I thought I'd run this article again; its from Baseball Rants, but it is still relevant, especially with the Astros run scoring trouble so far this year.
Update [2005-5-6 21:36:15 by Marc Normandin]: Adam Everett's 2005 line stands at .233/.310/.389; Lance Berkman has returned from the DL, so Luke Scott is not an option in the lineup any longer.
How can some clubs and managers still not understand the value of on-base percentage? Honestly, it boggles the mind. Here is the latest example: Adam Everett is leading off in Houston. Adam Everett, owner of a .315 career OBP; Adam Everett, who last year had 17 walks in almost 400 at-bats. Baseball Prospectus says the league OBP in the NL was .329, and Baseball-Reference says it was .341, but either way, Everett is far from it. Baseball-References' definition for lgOBP is "league's on-base percentage (pitchers removed) with the same home field." so that has something to do with the difference, getting rid of all of the auto-outs. Either way, Everett is not qualified to lead off. Garner's reasoning, broken down, is that 1) Everett is fast and 2) Biggio is a better hitter...all right. Let's break that down. Everett is fast, so Garner assumes Everett will steal bases (of course that hinges on his being on base). Everett had all of 13 steals (with only 2 caught stealing, at least he has a great percentage at 87%) in his in 485 plate appearances...so we could say he'll steal roughly 20 bases give or take a couple in a full season. Are those 20 steals worth having your leadoff hitter get on base only 32% of the time? Probably not...and this is my thinking:
If Everett, who does steal bases at a good percentage, although not many of them, was going to be used in the offense as a speed weapon, the best place for him is NOT the leadoff spot. Rather it would be the 8th spot. If you are a team that is going to continue to play small ball and steal bases while bunting, then you should let your stolen base "threat" Everett bat in the 8 hole. Why? Well, the pitcher is coming up after him correct? Everett could get on base (31-32% of the time anyways) and the pitcher could try to bunt him over after a steal of second, or Everett could just get a good jump on the bunt. Instead of trying to force runs where runs would be created with a correctly setup top of the order (Biggio-Lane or something to that effect), why not try to make runs where runs don't happen at the bottom of the order? Also, when you pinch hit for the pitcher later in the game, this would help setup the hit and run more effectively, with the speedy Everett on base. I am not saying this is the optimal plan to end all plans, but for a team like Houston who has lost a great deal of offense, it is a good idea I believe. Houston cannot afford to waste outs at the top of the lineup, and would be much better served trying to create runs down at the traditional black hole at the bottom of the order rather than making two separate ones. Now this is not to say that you should bunt early on in the game just because Everett is on first and the pitcher is up. Teams do not correctly use the small ball skills to score runs, as they kill rallies before they even have a chance to get going by wasting outs and the like. Later in the game where one run is all you need then it is ok to play some small ball and work the runners over; early on just because you want to squeeze a run in for the sake of a run is not the time to do it.
Here is another thought concerning this situation: are pitchers (specifically in this situation, Houston's pitchers) more prone to hitting groundballs than flyballs, making the steal more valuable in order to avoid the double play? We will look solely at Houston's pitchers for now, in order to avoid generalizing based on a small sample size. This way we can focus on this one scenario regarding Everett in order to make our decision, and maybe at some point come back to this as a whole. I will focus on the starting pitchers as well, just because they pitch and hit the most.
2004 GB/FB Ratio; GIDP
Roy Oswalt: 37/11 = 3.36; 1 GIDP
Roger Clemens: 23/9 = 2.56; 2 GIDP
Andy Pettitte: 9/5 = 1.80; 1 GIDP
Brandon Backe: 5/2 = 2.50; 0 GIDP
Brandon Duckworth: N/A
2003 GB/FB Ratio; GIDP
Roy Oswalt: 16/6 = 2.67; 0 GIDP
Roger Clemens: N/A (AL)
Andy Pettitte: N/A (AL)
Brandon Backe: N/A (AL)
Brandon Duckworth: 13/3 = 4.33; 1 GIDP
It should be noted that some of the career rates are higher than these few seasonal stats, but the GIDP's are scarce. This is a tough analysis considering Clemens and Pettitte spent their whole careers in the AL, and Backe has not even had much of a career yet.
