Tuesday, March 11, 2008

Review - Stat One

Demonstrating once again that I am not just a political and economics blogger, I picked up a book at the library called Stat One which purports to rate "baseball's all-time greatest players." (I am deliberately omitting the author's name for reasons that will become obvious.)

I have been a baseball fan since I was a kid, which is curious since I had to pick it up on my own (my father passed away when I was young, not that he was a big sports fan anyway). My first teacher was the great Cubs broadcaster Jack Brickhouse, too little-remembered today, even in Chicago. The advantages of daytime baseball and 144 televised games a year meant that the games were the only alternative to soap operas, and I picked up on it, and became a devoted fan (and how many young potential fans are being lost as the Cubs move to high concession-revenue night games and, now, only 60 games a year on free TV?).

My other passion as a lad was mathematics, so I was a natural stats-gatherer and -rememberer. I played table sports games, and derived just as much pleasure out of collecting the numbers as in playing the games. My first computer programs (heck, calculator programs) were written to collect and analyze baseball stats.

So when Bill James came along, I was ready to read what he had to say. Through his brilliant Baseball Abstracts I understood the game better, appreciated talents that were lost in the statistics we were given. (I don't mean to slight the few who were doing similar things before; in particular, Earnshaw Cook did some very early work which, while flawed, was trying to answer some of the same questions.) That you could find the true value of a player or a strategy, rather than accepting "conventional wisdom," was revolutionary.

Of course, there has been backlash against what has been perceived as the "geeks take over" mentality. Many people criticize James and those who have come after for an over-reliance on numbers. And there are those, but James has been consistent in using statistics as a tool to better understand and appreciate the game.

Let me steal from myself: the other day, Joe Posnanski posted one of his typically wonderfully long posts, something about wins earned by a pitcher. Along the way, in a footnote, he wrote
I want to point out that I got these statistics from the Bill James Website. I would also like to take a moment to say something about my friend Bill, something that he probably would not want me to say. But every couple of weeks, it seems I will see yet another person throw Bill out there as the essence of statistical evils and pajama-wearing baseball geekdom. It makes makes me pretty ill. True, part if it is because we are friends, but a much larger part is that if you read Bill’s work at all, if you look at his theories with anything resembling an open mind, if you consider at all what he’s getting at … you realize that the man LOVES baseball. I mean loves baseball, loves the game, loves the stories, loves the characters, loves the ins and outs of strategy, loves the moments, loves trying to figure out why things happen, and why so many people buy into stuff that is probably nonsense. I don’t mind people saying that Bill is full of crap — hell, we ALL have to deal with that (and Bill is never shy about saying that someone else is full of crap, including me). But the people who try to make it sound like Bill’s love and understanding of baseball are wrapped up in obscure mathematics and unworkable thoughts and cold data just don’t get it at all.
I wrote in a comment:

Not so much to say about true wins, it’s another way to see things, and that’s always useful (well, not always, but…).

But to pick up on the Bill James note, I absolutely agree. Those who blame him for “ruining” baseball with his consarned stats have not really read him.

There may be people now who love stats, and find baseball a good outlet (because of the massive number of numbers) for that love, and not really love the game, but one could never have that feeling about Bill James. Again, you have to read what he’s written, but his passion has always been apparent (and it is what made the Abstracts so popular).

I always got the feeling that statistics were simply a means to an end for Bill, that they provided the only way for him to get at truths about the sport he loved. And I have always admired the myth-busters, those people who advanced the outlandish simply because the facts took them that way. Bill fits into that group of inquirers.

If Bill is not considered seriously for the Baseball Hall of Fame simply because he challenged conventional wisdom, despite the fact that he changed the way in which we all look at the game, well, that would be one of the two greatest injustices perpetrated by that body (and we all know Buck O’Neil is the other). If they have room for Alex Pompez and Bowie Kuhn, then Bill absolutely belongs there.

To explain to anyone who doesn't know how influential James has been is nearly impossible. So to find a book that suggests that it will rate players in a new way, but whose author has clearly not read Bill James or any of the subsequent research, would seem to be impossible.

But Stat One is such a book. I won't beat the author into the ground, but he claims to have found, "a comprehensive statistic...to analyze and compare baseball players across history with one another and to help answer...countless [questions] that remain unanswered. To my knowledge, no statistic like this currently exists, or has ever existed for that matter." That statement is specious nonsense, many statistics to do this have been proposed (by the way, his list of questions include such perennials as comparing Mantle to Mays, Cobb to Hornsby, who is the greatest player in history?).

Then he proposes taking his P/E Average (for Offensive Production and Efficiency Average) and combining it with traditional statistics and subjective factors, then classifying each player as Category 1 through 5. Of course, if his new stat is really "the single best statistic available to measure a baseball player's offensive success and contribution," why does he have to look at other statistics? Why does he look at Most Valuable Player awards? How does he measure leadership when he hasn't seen many of these players? Why does postseason success matter so much, when that is so team-dependent? Why can't I find a new statistic to help me answer these questions?

There are two components to his average. The first is Net Runs, R + RBI - HR. This measurement has come up a lot over the years (I thought of it when I was about 10), and it has a certain appeal (for me, it was that you could calculate it off the back of a baseball card). The author explains why we subtract home runs: Willie Mays comes to bat with two men on, hits a three-run homer. The Giants lead 3-0. If you don't subtract the homer, Mays gets credit for four runs, but the Giants only scored three. So the author subtracts the home run itself.

But, wait. The two guys on base each get credit for scoring a run, so now you've given five runs for a 3-0 lead. Wouldn't it be better not to subtract the homer, but divide the total by 2? Then the two guys on base each get 0.5, Mays gets 2.0, and that seems a little more fair (after all, if the other guys don't get there, Mays doesn't account for 3 runs). But even that adjustment doesn't solve all the problems with this stat, which has been extremely well-chronicled. (We still have the issue that runs and runs batted in are poor measures of accomplishment, dependent as they both are on the performances of others.)

The other component is Complete Bases, which is total bases plus walks plus hit by pitch plus net stolen bases. This isn't a bad measurement of achievement, but it does mix together elements of on-base percentage and slugging percentage.

At this point the author goes off the rails, as he doubles Net Runs, adds it to Complete Bases, and divides the sum by plate appearances. Why does he double Net Runs? So it will be roughly equal to Complete Bases. That's it, he wants to value them both the same, so he arbitrarily doubles one.

At this point we're only at page 8. I won't take any specific examples (Babe Ruth comes out as the best player ever, what a surprise). None of his results is truly bizarre, though many of his comments are ("[Derek] Jeter is the pure embodiment of a winning ballplayer who rises to the occasion when the game is on the line....He doesn't think his team will win. He knows it."). This is pure emotional twaddle, and appears periodically through this long book.

What's worse is the author starts to believe his own self-hype, a common failing of books of this type. In his comment on Johnny Bench, he states, "Unfortunately for Reds fans everywhere, he's only second on my list." Oh, no, get the counselors out to Cincinnati, this writer has Bench second, watch for falling bodies.

This is the kind of book that came out in the early 1980s when Bill James proved there was a market for this. Most of these pretenders fell by the wayside quickly. How, with all the knowledge that has been accumulated, a book like this gets published, I can't begin to imagine.

And it's not out of a vanity press. It's published by McGraw-Hill. A major publishing company! And what's worse is that it was championed by (and the foreword was written by) Bill Madden of the New York Daily News, who says he's going to consult it before making his Hall of Fame selections. I'd give Ray Romano a vote before I'd give one to Madden.

You can spend $17.95 ($12.21 discounted) on this book, or you can look online for The New Bill James Historical Abstract. It's a little dated, but it is based on far more solid research, and each page has more entertainment value than the dry stat recital of Stat One. Agree or disagree with James' ratings, they come from the mind of someone who has done serious thinking about the game, and, more importantly, someone who loves the game.

No comments:

Clicky Web Analytics