The dreaded M word was mentioned in last week's article. It's gotten a lot of ink and I do not intend to review the book. There are two very important lessons from Moneyball that deserve discussion though. The irony is that neither idea is really in the book.
Point number one is the open ridicule that needs to be heaped upon those who have consistently down-played both Beane and his methods. Before the book came out the backlash was already building, before most of the over-the-top stuff even came out. One common emotion unites those who dismiss the analytical methods of the A's: fear. Upper management of teams that wed themselves strictly to the old ways are afraid of being replaced, failure, and being expected to put forward a contender on the same amount of money that the A's have. Arrogant dismissal is almost universally a cover of deep-rooted fear. Much of the ridicule comes from scouts. The fear here is even more obvious, because while management can change, the scouts fear their entire job will be replaced by a computer. The idea that scouts will be replaced by computers is silly. It reveals a fundamental lack of understanding about how in-depth analysis affects our understanding of the game. Statistics tells us about trends, probabilities, and group behavior, not individuals. You can't look at a pitcher's line score as a pitching coach and tell him that he needs to walk fewer men. He knows that he needs someone to help him find the right release point, or that repeatable delivery.
Coming back to the uses for statistics, you can appreciate that there are a million different little things that affect a game. With so many different things involved, it can be difficult to properly assign actual value to all of these little things. Emotion plays a dominant role in memory and perception. It hurts when the other team's star gets the game winning hit. The manager should have let someone else beat us! The problem is that letting someone else beat you may consistently result in more losses. It's the type of thing where you can figure out your best course of action by looking at trends in the game over years. However, you are never going to look at a guy's BA/OBP/SLG line and say he's got a hitch in his swing that he could fix.
This leads into the second point, which apparently Mr. Lewis forgot in his lessons from the A's: we don't know everything. This is a lesson that anyone trying to problem solve scientifically quickly learns. When you start trying to measure things and put numbers on them you discover how little you really know. In the case of GMs, they are trying to predict the future. This is something (by most people's definition) that can never be known. This brings us to the notion of balance between statistical analysis and conventional baseball. Statistics is great for determining best bets, both with player acquisition and in game strategy. When it comes to individuals though, there's no substitute for human interaction, teaching, and observing.
This may be a shorter column which is heavier on semantics than crunchy explanation, but with so many newer readers I felt some of these issues merited attention. Next week I'll talk about Baseball Prospectus' new Post Season Odds Report statistic.
Player of the Week
There were a number of players who had really good games this week and as some may remember, a signature game for the week is often critical to winning the award. John Vanderwal had an excellent series against Florida, and I thought he was going to win this week because no one else had really distinguished themselves early in the week. However, Keith Ginter came on strong in the Philly series with 5 RBI and was key to the Brewers lone win. He did everything but steal bases - he took walks, had multi-hit games, scored runs, and raised his OPS 50 points.