It Doesn't Always Add Up
August 30, 2005
As I scored today's Oriole's game against the A's, I couldn't help but
think about the events that I was recording and how the numbers they
generated would be analyzed. For example, in the 12th inning,
Eric Byrnes jogged to a flyball and fumbled the catch, resulting in a
three-base error. Normally, the official scorers in Baltimore are
reluctant to pass out errors, especially on home players. But
this one was a no-doubter - Byrnes had tried to make a play on a ball
that required reasonable effort to get to and had simply blown
it. The fact that a good number of bad plays in Camden Yards
would be called errors at RFK (the other park where I score quite a few
games) or many other parks reminded me of a thought I've had quite
often over the last five years of recording what happens in these
games. And that is this: stats simply don't tell the whole
story. So anyone using them exclusively to analyze what happens
on the field
In the Orioles' case, their pitchers get charged with more earned runs
than they might were they pitching elsewhere because of the generous
scoring of the home OS. I don't have any idea how significant the
difference is, only that there is a difference. I do know that
Camden is a good park for triples as far as park factors go despite the
fact that it doesn't have many quirky features that are typical of
triples parks. I believe the official scorer's preference for not
calling all misplays "errors" is a significant contributor in that
respect.
There was another play in this game that will be missed by the
boxscore. In the bottom of the eighth inning, Alejandro Freire
(pronounced Free-dee) hit a ground rule double. He was replaced
by pinch runner Luis Matos. David Newhan tried to sacrifice him
to third, but Justin Duchscherer made a throwing error that allowed him
to reach base. So now there are runners at first and third and
nobody out. Next, Sal Fasano hits a ground ball to short and even
though Marco Scutaro did a quick look to third to try to hold Matos at
third, Matos had an opportunity to try to score. At the very
least, he could have forced a throw and/or gotten himself into a
rundown. But he didn't. He just moved about 10 feet from
third base and just stood there as Scutaro stepped on second and threw
to first for a double play. So instead of the O's possibly
scoring the go ahead (and possibly winning) run, or at least have two
men on with only one out, now they only had one on with two out.
Brian Roberts followed with a soft grounder to second ending the inning
and leaving what would turn out to be the O's last chance of winning
the game stranded at third.
It was clearly a significant play in the game, yet no stat will record
the scoring opportunity squandered by Matos' indecisiveness.
Since errors are only a defensive statistic, there won't even be a
footnote as to why the O's didn't threaten more that inning. The
reason I bring this up is that these kinds of plays have a real
baseball impact, and sometimes players get benched for stuff like
this. But if one only looks at the numbers, the only thing they
show is Matos hitting .362 in August. Theoretically, he could
find himself riding the pine more often if he continues to make
baserunning blunders like this one despite what appear to be very good
offensive numbers. A player hitting .360 over a month but costs
his team 5 runs with bad baserunning decisions might not be worth much
more than a guy hitting .250 or .260 who is very efficient with his
baserunning opportunities.
Another instance where stats don't always tell the real story is pitch
counts. Not the total pitch counts necessarily, but the
strike/ball breakdowns. It's generally believed that a pitcher
should throw strikes and that the more often he throws strikes the
better off he'll be. But this isn't necessarily true. If a
team has a pitcher who has nasty stuff like Daniel Cabrera, then yeah,
sure, he should throw strikes as often as possible because the hitters
will have a tough time catching up with his stuff. But if a guy
doesn't throw anything nasty and tops out at 85 mph, say like Tomo
Ohka, then pounding the plate might not be the greatest idea. For
a pitcher like Ohka, he needs to find the plate just enough to make
hitters offer at stuff off the plate. Any more than that and he's
opening himself up for a beating.
To whit: can you guess which starting pitcher threw the greatest
percentage of strikes last year? If you said Paul Byrd, give
yourself a gold star. He's usually among the league leaders in
the category. Some years, like 2002 (17-11, 3.90, 1.15 WHIP),
that's a good thing. He threw 69.3% strikes that year.
Others, like 2001 (6-6, 4.05, 1.41 WHIP) when he threw 66.3%, it's not.
But last year he threw 71.4% yet wasn't quite as good overall as
2002. Jose Lima is also one of the yearly leaders but anyone
who's had him on his fantasy team knows what a wild ride that he can
be. It's true that more often than not the guys who do well tend
to throw more strikes than the guys who don't, but it's not a slam
dunk. For example, Freddie Garcia (60.5%), Jaret Wright (60.7%)
and Tom Glavine (58.1%) were near the bottom last year in percentage of
thrown strikes yet they had pretty solid years. Tomo Ohka had his
best success in 2002 when he threw 65.87% strikes. He was 13-8
that year with a 3.18 ERA and 1.24 WHIP. The next two years he
actually threw a greater percentage of strikes, yet his ERA and WHIP
were worse.
An interesting bit of trivia is that Curt Schilling led the majors in
strike percentage in 2001 and 2002 with 70.59% of his pitches going for
strikes... the kicker is that he threw the exact same number of pitches
in both years: 3709.
All this to say that even as methods and tools for statistical analysis
improve - and they have made significant strides in the last five years
- we're not yet to the point where we can just look at stats and know
which way to go with our fantasy acquisitions. One still has to
watch the games in order to better understand what they mean and what
they don't reveal in order to make the most prudent moves.