It's no question that Placido Polanco's struggled with the bat this year. Judging from his advanced numbers, it's a bit luck (low batting average on balls in play [BABIP]), but part of the low BABIP is due to hitting about 3% more fly balls than normal.
But, that aside, is it time to move him down in the order? He's at -5.7 Weighted Runs Above Average this year at the dish. Polanco's value with the bat is tied in his abilities to hit .320 or better because he doesn't walk much (just 5.4% of his plate appearances) but he also doesn't strikeout much (7.1% of all PA's). He also doesn't hit for power (.112 Isolated Power), so he's a guy who needs to hit for a high average to allow the few walks he does take to bump his on-base percentage into a reasonable state.
This year, he's hitting just .263 entering play this afternoon and that leaves his OBP at a measily .317. His normally higher average also bolster's his slugging percentage over the .400 mark normally, but due to his bad average, he's fledgling at .390.
The way I'm going to analyze the lineup is through David Pinto's Lineup Analysis tool over at Baseball Musings. The most used lineup for the Tigers in 2009 has been:
Granderson
Polanco
Ordonez
Cabrera
Guillen
Laird
Inge
Raburn
Everett
The Birchman brought this up in the comments, and proposed the following lineup:
Clete
Guillen
Cabrera
Granderson
Thames
Inge
Polanco
Laird
Everett
Now, I don't like Guillen in the 2 spot. He's struggled at the dish this year, so I'd put Inge (yes, even with his struggles) in the 2 spot.
How do these lineups stack up against each other and what would be the optimal lineup with both sets of personnel? Glad that you asked...
The most common lineup for Detroit would average 4.652 runs per game. Coming into this afternoon's tilt with the Baltimore Orioles, the Tigers had scored 488 runs in 106 games, for an average of 4.604.
If you extrapolate their actual rate of runs scored, you get 746 runs per 162 games. If you use the most common lineup would score about 754 runs over the course of a season -- remember, this doesn't count days off for guys or account for injuries, so that's the reason for the discrepancy. The Tigers have just 55 games left after today's series finale, which would mean if they used their most common lineup every game for the rest of the year, they would score around 256 more runs this year.
Now, if they employed the proposed lineup, they would average 4.792 runs per game. Extrapolate that over an entire season (a.k.a. everyone in that line up play 162 games, which isn't reasonable), you get 776 runs scored -- some 30 run improvement (or, about 3 wins worth). Now, just for simplicity's sake, the most common lineup averages 4.652 but the Tigers have averaged 4.604 which is a difference of 0.048 runs between the two for injuries and resting players. So, let's take that off of this lineup's projection and you now get 4.744. That comes out to 769 runs -- so just about 7 runs slashed from the 162 game average. Over the next 55 games, your'e looking at 261 runs scored which is just about 5 runs (half a win's worth) over the most common lineup.
But, that begs the question, what would be the best lineup?
Well, for the most common lineup Detroit's used this year, it'd be:
Cabrera
Inge
Ordonez
Polanco
Granderson
Raburn
Everett
Guillen
Laird
That batting order would score 4.814 runs per game and 780 runs -- a 4 run improvement over our proposed (and better) lineup. But, subtract the 0.048 for reserves/injuries and the best lineup using these nine would be 4.766 runs or 772 per 162 and 262 the rest of the way.
Using the proposed batting lineup, the most optimal one would be:
Cabrera
Inge
Laird
Thames
Clete
Granderson
Polanco
Everett
Guillen
And that lineup would produce 4.888 runs per game (792 per 162 games). Over a 55 game span the rest of the way, that lineup would produce 269 runs. Subtract 0.048 runs per game, though, for reserves/injuries over 162 games and that becomes 4.84 R/G which is 784 runs in 162 games and 266 runs the rest of the way.
For what it's worth, the most common lineup, if rearrange to score the fewest runs possible, would include this order:
Everett
Laird
Granderson
Guillen
Raburn
Ordonez
Inge
Cabrera
Polanco
And they would average just 4.369 runs per game (708 per 162/240 rest of the way).
And for the proposed lineup, the worst order would be:
Everett
Laird
Granderson
Guillen
Polanco
Clete
Inge
Cabrera
Thames
And that lineup would score 4.442 R/G (720 per 162/244 the rest of the way).