Considering the ratio of groundballs to flyballs here, we can assume that there were many at-bats with no one on base or no one on first since there are an extremely low number of GIDP's here, especially for guys who do not run out groundballs hard in order to retain stamina. We can check their at-bats with runners on base though just to make sure, but sadly I can't seem to find them anywhere. If someone knows where to find this information please tell me.
Looking to drum up some support from people who are smarter than I, I stumbled across this statement by Joe Sheehan from BP in early 2004:
"The batter at the plate is much more likely to score the runner from second than he is from first. Teams will often use their best base stealers at the top of the lineup, even players with low on-base percentages, in front of their most powerful batters. In fact, they should be using those players lower in the lineup, in front of their least powerful hitters. Risking an out to advance from first base to second base is much more important when the guy at the plate can't get the runner home from first base."
Bingo. Everett was born to hit in the 7th or 8th spot (I say seventh because it just dawned on me that Brad Ausmus also plays for this team, creating a vacuum of offensive suckitude in spots 7-9), and the pitchers were born to get out behind him trying to advance him or simply because of their inability to do anything but. The leadoff spot is not for Everett or his ilk in the major leagues, but who should leadoff for Houston? Let's look at their starting 8:
C - Brad Ausmus (No..never.)
1B: Jeff Bagwell
2B: Craig Biggio (the traditional pick)
3B: Morgan Ensberg/Mike Lamb
SS: Adam Everett (you know my thoughts)
LF: Luke Scott
CF: Willy Tavares
RF: Jason Lane
Biggio's OBP dropped to .337 last year, but he had an increase in power. Whether the power spike sticks around to offset the drop in OBP is the real question surrounding his performance. An interesting little piece of information: Willy Taveras had a .402 OBP in 2004 in the minors, which came out to an EqOBP of .355. BP claims he needs more seasoning in Triple-A before coming to the majors, but also says that they need him to get Biggio out of center. Well he did, and he is here minus the seasoning, but if he is capable of fulfilling his PECOTA projected stats for 2005 (.252/.330/.334) then he would be a better choice than Everett even before he starts to improve. Even if he regresses to his 2003 Eq form, the line was .261/.343/.341...better than Everett once again. Not to mention he has stolen 54, 57, and 55 bases in the last three minor league seasons. Of course, putting Taveras in the #1 spot would go against everything I have said thus far concerning putting your fast guys in front of the lightweights. But with the number of lightweights Houston has, maybe this is a necessary evil. Are you going to put Adam Everett in the 6th spot, followed by light hitting (understatement alert) Brad Ausmus, and follow it up with Taveras in the 8th spot to steal in front of the pitcher? That would not be the best idea, because it would stretch an already thin offense. Maybe they should stack the deck up front with Taveras and Biggio and try and play small ball at the back end with Everett and the pitcher or Everett and Ausmus. Ausmus' GB/FB ratio by the way is not as large as the pitchers (career 1.49), but his double play rate (probably the result of being an aging catcher) is large: 13, 30(!), 8, and 13 respectively since 2001. Maybe the lineup should look like this:
One last thing. Joe Sheehan also said that the worth of a man on first in runs with 0-1-2 outs is 0.9116, 0.5348 and 0.2349 respectively. If the runner goes to second, those numbers go up to 1.1811, 0.7125, and 0.3407. For a guy like Ausmus, who cannot drive a guy in from first without some help most of the time, having Everett go to second on a steal (remember the 87% success rate, with a career rate of 89%) would increase Ausmus' chances of driving in a run, thus bringing up his value slightly. Anything that can make Ausmus and Everett more valuable offensively, even slightly, would be a huge boost for a Houston team that all of a sudden lacks true multiple threats.