Sunday, June 14, 2009

After 60 games

My post-60 game analysis is a bit late – we’re at 64 games now – but I was trying to get a better source for the data that I am using for this analysis.

Specifically, the data I am referring to is Win Probability Added (WPA). What is WPA? When the game starts between two equally matched teams on a neutral field, each team has a 50% chance to win the game. (This is referred to as the Win Expectancy) If the visiting team’s leadoff hitter, Lyle Leadoff, hits a home run, now the probabilities of winning have changed. The visiting team now would have a better than 50% chance of winning and the home team’s chances are now less than 50%. The change in probabilities is the WPA. The increase in the visiting team’s winning probability is credited to Lyle Leadoff, while at the home team’s pitcher, Sherman Shagwell, is credited with the home team’s decrease in their probability of winning. This is repeated for every play. At the end of the game, the total of the individual WPA’s for the winning team will be 50% - since they started with a 50% chance of winning the highest the probability can go is 100%, when the game is over. Likewise, the sum of the individual WPA’s on the losing team is -50%, as their probability of winning dropped from 50% at the outset to 0% at the end of the game.

If you are watching or listening to the game and have Internet access, you might want to follow the live scoreboard with the evolving graph of the Win Expectancy. A particularly interesting graph is the extra inning 10-12 loss to the Orioles, May 27.

When Roy Halladay got Nick Markakis to hit a foul ball fly out with the bases loaded to end the seventh inning, the Blue Jays’ Win Expectation was 97.8%. By the time Aubrey Hull open the ninth with a single, the Jays’ probability of wining was down to 27.3%, but it was back to 90.2% when Aaron Hill homered to give the Jays a 10-8 lead in the eleventh. Even in the bottom of the eleventh, when Nolan Reimold came to the plate with two on, the Jays still had a 52% chance of winning the game. Three pitches later, the Orioles had won the game.

One limitation of the data that I am using is that only the pitcher and the hitter are credited with the WPA. Imagine the bases loaded with one out and Halladay pitching. The batter nails a screaming liner down the third base line that appears destined to drive in three runs, which miraculously gets knocked down by Scott Rolen, who grabs it and fires to Hill at second, who leaps to avoid the runner and gets it to Overbay who digs it out of the dirt just before the lumbering 38-year-old DH hits the bag, for the double play. There is a lot of WPA on that play which gets awarded to the pitcher. Certainly, Halladay deserves some credit for keeping the ball in play, but does he deserve as much as he would have received if it had been a routine two hop grounder DP ball to Scutaro? Rolen deserves a huge amount of credit, since if he had not kept the ball in the infield, three runs would have scored. Likewise, Hill and Overbay each deserve some credit as well. I will keep looking for WPA data that does a better job of attributing WPA to defense and other facets of the game.

So, the foregoing is an admission that the data is limited.

Since the Jays were 33-27 after game 60, the sum of the individual WPA’s has to be 6x50 = 300% = 3.00. Batters had an aggregate WPA of 4.80, Starting Pitching was at -0.45 and the Bullpen was at -1.36. This is a shocker, since some of the Pitching WPA should be rightly attributed to the defense, one of the Jays’ fortes.

Hitting (WPA = 4.80)

Which batters have been contributing to winning games? Are Rios and Wells costing us games? Here's a graph of the individual hitters' WPAs:

With the exception of Rios and Wells, the bats have been contributing. Rios has had a minimal positive contribution, while Wells has been a liability at the plate so far. Here is the cumulative performance of Rios and Wells:

These are graphs that should be trending upwards as a player's contribution to winning adds up over the season. The streakiness of these two is evident. Wells is in a 5-week+ slump, while Rios is bouncing all over the place. As Cito said, these two have to start contributing.

By comparison, Scutaro and Hill have been contributing:

You can see that Scutaro has leveled off somewhat after a great April, but the trend for both is upward as it should be.

The bottom of the order looks like this:

All have contributed, although Rolen, Like Scutaro, has leveled off since April. Lind and Overbay continue to contribute, while Barajas has tailed off a bit in the past couple of weeks.

Millar is overused, and the rest of the bench has been spotty. So far, Inglett does not appear to anywhere close to what he was last year.

Needs: Rios and Wells to live up to even a fraction of their expections and a dependable 9th bat to DH.

Starting Pitching (WPA = -0.45):

Here's the individual WPA data:

We knew that the Jays' rotation would be an issue this year, and the graph is not a pretty picture. Tallet has been better than his WPA suggests. Janssen has yet to rpove to me that he can keep the Jays in games and Richmond has to be more consistent.

Bullpen (WPA = -1.36)

There are really only two performers in the bullpen, Scott Downs and Jason Frasor. the rest of last year's best relief corp has just not been doing it:

League has been a enigma. About 70% of the time, he's been "lights out", but the rest of the time, he either makes me extremely nervous or he's shown us why I get nervous. His -0.659 in the May 4th 7-9 loss against Cleveland is the single-largest negative WPA this season.


I'm not about to make projects for this 20-game stretch. They could be 8-12 or 12-8. It's interleague play, too unpredictable.

No comments:

Post a Comment