontification - to wit: free ***s walk
a lot because they swing and miss a lot. That gives
the pitcher more chances to walk them. Some people
emailed me congratulations on my perspicacity. Roger
Lustig, like a damn fool, disagreed. The strength of
my logic was such that I saw no real need to look at
any data. Aristotle would have been proud of me. I
have a fiduciary interest in data however, and eventually
I did feel obligated to do something with the numbers.
I took the 1988 and 1989 AL and NL batting stats and
computed strikeout and walk frequencies.
Kfreq=K's/(AB+BB)
I don't have HBP or sacrifice data. BBfreq is defined
analogously. I combined the data for both leagues across
both years (a player can then appear twice) and correlated
the two frequencies (I only kept players with 400+
appearances. If I'm right we get a nice positive
correlation.  We (I) got something like .017 which isn't
even significant (statistically). If I ever get around to
adding the 1990 data I imagine the result will become
significant but there is no reason to believe that the
magnitude of the correlation will change substantially and
that kind of correlation isn't worth a dipity damn.
I do not give up easily, especially when I am behind, so
I decided to try another angle. Roughly, I wanted to put
in the pitcher's fear factor. Now they don't keep a stat on
that so I had to invent one or at least rename one. How
about HRfreq? You can guess the definition. Slugging
percentage is reasonable too. The fuzzy idea here is that
you've gotta be nuts to walk Felix Fermin when about the
worst thing he can do to you is hit a single and walking
Mark McGwire is not nearly the worst thing that can happen
to a you. Anyway, I did some multiple regressions and
certain graphical things to look at the K/BB frequencies
after factoring other things out. Guess what? The association
between K's and BB's is NEGATIVE. I tried it with all kinds
and combinations of regressors, looked at all kinds of
collinearity diagnositics and generally tried everything. The
relationship between K's and BB's is consistantly negative
and usually significantly so. Now please understand, we
aren't talking about real tight relationships here, but still
they were negative. A posteriori, I can think of some good
reasons (so maybe I should be financial advisor) BUT ...
Gerry

Strikeouts and Walks

The following is NOT a flame, please insert smileys if you think I was being
mean.  I just read the thing, was really interested in the subject, and got
really confused reading a lot of statistical
mumbo-jumbo-that-I-didn't-understand.

Quote:
>Some time ago, actually a long time ago, I posted a
>brilliant pontification - to wit: free ***s walk
>a lot because they swing and miss a lot. That gives

I remember that, and it sounded reasonable to me at the time.

Quote:
> [bunch of words I vaguely remember from my statistical/probabilities class
>  two years ago are clumped with some baseball words]
>correlation.  We (I) got something like .017 which isn't
>even significant (statistically). If I ever get around to

Translated to English: There is no relationship.  Seems odd...

Quote:
>  I do not give up easily, especially when I am behind, so
>I decided to try another angle. Roughly, I wanted to put
>in the pitcher's fear factor. Now they don't keep a stat on
>that so I had to invent one or at least rename one. How
>about HRfreq? You can guess the definition. Slugging
>percentage is reasonable too. The fuzzy idea here is that

I like HRfreq, I think it pretty accurately describes the fear.  The only
problem is someone like Wade Boggs, who is walked by frightened pitchers all
the time.  He is really only an anomaly, I am not sure there are many other
singles hitters whom pitchers fear.

Quote:
>after factoring other things out. Guess what? The association
>between K's and BB's is NEGATIVE. I tried it with all kinds
>and combinations of regressors, looked at all kinds of
>collinearity diagnositics and generally tried everything. The
>relationship between K's and BB's is consistantly negative
>and usually significantly so. Now please understand, we
>aren't talking about real tight relationships here, but still
>they were negative. A posteriori, I can think of some good
>reasons (so maybe I should be financial advisor) BUT ...
>                     Gerry

So what does that mean?  That hitters who strikeout a lot walk LESS, given that
we have factored out all the times they have been walked by scared pitchers.
Hmm, that sounds bizarre to me, I always thought as you did in your initial
reaction to the idea.  But then again, all those players who strikeout and walk
alot are mostly power hitters, so a lot of their walks WOULD be from fear.
However, I might also guess that your definition of fear factor has unfairly
biased the correlation against these very types of hitters.  Hmmm...  does
anyone out there have a logical explanation as to WHY the stats are negatively
related?