The results aren't that dramatic for the most common and the proposed lineup's (about 4 runs), but any little bit helps in such a close division race.
Thursday, August 6, 2009
Subscribe to:
Post Comments (Atom)
Posts
Wow. Seriously well done Mike. I've made the argument with non-SABRmetrites (?) before that Cabrera leading off would be a seriously awesome move but really had no way to prove it. I've settled for the argument that he should be at least batting three to guarantee he gets the largest number of at-bats over the course of a game/season.
ReplyDeleteI'd also like to agree with putting Bingo in the two spot. My thinking was if Inge is as hurt as he is, maybe the numbers (even the OBP) may tend to dip. I dunno if I mentioned this in my last comment, but most likely, his power numbers will fall off yet he'll still tend to get on-base at a reasonable clip, making him a perfect fit for the 2 spot.
I think Guillen's a better fit as a 5 or 6 hitter, it just seemed like as he's getting healthier and Inge may be moving the opposite way, it might make more sense to switch them down the stretch. Based on numbers alone, Bingo's a definite 2 hitter and CG's a 6.
That brings me to my next question: Should Marcus Thames be playing everyday? He mashes, but doesn't walk a whole ton. I know Baseball Prospectus has been expecting him to be an everyday player in this lineup each of the last 3 years, but every single season he gets moved to the back of the outfield rotation. I know part of it has to do with his terrible defensive ability and Clete's very good UZR, and Guillen's limited positional flexibility, but at what point does Thames's power alone make him a necessity for a team struggling to get runs across?
Thanks again for your quick/detailed post regarding my question Mike. Now I can actually quote some numbers when making my lineup argument. Plus, you turned me on to that very good Lineup Analysis tool -- thanks a ton. There's a reason you're writing for BTB -- congratulations!
Thanks, Birchman. It was no problem and didn't take too long at all.
ReplyDeleteThat lineup tool always puts the best hitter at the top of the order and it makes sense. To score in the NBA or NHL you want you best player to control the ball or the puck. In baseball, you want your best player to have the most PA's because he'll do the most good things of anyone on the team.
It would be helpful to do this for L/R splits, since that obviously changes the lineup drastically.
ReplyDeleteInge's stats are heavily propped up by his success vs. LHP. He's a mediocre hitter against RHP, so I don't think he should be anywhere near the top of the lineup vs. RHP.
As for Polanco, he should be towards the bottom of the lineup, but apparently the tradition of having a high-contact, hustle guy in the 2-spot should prevail even if it flies in the face of any statistical support.
Jeff, L/R splits probably could be done. I just did this as a quick-and-dirty general look at the lineup.
ReplyDeleteanyone who thinks cabrera should bat lead off needs their brain checked... period.
ReplyDeleteThese stats are fun and all but not practical. The idea that you should bat your best hitter first automatically doesn't make sense, especially if he is slow. And Brandon Inge would not be a good second hitter as he isn't someone who hits behind runners. And yes you want your best hitter to get more plate appearances but you want them to be the most important plate appearances. By hitting him first you negate the three run homer threat with his first at bat every single game.
ReplyDeleteAnonymous 1: Thanks. I'll look into that. Should have an appointment set up around Thursday of next week.
ReplyDeleteAnonymous 2: They aren't practical because they won't ever be used, correct. But to say they aren't practical because they are wrong is a fallacy, given the statistical evidence.
And I'm not sure what "he's not a good hitter to hit behind runners." Doesn't everyone hit behind runners in a lineup? Also, there are 8 other innings in a baseball game for Miguel Cabrera to come up with runners on base in the lead off spot. Something that has a chance of happening in the first inning is not a reason to not do something, given there's 8 more innings in which the occasion can happen in any given game. 8 is greater than 1.
Anonymous's post made my point perfectly (though by irony) - there's a long way to go in getting people away from certain game/lineup management ideas that are based on intuition and tradition and no numeric evidence.
ReplyDelete"These stats are fun and all but not practical."
No, they're the opposite of not practical. What the program does (I believe) is simulates games based on players' numbers.
