If you ask anyone who follows baseball particularly closely what Daric Barton does best offensively, he or she will respond that patience is his most powerful virtue. Yes, the man who drew 110 walks for the punchless Oakland A’s offense in 2010 does seem to have a gift for taking pitches and working the count. This is useful because it invariably leads to either him a) getting a better pitch to hit or b) walking, and always drives up a pitcher’s pitch count. To assess his value, yet not adequately reference how exceptional he is at forcing pitchers to throw pitches, is an injustice to Barton and the exercise in general.

The question becomes: “How does one measure this ability to work over pitchers and its impact on run production?”

In order to find this out, it is necessary to first determine when an average pitcher is most strongly affected by his pitch count. Using Baseball-Reference’s League splits page, one can determine the wOBAA of the average pitcher sorted by the amount of pitches thrown in intervals of 25. In 2010:

{exp:list_maker}From pitches 1-25, wOBA was .326 (wRAA -87.29)

From pitches 26 to 50, wOBA was .321 (-183.63)

From pitches 51 to 75, wOBA was .329 (59.25)

From pitches 76 to 100, wOBA was .341 (299.54) {/exp:list_maker}There are two fairly large gaps here. The first is the jump between pitches 26 to 50 and 51 to 75; the second, wider gap is between 51 to 75 and 76 to 100, measuring .012 points of wOBA. Pitchers are generally at their worst after they hit the 75-pitch count. One can assert that having a pitcher reach that total of pitches has a clear value in terms of runs scored; this is where wRAA comes into play.

In calculating each value for the individual states per game and season, it was found that the difference between a pitcher who has thrown between 51 to 75 pitches (1.99 wRAA) and one from 76 to 100 pitches (10.90 wRAA) was 8.9 weighted runs above average. In essence, every 75 pitches a hitter sees generates roughly 8.9 additional runs for his team.

Examining the data reveals a glaring problem, though. Of what use are the extra pitches Barton made an opponent’s closer throw in the 9th inning? Unless the game is tied or becomes so, the pitcher will not pitch beyond that inning. Even if he did, he certainly wouldn’t be asked to throw 75 pitches (this isn’t college baseball, after all).

So in calculating each hitter’s worth from a “pitches seen” perspective, each batter’s total number of pitches seen was reduced by one-third, so as to accurately reflect the amount of innings thrown by starting pitchers as opposed to relievers. These new totals of pitches seen from starters were multiplied by the total of wRAA per PA, and 162, to show the total number of runs each player added to his team. The top 10 for the 2010 season are as follows:

Daric Barton OAK 43.98 Rickie Weeks MIL 43.12 Chone Figgins SEA 42.38 Jose Bautista TOR 41.98 Bobby Abreu LAA 41.81 Nick Markakis BAL 41.78 Jayson Werth PHI 41.56 Dan Uggla FLA 41.25 Albert Pujols STL 41.15 Mark Teixeira NYY 41.06

The final two months of the 2011 season will be very interesting because, for the first time in several years, it appears the difference between the quality of a pitcher at 51-75 pitches and one at 76-100 pitches has lessened considerably. After a fairly consistent 30-point OPS difference in the previous three seasons between the two states, this season it has fallen to just about seven points. Whether this is an anomaly or perhaps the result of the league becoming more pitching-dominant remains to be seen. Further, it is possible that this will correct itself over the season’s last two months.

All in all, it seems certain that the ability to drive up a pitcher’s count has a strong value to an offense. I hope that this lends some clarity to the subject, perhaps helping to pave the way for an even better look at what appears to be an ambiguously-valued part of run construction.

Jono said...

So how much of the performance difference based on pitches thrown is actually from number of pitches thrown as opposed to from the times through the order effect? I would suspect that most (if not all) of it comes from facing batters the 2nd or 3rd time as opposed to from actual tiring.

Tom said...

This is an interesting point, however I would point to the decrease in control (or increase in pitching carefully late) along with the even slight loss of velocity and/or command. This is speculative, but I would reason that seeing pitchers more often is most beneficial when the hitter has never seen the pitcher before, and less so in other situations.

Jono said...

I don’t know. I think it makes more sense that seeing a pitcher more times in a row lets the batter adapt than that starting pitchers are hitting a wall after 75 pitches thrown. At the very least, those numbers you got seem way too big to be right (do you really think Barton saved his team 40 runs by being more selective than the average hitter? that’s like the difference between Brett Gardner and Manny Ramirez playing the outfield).

At the very least, it seems likely that there are other things going on here that you aren’t taking into account that are skewing your results.

Peter Jensen said...

In calculating each value for the individual states per game and season, it was found that the difference between a pitcher who has thrown between 51 to 75 pitches (1.99 wRAA) and one from 76 to 100 pitches (10.90 wRAA) was 8.9 weighted runs above average. In essence, every 75 pitches a hitter sees generates roughly 8.9 additional runs for his team.Tom – Jono is quite right you have to have a “smell” test of the numbers you generate. Adding 40 runs per year by seeing an additional .5 pitches per PA should have alerted your smell test. You also have to have a smell test for the numbers that you use in your metric. The above quoted paragraph should have alerted you that something was wrong. A wRAA of 10.9 runs for pitches 75 to 100? That would mean that every starting pitcher that pitched 100 pitches or more had let up more than 11 runs on average! I think you must have missed the E-3 at the end of the scientific notation, but you may have made some other mistake in your query.

Let me run you through the numbers for you for 2010 for Daric Barton. He saw 4.38 pitches per PS on average. The league average non pitcher for a non intentional walf or hit by pitcher saw 3.82 pitches on average. So Barton is a patient hitter who added .56 pitches to every PA. As you correctly pointed out it is only the PAs that Daric had against starting pitchers that mattered. It should also be limited to PAs against a starting pitcher when that pitcher actually reached 75 pitches in that game. In 2010 Barton had 477 such PAs. That means that he added a total of 267 pitches. Now to figure the added runs. In 2010 a PA that ended after the starters 74 pitch had a run value 0f .008 runs. A PA that ended between pitch 50 and pitch 74 had a run value of -.003 runs. So pushing a PA into the 75 to 100 bracket adds .011 runs. But it takes adding the league average 3.82 pitches to push a PA up a bracket. 267/3.82 = 70. 70*.011 = .77 runs. And that was Barton’s best year.