This season, it seems like when the Cardinals lose, it is in the most frustrating way possible. I tossed and turned all night just replaying Carlos Martinez hanging that changeup over and over and over again. The good news, if you prefer to think this way like I do, is that the offense did score four runs.
Last Thursday, Joe wrote about Matt Carpenter & his early impact on the offense. In that post, Joe found that when Carpenter got one hit, the team has scored 1.67 more runs on average than when Marp goes hitless. This makes sense as more hits generally lead to more runs in general, seeing as that is a main way to move around the bases and score. This has been no exception for the Cardinals this year.
Throughout the first 35 games of the season, the Cardinals have registered 283 hits for a total of 129 runs.
Matt Carpenter is the leadoff hitter and the spark plug of the offense, so it seems a safe assumption that whether he hits or not plays an important role for how many runs are scored during the game. This assumption is supported by the numbers.
The only way to know the exact significance of Carpenter's contributions is by comparing him to his teammates. I did this by going through the game logs of the past 35 games, and recorded the total runs scored and the number of hits each player recorded in the game. Then I added up all the one hit, two hit, and three plus hit games of each player, and the runs scored in those games by the team and divided the total runs by the number of games. (Naturally, the next step will be to do the same process with times on base, but let's start small.) My complete results can be found in this spreadsheet in the tab "Individual".
My calculations matched Joe's and showed the Cardinals score on average 2.000 runs per game when Matt Carpenter fails to record a hit. This is the least amount for any player with the exception of one.
The difference in runs scored when Matt Holliday goes 'O' fer and when he records a hit is 2.7821, which is the greatest on the team by a large margin, with the third Matt, Matt Adams in second place at 1.889, Peter Bourjos in third at 1.750, Allen Craig in fourth at at 1.700 and Matt Carpenter back in fifth with a difference of 1.667*.
The two Cardinals whose lack of hits lead to the lowest average of runs are Matt Carpenter and Matt Holliday**. Looking at the two together paints an even bleaker or brighter picture, depending on your perspective.
The sample size is indeed small at only 35 games and is by no mean predictive. Just because Matt Carpenter records a hit does not mean that the Cardinals are necessarily going to score four runs. When discussing this post (which is my first solo post that is based on actual events and not just stuff I made up), Ben shared with me something that makes a lot of sense.
I've always thought of lineups as a loop that you want to run uninterrupted as long as possible, so you want players who make the lowest shares of outs (i.e., highest OBP) at the top to keep the loop running. Carpenter and Holliday being stalwarts at the top (whereas there's been a lot of fluctuation in the second spot) probably contribute to the team's run-scoring fortunes correlating with them hitting. They keep the loop going.
Most of us are aware of the Markov Process, which basically shows that lineup construction matters very little, but maybe it Matters (heh) a little, tiny bit as there can be a slight improvement in runs scored depending on how hitters are alined. Matt Holliday, being third in the order, not only drives in runs (leads the team with 22) with good hitters in front of him like Carpenter***, but also is able to score runs (third on the team with 15) with other good hitters like Adams and honorary Matt, Yadier Molina, hitting behind him. That is the second most runs created with 37 (Matt Carpenter leads with 40). Does this explain the nearly three run improvement when Matt Holliday records just one hit? Is it just a case of small sample size noise? The answer is probably a little of both.
*What about Yadier Molina you ask? Check out the spreadsheet. The difference in run average from when Molina goes hittless and records one hit is 0.1667 and the average run total when Molina is hitless is 3.250.
**When I add Matt Adams to this, the sample gets too small to gather any significant data. Sorry, Matt. I did include it in the spreadsheet linked above, however, under the "Matts" tab.
***68%, or 15, of Matt Holliday's RBI were from Matting in Matt Carpenter. (Via Vexedtechie here.)
****If you are having trouble viewing the charts you can do so with these links (hopefully): Hits per game, Matt Carpenter, Matt Holliday, Carpenter and Holliday