Saturday, August 18, 2007
Lineup Permutations
With Jason Giambi back in the fold, the Yankees have a problem now. It's not a bad problem, because having more talent than you have spots for is a good thing. The key will be using the talent on hand in the way that is most likely to help you win games.This is not as clear-cut as the numbers that will follow may make it seem, because you're dealing with egos, over/under performances, possible injuries and fatigue, and many other factors(base clogging, chemistry of the lineup,etc.,). However, since I don't have any way to quantify any of that, I'm just going to look at the data that is available to me and see how the different options shake out.
For offense, I'm using a runs created formula that multiplies OBP times SLG times .92. This is then multiplied by 4.4 PA per player, which is the average PA per game for the Yankees's lineup. This doesn't consider lineup order. OBP and SLG are 2/3 based on the players' projections coming into 2007 and 1/3 based on their YTD performance. For Shelley Duncan I've included his MLEs so his stats aren't skewed by his small sample size MLB performance to this point. Since no one projected Duncan coming into the season, here's the projection I used for him:
| Year | Team | Last | First | G | AB | R | H | 2B | 3B | HR | RBI | BB | K | BA | OBP | SLG |
| 2005 | Trenton | Duncan | Shelley | 142 | 521 | 76 | 113 | 23 | 1 | 27 | 71 | 43 | 147 | .216 | .277 | .422 |
| 2006 | Tre-Col | Duncan | Shelley | 104 | 384 | 43 | 88 | 21 | 0 | 18 | 51 | 31 | 91 | .230 | .287 | .426 |
| 2007 | Scranton | Duncan | Shelley | 91 | 325 | 53 | 88 | 17 | 1 | 23 | 66 | 38 | 88 | .271 | .350 | .537 |
| 2007 | Yankees | Duncan | Shelley | 17 | 42 | 9 | 13 | 0 | 0 | 6 | 13 | 5 | 12 | .310 | .383 | .738 |
| Projection | 109 | 416 | 59 | 106 | 26 | 1 | 25 | 67 | 43 | 109 | .255 | .325 | .501 |
I just figured out my own versions of Duncan's MLEs (major league equivalencies) then weighed each season to get a .325 OBP and .501 SLG, which doesn't seem out of line.
For defense, I'm using a combination of career defense as rated by zone rating combined with YTD 2007 performance as far as runs saved per game. For players with small sample size I made some assumptions based on scouting reports and my visual observations. I realize this turns what is supposed to be an objective exercise into a somewhat subjective one, but I think that the logic I used here is at least somewhat sound.
So the numbers that follow are the overall runs created on offense and the runs saved above/below average on defense for a single game. Multiply by 162 if you want to see how that translates to a full season.
Let's first look at the lineup that had become the basic lineup of choice prior to Giambi's return.
| Pos | Player | RC | Def |
| 1B | Phillips | .537 | .037 |
| 2B | Cano | .686 | .077 |
| 3B | Rodriguez | .900 | .073 |
| SS | Jeter | .721 | -.164 |
| LF | Matsui | .741 | -.062 |
| CF | Cabrera | .588 | .042 |
| RF | Abreu | .677 | -.018 |
| DH | Damon | .620 | .000 |
| Off | 5.469 | -.015 | |
| Over 162 | 886 | -2 | |
| Per Game | 5.454 |
We'd expect this lineup to score 5.469 runs per game using the runs created formula I'm using assuming 40 plate appearances per game. We'd expect them to be .015 runs below average defensively per game. So net, offense plus defense this lineup is worth 5.454 runs per games. Over a full season, we'd expect them to score around 886 runs and be around 2 runs above average on defense. Bear in mind that Melky's projection coming into the year was a little a low and that his YTD performance would be worth .684 RC for ever 4.4 plate appearances. If we swap that out with his .575 combined projection you get an extra 17 runs over a full season. However, as much as I'd like to that, that's not objective analysis, so the projection stays as is with the understanding that it's possibly not indicative of a likely change in Melky's talent level (or a decline by Damon which makes him worse than projected).
So, let's look at a few other combinations.
