Wednesday, March 31, 2010

Baseball Prospectus at the Ballpark

Baseball Prospectus will be hosting a Ballpark Event at Camden Yards on Monday, April 12. Tickets are just $25 (including your game ticket - this sounds like a fantastic deal), and the event will include none other than Andy MacPhail. I really wish I could go to the event and ask him questions about player valuation methodologies, lineup construction and the relationship between front office and field personnel, but alas, I will instead be studying International Economics around the corner. If you go, I'd love to have you write a guest post about the event.

Here are the details:

Monday, April 12, 5:15 PM at Camden Yards
BPro writers: Clay Davenport, John Perrotto, Will Carroll, Jay Jaffe, Steve Goldman, and more
Celebrity guests: O's President Andy MacPhail, Sirius/XM's Mike Ferrin
Tickets: $25, call Dan Stahl at 410-547-6239

Tuesday, March 30, 2010

Lineup Analysis

Yesterday's Orioles lineup seems to be more or less what the club is planning to roll out for Opening Day. How did Dave Trembley do?

The lineup, with CHONE projections for OBP and SLG:

1) Roberts .358/.427
2) Jones .349/.497
3) Markakis .373/.484
4) Tejada .330/.435
5) Scott .337/.471
6) Wieters .355/.460
7) Atkins .327/.412
8) Pie .332/.435
9) Izturis .301/.334

According to the Baseball Musings Lineup Analysis tool, Trembley didn't do terribly, but there is room for improvement. This lineup projects for 5.133 runs per game, while the optimal lineup projects for 5.157, a difference of 3.888 runs over the course of a season: a bit less than half a win.

How can we reasonably improve the above lineup? Well, Roberts leading off and Cesar Izturis batting 9th are no brainers. Given what we saw for the importance of the number two slot in the lineup, we should also move Nick Markakis into the number two hole. The tool then likes some combination of Scott, Wieters and Jones in the numbers 3, 4 and 5 slots, with Tejada, Atkins and Pie rounding out the bottom half of the order. Since we want to break up lefties, let's put those six guys in this order: Wieters, Jones, Scott, Tejada, Atkins, Pie (feel free to swap Pie and Atkins if you would like). That lineup works out to 5.155 runs per game, tied for the fourth most optimal lineup. The moral of the story? Batting Tejada cleanup instead of further down in the order is likely to cost the Orioles somewhere between 3 and 5 runs this season. Keep that in mind when Jim Palmer extols the virtues of Tejada's proven run producing abilities right before another inning-ending double play.

More on Tillman

This tweet arrived on my BlackBerry this morning, and it gives an indication as to why Chris Tillman is headed for Norfolk instead of Baltimore next week:
Tillman: "If I were 2 stick w/ my fastball, change, curve it woulda been a totally different outcome, results-wise." worked on cutter instead
Now, the immediate reaction might be to say "Well, that's dumb. If you want to make the team, you need to try and pitch your best." However, over the long run, Tillman is likely to be a better pitcher if he can master the cutter. Should a 22-year old prospect on a team staring up at what might well be the three best teams in baseball work on his cutter and risk starting the year in Norfolk, or should he try get outs any way possible? Yeah, I'm voting for the former. I hope the team has communicated to him, though, that his promotion depends not on "outcomes" but on his progression in mastering the cutter and the strike zone. We know that he can get AAA hitters out with his three traditional pitches, and if that's his focus in April, he'll likely do so again. Still, this is just another reason that we should all take spring training results and statistics with a giant mine of salt, and one more indication that his starting the year in Norfolk is far from the end of the world.

Monday, March 29, 2010

Chris Tillman

There has been a bit of a kerfuffle over the past few days as news broke that David Hernandez may break camp as the Orioles 5th starter while Chris Tillman begins 2010 in Norfolk. I have no problem with this. Here are some reasons, but they all surround two facts: David Hernandez will pitch 2010 at age 25; Tillman at age 22.

1) Tillman threw 161.2 IP across AAA and MLB last season after throwing 135.2 in AA in 2009. That's an increase of nearly 20%, and a corresponding increase for 2010 would bring Tillman to 193 IP. I sincerely doubt the Orioles will allow him so many innings; much more likely his target is 175 to 180 IP. Managing his ability to reach that target is much easier if he spends a few weeks in the minor leagues, where pushing his start back a day or lifting him in the sixth inning is much more justifiable.
2) Hernandez, by virtue of being older, will likely have a less stringent innings constraint. By virtue of passing the "injury nexus," the Orioles can push him a bit harder.
3) Since Hernandez is older, he also is more likely to be close to a finished product. If he pitches poorly as a starter this season, the Orioles can move him to the bullpen without wondering what might have been.
4) If Hernandez succeeds as a starter, then the Orioles have a further flexibility with Tillman. He could replace an injured or ineffective pitcher (and someone will get injured), he could piggyback with Matusz if necessary to manage workloads, or he could be used out of the bullpen.
5) If Hernandez does not succeed, then the original plan (Tillman in the rotation, Hernandez in the bullpen) can still easily be implemented.

