The Search for the Perfect Stat

Statistics are, and always have been, an integral part of baseball. Numbers are intertwined through all of its history, defining legend, woven into legend. Numbers like 60 and 61, .406, 5714, 56, 1.12, or 714 are ubiquitous; inseparable from the records and the men who set them.

We had the basic stats really from the beginning. Batting average, ERA, hits, stolen bases, strikeouts, and pitcher's wins and losses were all commonly tracked and known before 1900, and home runs and RBI, although not unknown, came into common use by 1920.

It's somewhat surprising, then, that it wasn't till the 1970s, sparked by things like when a (then) young night watchman from Kansas named Bill James began writing about baseball in his garage, that the world of baseball analysis began to develop itself. Not a lot had changed; for the most part, the same basic stats were still the basic benchmark in the baseball world.

This will, of course, never completely change, merely because those basic stats ARE the basic stats. Hits and walks will remain the easiest way to score runs. Runs will always decide who wins the game. Wins will always be the sole priority, and home runs will still be the most immediately impressive individual feat anyone can perform.

And there were some exceptions, too. In the 1940s and 50s, Allan Roth and Branch Rickey came up with what would become slugging percentage and on-base percentage. You may add that to the seemingly infinite list of contributions Rickey made to baseball as a whole, but that's beside our point here. Rickey, years ahead of his time (as was his wont), was, as the (then)-Brooklyn Dodgers' general manager, seeking to better quantify a player's production. Slugging percentage and on-base percentage were a step above the traditional batting average, home runs, and RBIs that had served as the standard for so many years.

But for the most part, the world of statistics was relatively static for those eighty years or so; perhaps remarkably so, in fact, as I previously remarked. At the most basic level, stats are used for two different purposes: first, to measure past production; second, to predict future production. These are, of course, interrelated, but those are the two questions, "How good was he?" and "How good will he be?", both tying directly into "How good is he?"

It was in the 1970s and 80s that, truly begun by the work of Bill James, as I said above, that there became a real movement on this part; that looked through the traditional benchmarks of the standard stats to find the things that more accurately marked and predicted success. James' works developed into the Bill James Baseball Abstract, a yearly publication where he ... wrote about baseball. They're pretty cool; I've read some of them. It was the very beginning of the revolution; new things were being discovered, and old things challenged, often for the first time. And this spawned a whole community, a community of baseball nerds, some of which organized itself into the Society for American Baseball Research (SABR). Men like Bob Davids, Pete Palmer, Rob Neyer, James, and John Thorn, among many others, led these first charges into advanced evaluation, talking about things like runs created, or fielding-independent pitching.

This was, however, still very much a nerdy thing. Sabermetricians (so named for SABR, which published journals and articles about these new concepts) were something of an odd, elitist group, with odd, theoretical ideas about the game which were just that: theoretical. The professional organizations surely took little notice; baseball was played on the field much as it always had been and sabermetrics had little significance in the public eye.

That all changed with Michael Lewis' 2003 bestseller, Moneyball, where Lewis got a detailed, behind-the-scenes look at the Oakland Athletics' front office, run by now-legendary Billy Beane. The book (now a movie, starring Brad Pitt, for some reason. I don't really recommend the movie (it's not that bad, really, just not that great;) I would recommend the book but for quite a lot of the coarse language so prevalent in the masculine atmosphere of a baseball setting. If you can handle that then it's a pretty good read; or it was, for me, who is very much interested in baseball and stuff like that.)

...Lewis' book sought to find out how the A's, in the small market of Oakland, were consistently able to compete with the juggernauts of the American League, dominated by the massive wallets of the Yankees and Red Sox. Obviously, with such a disadvantage in glamor and budget room, the A's had to do some major things right to compete at the level they seemingly could consistently do. (Moneyball looked in detail at the 2002 MLB season, where the A's won 103 games, including twenty in a row, with literally the lowest payroll in all of major league baseball.) And what Lewis found was that Billy Beane, heavily influenced by the writings and research of James and others, understood player evaluation so much better than other teams and could find a player, undervalued by others, who could contribute valuable production at a low cost and, on the other hand, pass on players who had eye-catching stats in some departments (say, forty home runs) but did relatively little besides that. Beane is known for his high dependence on on-base percentage; realizing (rightly) that the best way to score runs is to not get out, and that a walk is as good as a hit, he targeted players who could use those abilities to gain those advantages.

Moneyball was, for these reasons, a huge event in the history of baseball. It changed the game, honestly and truly. Almost overnight, nerdy statistics like on-base percentage migrated from the realm of the obscure and interesting-but-nerdy to the cool and very trendy. Every team was looking for the next Billy Beane, every team began striving to catch up in this area of accurate player evaluation. And sabermetrics went from the background to the mainstream.

This all sparked what I here call the 'search for the perfect stat'. The stats world is now a constant turmoil of new developments. OBP was 'the thing' in the early part of this century. Bill James' original Runs Created has been eclipsed and improved into wRC and wRC+. Wins Above Replacement (WAR) seeks to measure the actual number of wins a player contributes to his team. And those are only the tip of the iceberg. In this decade, especially since the introduction of StatCast in 2015, we have launch angle, exit velocity, and spin rate. WAR is now on the back of Topps baseball cards. And with these new developments, we have moved on from the older, less useful, less accurate stats like pitching wins and runs batted in.

Or have we?

A stat can only tell us so much. The 'wins' stat for pitchers is admittedly very much dependent on things outside of the individual pitcher's control, rendering it that much less accurate when measuring that pitcher's results, let alone ability. This came to the forefront in 2018, when Jacob deGrom, pitching for the hapless Mets, led the league with an astonishing 1.70 ERA in 217 innings, yet finished with a W-L record of a mere 10-9. deGrom won the Cy Young, as he should have; he was no-question the best pitcher in baseball, yet he only won ten games.

And so it's true: wins are not a good method of measuring pitcher's ability. 'Pitcher A has 10 wins; Pitcher B has 9' is basically meaningless. But wins do tell us something; perhaps it's not a great stat for many things, but it still does mean something. 'Pitcher A has 17 wins; Pitcher B has 4.' That's a solid indicator that A is better than B; at least; A won 17 games, B only won 4. Cy Young [the original] had 511 wins in his career. That's astonishing. Denny McLain was the last to win 30 in a single season; Dizzy Dean was the last in the National League. That means something. Those are significant feats, no matter how dated or unwieldy the stat. Yes, there are extenuating circumstances, but there always are.

Because no stat is perfect.

Oh, perhaps we're getting closer. Perhaps WAR is one of the most accurate measures out there, perhaps not. I'd say it's more accurate than many, but still far from perfection. And here's the wonderful thing: we'll never find it. No one will ever be able to reduce baseball to a number. Some numbers will tell us things. Some will tell us more, some will tell us less. But we will never solve it, because it's not meant to be solved.

"We haven't figured out anything yet. A hundred years from now, we won't have begun to have the game figured out."

-- Bill James