First up, let's just swap out Damon for Giambi.
| Pos | Player | RC | Def |
| 1B | Phillips | .537 | .037 |
| 2B | Cano | .686 | .077 |
| 3B | Rodriguez | .900 | .073 |
| SS | Jeter | .721 | -.164 |
| LF | Matsui | .741 | -.062 |
| CF | Cabrera | .588 | .042 |
| RF | Abreu | .677 | -.018 |
| DH | Giambi | .818 | .000 |
| Off | 5.668 | -.015 | |
| Over 162 | 918 | -2 | |
| Total | 5.653 |
No impact on the defense obviously, but over a full season it'd be about a 30 run upgrade on offense.
How about Damon at first and Giambi at DH? To make this work I assumed that Damon would be a slightly below average defensive 1B, but not too bad of one. I have no idea if that's a reasonable assumption or not.
| Pos | Player | RC | Def |
| 1B | Damon | .620 | -.006 |
| 2B | Cano | .686 | .077 |
| 3B | Rodriguez | .900 | .073 |
| SS | Jeter | .721 | -.164 |
| LF | Matsui | .741 | -.062 |
| CF | Cabrera | .588 | .042 |
| RF | Abreu | .677 | -.018 |
| DH | Giambi | .818 | .000 |
| Off | 5.751 | -.059 | |
| Over 162 | 932 | -10 | |
| Total | 5.692 |
It's too bad this isn't a realistic option, because it's better than the first two. Unfortunately, it likely isn't or we'd have seen it already.
We also have to consider Wilson Betemit and Shelley Duncan in the mix. How about Duncan at first and Giambi at DH?
| Pos | Player | RC | Def |
| 1B | Duncan | .659 | -.059 |
| 2B | Cano | .686 | .077 |
| 3B | Rodriguez | .900 | .073 |
| SS | Jeter | .721 | -.164 |
| LF | Damon | .620 | -.003 |
| CF | Cabrera | .588 | .042 |
| RF | Abreu | .677 | -.018 |
| DH | Giambi | .818 | .000 |
| Off | 5.668 | -.052 | |
| Over 162 | 918 | -8 | |
| Total | 5.616 |
We don't have enough data to have a good read on Duncan's defense at first although the consensus is it's not good, so I gave him a rating equivalent to a -10 over a full season.
Another option is, Betemit at first, Giambi still at DH.
| Pos | Player | RC | Def |
| 1B | Betemit | .612 | -.006 |
| 2B | Cano | .686 | .077 |
| 3B | Rodriguez | .900 | .073 |
| SS | Jeter | .721 | -.164 |
| LF | Matsui | .741 | -.062 |
| CF | Cabrera | .588 | .042 |
| RF | Abreu | .677 | -.018 |
| DH | Giambi | .818 | .000 |
| Off | 5.743 | -.059 | |
| Over 162 | 930 | -10 | |
| Total | 5.684 |
We have a winner. According to these numbers, this is the best combination the Yankees can run out there if Damon at first is not an option, at least against righties. The nice thing is that since Betemit doesn't hit lefties they could use Duncan against lefties. I'd also assume that Betemit is a better defender at first base than Duncan, although without enough data to know I made him touch below average for this exercise.
The Yankees can also use Duncan in the OF corners although he's supposedly going to be pretty bad out there.
The best defensive team is probably the one below.
| Pos | Player | RC | Def |
| 1B | Phillips | .537 | .037 |
| 2B | Cano | .686 | .077 |
| 3B | Rodriguez | .900 | .073 |
| SS | Jeter | .721 | -.164 |
| LF | Damon | .620 | -.003 |
| CF | Cabrera | .588 | .042 |
| RF | Abreu | .677 | -.018 |
| DH | Giambi | .818 | .000 |
| Off | 5.547 | .044 | |
| Over 162 | 899 | 7 | |
| Total | 5.591 |
For all the crap Andy Phillips has been getting lately, offense plus defense he's not really that big of a problem.
There are way too many potential combinations for me to run through in this space, so I've uploaded the spreadsheet that I used for this if you want to play around with it yourselves. Just select a player's name and you'll get a pulldown which lets you change the player and their values will fill in automatically.
There are 39 games left in the season. I'd try to use the Duncan/Betemit platoon at first in at least half of them. I'd use Damon to rest Melky, Abreu, and Matsui once a week, and I'd save Andy Phillips for late inning defense. The nice thing about the current Yankee bench is they can rest almost anyone on the team without losing a ton of value. It seems to me that the main thing the Yankees have to avoid is starting Phillips at first base on the same day they start Damon at DH. If they avoid that, they'll be putting a pretty strong lineup out there no matter what combination they use. Except for House Money day of course.
Page 1 of 1 pages: