Those of you who think that advanced metrics deserve a spot alongside chips, dips and dorks might want to take a moment to stick a couple of breakfast sausages in your ears (I'd recommend using the ones on the Residence Inn breakfast buffet, since they do not belong in your mouth). Actually, don't do it yet, because I'm going to throw you a bone: trying to quantify defense through statistics is about as worthless as trying to avoid french fries and cole slaw on a trip through Pittsburgh. You might be able to cobble together a healthy using dishes from different restaurants, but you aren't going to find everything that you need in one place. Advanced defensive metrics, at least the ones that are available to us peons in the general public, are deeply flawed. Even people who study the stuff will admit to that.
Now, let's talk about Win Probability Added. This isn't a perfect metric by any means. But the data it uses is a lot less subjective than the stuff that provides the basis for a lot of the defensive ratings you will find. Essentially, WPA measures the impact a specific play in a game had on the likelihood that a player's team would win that game. I'm not going to explain the methodology. For the time being, just go with it. It isn't definitive, but it at least provides one prism through which to view a game's events.
Now, let's talk about Freddy Galvis. In Thursday's opener, which happened to be his major league debut, the 22-year-old second baseman happened to be involved in three of the five plays that WPA identified as having the most impact on the game's end result.
Those three plays, according to Baseball-Reference.com, were the three double plays that featured Galvis. As you will see, they represent a microcosm of the two sides of the Phillies' decision to make him their Opening Day second baseman.
How important is it for the Phillies to find a second baseman who can provide some offense?
|| 59 (17.7%)
|| 56 (16.8%)
|| 40 (12.0%)
|| 40 (12.0%)
|Not very important|
|| 30 (9.0%)
|Let the kid play, because I don't know yet|
|| 108 (32.4%)
Total votes = 333
According to WPA, the double play that Galvis turned with Jimmy Rollins in the first inning was the second-most important play of the game, trailing only John Mayberry Jr.'s double to right field in the seventh inning, which set up Carlos Ruiz's sacrifice fly, which scored the game's only run.
Mayberry's double increased the Phillies' probability of winning by 15 percent. The Rollins-to-Galvis-to-Wigginton double play increased their odds by 10 percent.
What WPA does not tell us is the technical difficulty of the pivot that Galvis made at second base, or the throw that he made to first, or whether a lesser defender would have been able to nail Andrew McCutchen by a half step, thus preventing Roy Halladay from having to deal with runners on the corners and one out in what would prove to be a 1-0 Phillies victory.
It also does not tell us whether an eight-hole hitter with a better bat would have avoided grounding into double plays in the second and fifth innings, which is what Galvis did in his first two at-bats against veteran lefty Erik Bedard. The first double play came with runners on first and second and one out. The second one came with a runner on first and one out.
According to WPA, the first double play decreased the Phillies' chances of winning by nine percent. The second one decreased their odds by seven percent.
WPA, at least to my knowledge, also does not consider the fact that even if Galvis struck out, or moved the runners, the Phillies still would have had the pitcher batting with two out in both situations.
Again, WPA is not definitive. But it does crystalize the two options the Phillies pondered when deciding how to replace Chase Utley:
1) Give the job to Galvis, whom they believed was the best defensive option they had at their disposal.
2) Find a better hitter and sacrifice whatever defensive ability such a move required.
All told, Galvis' four at-bats decreased the Phillies' odds of winning by 21 percent. The six plays he made in the field -- five assists, one pop fly put out -- increased their odds by 23 percent. Apart from the double play, the other five plays were routine. But the one non-routine play he made was the second-biggest play of the game (at least according to WPA).
The Phillies can't expect to shut-out every opponent, which means they won't be able to afford a second baseman who makes one-and-a-half outs during every at-bat. Then again, they also can't expect the top four hitters in their order to go 1-for-14 with three strikeouts and two walks every game.
It is clear that the Phillies will spend the entire season monitoring the infield bats that are available to acquire. But it only makes sense to replace Galvis if the new bat's offensive capabilities off-set his defensive liabilities. And players like that are not readily available. It wouldn't surprise me to see the Phillies add a guy like Mike Fontenot to replace Pete Orr on the bench, giving them a back-up plan in case Galvis' bat turns into a huge liability.
But the liability of Galvis' bat also depends on the viability of the bats in front of him. If John Mayberry Jr. and Carlos Ruiz give their Opening Day production on a regular basis, then the Phillies will be able to afford a mulligan at another spot in the order.
That might not be the case if they did not have three pitchers who hold opponents under three runs in the majority of their starts. But right now, the Phillies offense has enough potential that it makes sense to err on the side of preventing those pitchers from incurring extra runs.
In today's SportsWeek piece on Galvis, I referenced the winning percentage that National League teams posted when allowing X amount of runs in 2011. I'm not sure if these numbers make mathematical sense. But they make intuitive sense, so I'll throw them out there.
Below is the winning percentage in each run situation -- when a team allows one run, two runs, three runs, etc. -- as well as the decrease in winning percentage from one run to the next. In other words, when a team held an opponent to one run, it won 89.4 percent of its games. When it allowed two runs instead of one run, that winning percentage decreased by 15.1 percent to 74.3 percent.
1: .894 (-.106)
2: .743 (-.151)
3: .593 (-.150)
4: .481 (-.112)
5: .350 (-.131)
6: .250 (-.100)
7: .180 (-.070)
8: .130 (-.050)
As you can see, the two most "important" runs were the second run and the third run, both of which lowered a team's odds of winning by about 15 percent. Since the Phillies have pitchers who allow between 2 runs and three runs per game, the prevention of an extra run can be huge.
Again, not sure if that makes any sense from an empirical perspective. But it made sense to me.
Anyway, that's all the math we need for one day. Game 2 is seven hours away.