"The idea that you should bat your best hitter first automatically doesn't make sense, especially if he is slow."
"Brandon Inge would not be a good second hitter as he isn't someone who hits behind runners."
"By hitting [Cabrera] first you negate the three run homer threat with his first at bat every single game."
These are the kind of ideas that most people subscribe to that, while they make perfect sense logically, don't play out statistically. They also show why Leyland's a terrible on-field manager (though so are most of his colleagues). There's not a lot of reasons to have your leadoff hitter be fast - maybe if you have a guy who's constantly stealing bases at a high rate. It's much more important to have someone who gets on base a lot. And why does the #2 hitter need to be better at hitting behind runners than the #6 hitter? I'd rather have someone who gets hits - isn't that better than hitting a groundball to 2B so the runner can advance to 3rd? As for Cabrera, as Mike said you'd negate the possibility of a 3-run HR only in the 1st inning. What are the odds that both guys get on ahead of him? Maybe 10-15%? And how often does Cabrera hit a HR - maybe 5% of the time he steps to the plate? So we're going to put him in the 3 spot so he can hit a 1st inning 3-run HR less than 1% of the time. I'd rather him get as many at-bats as possible, and this program shows what would happen when you put your best on-base guy in the leadoff spot and forget about the antiquated traditions about lineups that fly in the face of the evidence and that we need to started advancing beyond.
Well hitting behind runners means moving a runner on second over to third with a ground ball or a deep fly to the right side, thereby giving your third hitter a chance to score him with a fly ball.
ReplyDeleteAnother flaw is that you don't allow for the different kind of pitches players see at different spots in the order. One way in which it might be beneficial in batting Inge second would be that he would see more fastballs if there were a liitle speed on the bases (not so likely with Cabrera on first). It also doesn't deal with the different mental approach hitters need to take in different slots in the order. Are'nt you automatically assuming that all the players will have the same stats no matter where they hit in the lineup. And as far as the negativity about Leyland and the rest of his "terrible" colleagues, you might want to check Laird's slugging and ops and realize you are defending a system that declares him the "optimal" third hitter. Or better yet, close the computer and watch some games.
Anonymous: I don't see your first point as a 'flaw' in the system. Outs are how this game is timed and the most precious commodity in baseball. Therefore, the run state actually decreases with things like sacrifice bunts/hits -- so I don't view them as positive things until late in games when down by a run/up by a run.
ReplyDeleteYour second 'flaw' you've pointed out doesn't make sense to me. Is Miguel Cabrera going to get more breaking balls than Placido Polanco? Most likely. Is it because of where they hit in the lineup? No. It's because Polanco is a light hitter (who's hitting as poorly as Gerald Laird, who you've voiced your displeasure with) that's unlikely to drive a fastball deep into the LF bleachers. I don't believe in the theory that pitchers throw different pitches based on the line-up spot. My thoughts are that it's coincidence considering nearly every manager has a light-hitting batter in the 2 spot ahead of their thunder bats. That's why it seems like the 2nd spot in the order gets "less fastballs" than the 3-4-5 spots -- because of the players the managers put in those spots. With Brandon Inge's power surge (and big time slump recently), he's likely to see more breaking balls than fastballs when compared to, say, Ramon Santiago or Placido Polanco -- again, because he's a guy with power in his bat and they are not. This would not change if he were batting 2nd.
Yes, I am assuming that the stats would largely stay unchanged for the most part. The stats that I look at are the triple slash of average/on-base percentage/slugging percentage. Things like RBI's would be altered because those are predicated on what your teammates do -- those are a team thing, not an individual thing. You need to be in a spot in the lineup behind great OBP guys that would give you a lot of RBI opportunities.
Lastly, this model is rooted in thousands upon thousands of statistical samples. You can leave your condescension off of this blog when it's clearly advertised as a numbers-based Tigers blog. If you feel that I've been condescending in my responses to you, then I apologize, but it's not welcome here. If you disagree with the things I write, a discourse on it is welcomed -- but only in an adult manner. Things to end comments like "close the computer and watch the games" like I'm not an avid baseball fan that soaks up any baseball telecast he can, is not only a slap in my face, but just an unnecessary comment. I haven't ever elected to not approve any comment made on this site, but have no issues in ignoring the comments that bring an holier-than-thou attitude. This blog is no different than the commenting guidelines on all other blogs. Respect. Another comment of that ilk, and I'll just elect to not publish your comments or continue in a fruitless discourse that's on a slippery-slope to name calling. Thanks.
