Friday, June 10, 2005

Why Frank Robinson is Like Saudi Arabia
But Craig Biggio is Like North Korea  

Statistics are like a girl in a fine bikini.
They show a lot but not everything -- Toby Harrah

As a player without five tools myself, I always have appreciated the Hit By Pitch (HBP) as a way to get on base. When I was a kid, I needed all the walks and HBPs I could get because for the longest time, I was a dead-pull hitter without a lot of power. Opponents would recognize pretty quickly that everything I connected with landed in about the same place, so any way I could get on that didn't require my soft liner to elude a fielder was a net credit to my middling efforts. BTW: I still let myself get hit in slow-pitch softball...it doesn't hurt and the lousy umps who normally call everything that doesn't bounce before it reaches home plate a strike have a really hard time calling a pitch that hits a batter in the box a strike, whereas they have no qualms about calling it a strike if you jump out of the way.

I usually was near the top of the league in HBP, though no one else, I think, wanted to be. The first major leaguer I ever saw who played it roughly my way was Ron Hunt, and I was at two games where I saw him get plunked. Twenty-five HBP/season were not unknown for him, and one year, 1970, he harvested 50 of them.

Like me, he was a mistake hit-by...Hunt intentionally leaned into pitches that were mistakenly thrown too close to him -- a tribute to his greater skill because major league pitchers make far fewer mistakes than little league ones do (and throw harder, obviously). The biggest difference though, was I always pretended it was my lack of mobility, whereas he proudly showed he was doing it on purpose, and that difference is an important part of this discussion.

Mike at Mike's Baseball Rants this week did a swell era-adjusted study of all-time major league HBP leaders (Hunt is in there), and it makes for a great example around which to discuss both metrics and interpretation of what they mean.

One of the best things about Mike's presentation is, he doesn't even bother to lead with the raw totals for players...a traditional leaders' table ordered by career or single season highs. Because over the decades, the frequency of HBP occurence has shifted. His very first, marquee, stat table is contextualized to the seasons the player labored in. So he creates a measure, the number of HBP by which the player exceeded the league average HBP over the number of plate appearances. Let's call it "career exceeded expectations" or CEE. Here's the top of his table, with the key columns (read his article...really interesting and more complete data) for the players who got 100 or more HBPs over what the league average would have been for his career.

Name First Last   CEE
Hughie Jennings 1891 1918   226.77
Don Baylor 1970 1988   218.43
Ron Hunt 1963 1974   208.91
Tommy Tucker 1887 1899   195.72
Craig Biggio 1988 2004   174.80
Dan McGann 1896 1908   170.32
Minnie Minoso 1949 1980   150.67
Frank Robinson 1956 1976   132.81
Jason Kendall 1996 2004   129.82
Curt Welch 1884 1893   121.14

source: Mike's Baseball Rants

If you know much about these players, you can see they are not all that similar. A common mistake people in business or government make too frequently is to look at a table and think every case/person/country/department that is on the high end or the low end is there for the same or similar reasons. In this table, and in a subsequent beyond-baseball example, there's a lot of variation.

In the case of this table, the players fall into three clusters, which I'll call the Ends cluster (which includes Hunt & Biggio), the Means I cluster (includes Minoso, Jennings, Welch, Tucker & Kendall) and the Means II cluster (which includes Robinson & Baylor).

The players in the Ends cluster get hit as an end. They just want to get on base as often as possible so they can be a potential run. In Hunt's case, it was a major part of his OPS, in Biggio's less so, but the HBP itself was the goal.

For the players in the Means I cluster, the HBP is a technique to terrorize the pitcher an unnerve the infielders. Jennings, Welch and Tucker all played for the brawling, effervescent 19th century Baltimore Orioles whose core strategy was actually a tactic -- cognitive terror and raw aggression designed to make sure the umps and opponents were distracted (Welch stole home to win a "world series" like post-season championship, allegedly after telling the pitcher and catcher he was going to do it). Minoso (The Cuban Comet), Jennings (Eeh-Yah) and Welch were all excellent base-stealers who made a show of working the pitcher's attention away from the hitter when they were on base. For them, yes, getting on base was worthwhile, but the HBP was just the prologue to other, more irritating things they could inflict on their opponents.

For the Means II cluster, Baylor and F. Robinson, getting hit was also a means, but to a different end. Both were serious power hitters who knew they could hit better if they could get the fat part of the bat on a pitch and both found that harder on inside pitches. So by arranging to make it incrementally more dangerous for a pitcher to come inside against them, they created a bit of incentive for the pitcher to try to work him away (where they felt they were more effective). A subtle point...Baylor, in the early part of his career was more of a member of the Mean I cluster; when in his mid-30s, his bat slowed a tad and his success with inside pitches was not up to what he wanted, he notched up his technique as an extension of his game. Many are critical of Baylor as a manager, but he was an extraordinarily smart batter.

Being on the leaders table, in itself, doesn't really tell you what the player was aiming at, and if you presume there's a uniform strategy or reason behind that, you not only might be wrong on an individual case, you might be playing into their hands. Want to defeat Hunt? Pitch him away and that will defuse this tactic. Pitch Baylor or Robinson outside, and you're merely doing exactly what they're trying to get you to do.

Too many educated managers make the same mistake. They gather or get presented data about competitors or markets and while the facts are they, the managers make a hash of it by assuming uniformity. And like throwing outside to Baylor, they're somethimes playing right into the hands of their opponents or competitors.

A marketing manager looks at her competitors' marketing and advertising expenditures to see how her next year's budget should get tweaked. Maybe she even accounts for relative cash flow to see the level of commitment each has to marketing effort, instead of looking at just gross expenditures (which don't account for the gross amount each company has to draw upon). But there's a strong need for context -- some of those competitors are going to be new and trying to make themselves known in the field, some don't have new product to push, some are having a temporary slowdown in marketing expenditure to make their books look better in bad times, some pre-paid for advertising and the expenditures numbers don't indicate that.

Without context, the numbers don't add the full value they could. In most measures there's almost never one single reason why all "leaders" or "trailers" appear on the board.

Here's one last example. Recent military expenditures as a percentage of GNP. The CIA invented this figure for positioning reasons during the Cold War and it's still in use. Here's a recent table from www.nationmaster.com for countries' military expenditures as a percentage of their GNP

  Country Description
1. Korea, North 34% (FY02)
2. Mali 15% (FY02)
3. Saudi Arabia 13% (FY00)
4. Ethiopia 13% (FY00)
5. Oman 13% (FY01)
6. Eritrea 12% (FY02)
7. Qatar 10% (FY00)
8. Israel 9% (FY02)
9. Jordan 9% (FY01)
10. Maldives 9% (FY02)
11. Afghanistan 8% (FY02)
12. Bahrain 7% (FY01)
13. Armenia 7% (FY01)
14. Macedonia 6% (FY01/02 est.)

The #1 is a country run by a functionally-demented cadre, plus a significant chunk of their GNP isn't measured by GNP measures (meaning their extraordinary level of military spending becomes an even more exceptional percentage of the part of the GNP that gets included). #2 is one of the poorest countries in the world and their currency got devalued and they spend military $$ on dealing with a neighbor's insurrection troubles. Four of the members don't need their militaries to be significant, but they have a lot of revenue from U.S. oil and return some of it to the U.S. in arms purchases as a diplomatic kickback. A few of these countries are scared spitless about ongoing border disputes and potential insurrections and wars. #14 is probably very different today...at that time, one of its neighbors claimed it had no right to exist as a country and another was having a major war that was threatening to leak over its borders, but that conflict is not going on any longer while the table is old enough to represent a changed reality.

And again "as a percentage of GNP" isn't a real military measure, more of a government-choice indicator. If you were trying to judge how "war-like" or "war-prepared" or "dangerous" any of these countries actually are, you'd have to look at their actual military budget. Take Afghanistan...their military $/GNP is about twice the US', but their expenditures are actually 1/745th of the part of the U.S. military budget that's on the books.

As with the HBP table, sometimes there's intent that's an end, sometimes there's intent that's a means, sometimes it's inadvertent.

There are a whole host of reasons to take presented measures carefully, to make sure you understand the context for a member's inclusion before you do much to shape a competitive or analytical approach.

Doing anything else is too often the equivalent of giving Don Baylor a pitch over the outside half of the plate.

This page is powered by Blogger. Isn't yours?

free website counter