------------------------------------------------------------------------------  | Dave Donat                   | "You can't even swing a dead cat without    |
| Interactive Fiction Buff     |  hitting some bum who dumped on me"         |

|                              |        - Carla, "Cheers"                    |  ------------------------------------------------------------------------------

Strikeouts and Walks

Quote:

>>after factoring other things out. Guess what? The association
>>between K's and BB's is NEGATIVE. I tried it with all kinds
>Hmmm...  does anyone out there have a logical explanation as to WHY the
>stats are negatively related?

Simple.  "Good" hitters will swing at strikes, and not swing at balls.
"Bad" hitters wil swing at strikes and balls.  Swinging at balls results
in more strikes and less balls, which means less walks and more strikeouts.
So "bad" hitters will have lower BB numbers, and higher K numbers, than
equivalent "good" hitters. Negative correlation.

He waded through the ankle-deep debris and shreds of crumpled paper on the
table on which Saavik was curled asleep, clutching a mangled page in her
hand.  The words were malformed, torturously carved, but they were legible
and correctly spelled: SPOCK NOTGO MY NAME IS SAAVIK.

Strikeouts and Walks

With regard to the relationship between strikeout
frequency and walk frequency - I suspect that if
we could truely "hold everything else constant" we
could establish a positive relationship between the
two frequencies (at this point I'm not sure though).
The problem is that the data don't really let us hold
other things constant. The posting by Billones makes
pitches. I agree with that. I'd go a step beyond though.
Power hitters, in order to generate greater bat speed,
have to hit the ball "out in front" (pull the ball). That
means they have to commit to swing earlier and are more
prone to be fooled (one of the aspects of which is swinging
at bad pitches). So there should be a positive relation
between power and strikeout frequency. I know this all
sounds like gooblety-gook but the point is that if that's
true then it is hard to untangle the relationships between
power which is associated with kfreq and with pitcher fear,
walk frequency which is associated with pitcher fear and
batter patience and; kfreq which is associated with just
about everything else mentioned so far.
Gerry

Strikeouts and Walks

Quote:
Dave Donat writes:
>Gerald R Hobbs writes:
>>Some time ago, actually a long time ago, I posted a
>>brilliant pontification - to wit: free ***s walk
>>a lot because they swing and miss a lot. That gives
>>the pitcher more chances to walk them...
>>after factoring other things out. Guess what? The association
>>between K's and BB's is NEGATIVE. I tried it with all kinds
>>and combinations of regressors, looked at all kinds of
>>collinearity diagnositics and generally tried everything. The
>>relationship between K's and BB's is consistantly negative
>>and usually significantly so.
>Hmmm...  does anyone out there have a logical explanation as to WHY the
>stats are negatively related?

I'd guess it is because free ***s swing and miss at pitches that
more disciplined hitters would take for balls.  Take someone like
Andres Thomas.  (Please!)  He is consistently swinging at pitches
that hitters with an ounce or more of judgement would ignore.  So,
high Ks, low BBs.  This happens with some reputable hitters, too
(Willie McGee comes to mind).  They swing and miss, giving the pitcher
more chances to walk them.  However, they also swing and miss at the
pitches which would result in a walk.  So you wind up with a negative
correlation, in spite of guys like McGriff, McGwire, and McCanseco who
roll up big numbers in both categories.

Bob Davis, Sledderbrenner

Strikeouts and Walks

Quote:
>Simple.  "Good" hitters will swing at strikes, and not swing at balls.
>"Bad" hitters wil swing at strikes and balls.  Swinging at balls results
>in more strikes and less balls, which means less walks and more >strikeouts.
>So "bad" hitters will have lower BB numbers, and higher K numbers, than
>equivalent "good" hitters. Negative correlation.

Not necessarily.  A "good" hitter (at least a good contact hitter--ala
Gwynn, Larkin, Pucket, etc) might swing at "both"  (or, in any event,
anything close) in an attempt to actually get a swip at as many pitches
as possible (and thus not "waste" a pitch by eliminating the possibility
of watching it float by as a strike--when you could have swung at
it--and missed--and watched the same thing happen>).  Thus, someone who
is after contact might do so to the SEVERE detriment of their walk total
(they'd trade in about 50 BB or so for the extra 15-20 hits they might
get in those 50 opportunities, instead.....).  Thus, the contention that
someobody is a better hitter if they get more walks is really rather
sillly in many cases (especially if that person hits #2 or #3  in a
lineup--where they want to get on base, but more often want to advance
runners).

Rob