I didn't mean to offend, but I do think sometimes numbers while very valuable, can be over analyzed. And with the blog titled "Fire Jim Leyland" I didn't think my words were too out of line. My apologies if you took offense. I do politely disagree with the idea that Inge wouldn't see more fastballs in the second spot as long as base stealer like Granderson is on first. Also I do like the idea of a better RBI man in the second spot, I would just prefer less of a fly ball hitter than Inge. Tony Larussa has long been a believer of a strong RBI man in the second spot and I have enormous respect for him as a manager.
ReplyDeleteMike,
ReplyDeleteI say let his comments stay - it's good to expose people to some stats-based thinking.
Anonymous,
The problem with your argument is that it presupposes certain things that the program is trying keep independent. For example, "you might want to check Laird's slugging and ops and realize you are defending a system that declares him the "optimal" third hitter." That's what's valuable about this system - it doesn't "assume" that a third hitter should be great, and statistically, it shows what would be the best lineups to use based on simulations of games based on real-life numbers.
If you want to criticize the methodology of the program, fine. But when you do research (on anything, not just baseball), you can't dismiss the validity of the results simply because you don't like them, or because they aren't in line with traditional ideas about baseball.
As for a player's mental approach based on lineup spot, it may make a difference in a few cases. But is there any reason to believe that the way things are done now is optimal?
Anonymous: It's cool, I just took it the wrong way.
ReplyDeleteThis blog started as a fire Leyland blog, but I've moved away from that for the most part. I used to have something about that in the about me but took it down recently. No biggie. I'm pretty loose in my comments, but was a bit touchy earlier for reasons unrelated to this. No worries!
As far as Inge goes: I want someone who'll get on base and hit the ball hard. This year, Inge (until his slump) had been that guy. Though, his struggles and bum knee, I see reasoning not to have him in the 2 spot.
Jeff: His comments for sure are staying.
I confess to be a traditionalist when it comes to baseball but I try to keep an open mind. The hardest part of picking the best Tigers lineup is a lack of high OPS guys to choose from! Eric
ReplyDeleteWe agree there, Eric. The Tigers offense has stunk it up this year, no matter what angle you look from.
ReplyDeleteMy biggest beef with Leyland is batting Granderson leadoff vs. lefties. There's a strong argument that he should be a platoon player anyway, but when you put a 270-280 OBP guy at the top, that's just inexcusable.
ReplyDeleteIt's things like this (putting your "fastest" guy at the top) that just make me think he's a traditionalist who isn't willing to look at different ways of doing things.
Mike,
ReplyDeleteYou haven't posted anything about this, so I'll just chime in here. Sorry about that. Do you think the Tigers ought to sign John Smoltz once his 10 day "designation for assignment" runs out (assuming nobody trades for him in the interim)? Personally, I'd say hell yes they should. Imagine how fired up and motivated he would be to play for his home town team in the heat of a pennant race. That 1987 trade has haunted the Tigers ever since, and it would be a magical moment if John can somehow get his act together and produce some good work out of the bullpen in these last 6 weeks. Maybe the guy really is finished at the age of 42, but it'll cost the Tigers next to nothing to find out. Miner has been a disappointment lately and John can take his place. It just seems to me that the Tigers have nothing to lose and everything to gain. I think it's about time Johnny comes marching home with the fans cheering him on!
Robert: I think it's worth exploring. He'll be as cheap as can be and he's not going to be worse than Willis/Robertson/Bonderman/Zumaya, so why not. I think bullpens should be built as cheaply as humanly possible, and maybe he finds his form (I doubt it) in the Mud Hens bullpen (I'd like to see him give it a go down there first).
ReplyDeleteHow about that! We finally agree on something. Who knows, he just might retire, but I would sure like to finally see this guy in a Tiger uniform. I think many Tiger fans would. Here's hoping it happens.
ReplyDelete