While I'm obviously disappointed that Tillman has not pitched well enough this spring to win the job outright, I am far from lamenting this development. We all know that young pitchers do not develop in a straight line. That the Orioles can suffer a set-back and still have five viable starters is a big change from years past. That this is the debate we're having less than two weeks until Opening Day is a good sign.

Wednesday, March 3, 2010


Spring Training games begin today! And that means we get our first Orioles lineup of the season. As just tweeted by beat writer @Britt_Ghiroli, here's how the O's will look today (positions guesstimated for now):

CF Adam Jones
SS Cesar Izturis
RF Nick Markakis
3B Miguel Tejada
CA Matt Wieters
LF Luke Scott
1B Garrett Atkins
DH Josh Bell
2B Justin Turner

The Orioles are obviously without Brian Roberts and Nolan Reimold as those two ease into spring training while recovering from injuries, but once those guys are back, what will the lineup look like for Opening Day? Most importantly, what should the lineup look like for Opening Day?

Let's make a few assumptions:
1) Reimold will start in left field
2) Atkins will start at first base
3) Scott will DH
3) Everyone is healthy

I've been reading The Book: Playing the Percentages in Baseball by noted sabermetricians Tom Tango, Mitchel Lichtman, and Andrew Dolphin, and these guys have quite a bit to say about lineup optimization. Let's highlight their key point:
Your three best hitters should bat somewhere in the #1, #2, and #4 slots. Your fourth- and fifth-best hitters should occupy the #3 and #5 slots. From slot #6 through #9, put the players in descending order of quality.

They draw several other optimization conclusions regarding speed, baserunning, and strikeouts, but the above rule of thumb should be simple enough that everyone can understand it. So, who are the Orioles best hitters? The diagram above shows each starter's projected wOBA according to CHONE.

Nick Markakis is the Orioles best hitter. We'll stick him in the #2 slot to get him the most at-bats. Next up is Adam Jones, who derives more of his value from power than OBP; sounds like a clean-up hitter to me. Reimold, Wieters, Scott, and Roberts are up next. Where should each fit? Here are the OBP and SLG projections for each:

Reimold: .355/.477
Wieters: .355/.460
Scott: .337/.471
Roberts: .358/.427

Since Roberts derives more of his value from his on-base abilities, we'll put him lead-off (and, also, I don't think it would be possible to convince Dave Trembley that he should bat anywhere else). Reimold and Wieters project pretty evenly, so we'll put Matt Wieters in the #3 hole since that has traditionally been the place managers slot their best hitter. Reimold gets penciled in at #5. Rounding out the lineup we'll just go straight down the list: Scott, Tejada, Atkins and Izturis. Here's what we have:

2B Brian Roberts (S)
RF Nick Markakis (L)
CA Matt Wieters (S)
CF Adam Jones (R)
LF Nolan Reimold (R)
DH Luke Scott (L)
3B Miguel Tejada (R)
1B Garrett Atkins (R)
SS Cesar Izturis (S)

According to the awesome lineup analysis tool at Baseball Musings, this lineup would be good for 5.239 runs per game (849 runs for the season). This may seem high, but keep in mind that this lineup wouldn't play together every day, and the expected performance of replacements would be lower. How does this compare to the optimal lineup (according the tool)? The optimal lineup would score 5.249 runs per, or 16 runs more over the course of the season, a difference of a win and a half. What's that optimal lineup?


That's pretty close to the lineup we constructed above, and I'm pretty surprised that swapping Reimold and Jones and moving Scott from 6th to 8th makes that big of a difference.

Looking at the optimized lineups, comparing them to Trembley's creation, and drawing a few (perhaps hasty) conclusions, what can we say? I think we can make two: First and foremost, Dave Trembley does not understand the importance of the #2 slot in the lineup. Cesar Izturis batted second 28 times in 2009. TWENTY-EIGHT TIMES. He posted a .294 on-base percentage. He. Should. Not. Bat. Second. Ever. Even when Trembley got the top of the lineup right last season, he got the reason wrong. He would move Markakis and Jones to the second slot in order to protect them in the lineup, often when they were perceived to be struggling in the middle. In fact, he should have been moving the hot hand to the second slot (we won't get into streaks right now). Secondly, I think Trembley is overrating the "veteran presence" and "run producer" reputation Tejada brings to the middle of the order. Tejada accumulated RBIs in the past because he was a good hitter; he was not a good hitter because he had some special RBI generating power. I think his days as a good hitter are pretty much over, so he should be hitting lower in the order, letting Reimold and Wieters clear the table set by Roberts and Markakis.

In 2010, I'm going to keep track of loony lineup construction, bad bullpenning and silly strategizing with the Twitter tag #Trembleyfail. Follow @theorioleway to keep track, and let me know when you see the same. Together, we can be the force that brings the Orioles an extra win just by not screwing up the lineup.