Larry Mahnken and SG's

Replacement Level Yankees Weblog

"Hey, it's free!"

The Replacement Level Yankees Weblog has moved!  Our new home is:

Larry Mahnken
Sean McNally
Fabian McNally
John Brattain

This is an awesome FREE site, where you can win money and gift certificates with no skill involved! If you're bored, I HIGHLY recommend checking it out!


Disclaimer: If you think this is the official website of the New York Yankees, you're an idiot. Go away.

February 28, 2006

Spring Training Notes - 2/28
by SG

I'm taking a break from the looking ahead stuff for a day or two, so here's the lowdown on camp Yankee.

Gary Sheffield has apparently calmed down, and after a back spasm issue, looks to be fine.

"I have to take it easy and not overdo it," the 37-year-old Sheffield said. "It's a little tight, but I don't have the soreness. It's not big deal."

Hideki Matsui has a minor knee problem, but it doesn't sound so bad.

It's an old scar that slightly swelled up," said Matsui, who had blood seeping from his uniform on the left knee. "It's nothing specific, it happens during the season. It's not a big deal."

Today is the first intra-squad game. On Thursday, Jeter, Rodriguez, Damon, and Bernie Williams will leave camp for the WBC. Thankfully they have no pitchers playing. From the linked article:

Are they ready to play games that are being billed as having more intensity than the garden-variety spring training game? Can the stars deliver their best performances in early March?

That depends on whom you ask in the Yankees clubhouse. Interestingly, players who aren't participating have doubts, as does Williams. A-Rod, Jeter and Damon believe the level of play will be good, but can't guarantee it.

"No, not ever," said Mike Mussina (who is not participating) when asked if he would be ready to pitch against Mexico in Game 1 on March 7 at Chase Field. "None of us are ready to go. You might think you are but you aren't. You are in [camp] trying to get ready and now they are asking you to turn it up a notch."

This still just seems like a bad idea to me, mainly because of the timing. I just get the feeling someone's going to get injured.

Yahoo's got some spring training photos up for those who can't wait for winter to end.

February 26, 2006

Looking Ahead to 2006 - Third Base
by SG

Alex Rodriguez is the object of scorn by people everywhere, because he has committed the crime of being remarkably talented and very well-compensated for that talent. It is my opinion that the team owners use their media mouthpieces to ridicule Rodriguez for having had the audacity to sign the most lucrative contract in baseball history, ignoring the fact that it was one of them that offered it to him in the first place, in a thinly veiled attempt at keeping other players' contract demands down. Then the unthinking masses follow suit because they have been conditioned to by the Mike Lupicas of the world. It's easy enough to attack the character of Rodriguez, because it's pretty damn hard to attack his baseball ability, which should be the only thing that matters to sports fans.

Rodriguez had a down year by his standards in 2004. Part of it was certainly the move from Texas to Yankee Stadium. Part of it was probably the controversial position change that saw him move to third base even though he is by all accounts a better defensive shortstop than the incumbent Derek Jeter. Part of it is probably just the general volatility of baseball, and maybe part of it was adjusting to being under the microscope in New York. There's certainly nothing wrong with a .286/.375/.512 line with 36 HRs, unless you compared it to his previous 6 seasons.

Rodriguez came back with a bang in 2005. He hit .304/.349/.618 in April, and then improved as the season went on. He was really shaky in the field in the first half before recovering his ability in the second half. Rodriguez's breakdown by month:

April: .304/.349/.618 with 9 HR
May: .349/.513/.686 with 8 HR
June: .337/.407/.481 with 3 HR
July: .281/.400/.552 with 8 HR
Aug: .324/.429/.733 with 12 HR
Sep: .317/.419/.567 with 7 HR
Oct: .500/.500/1.500 with 1 HR

I realize that there are stats which I had never heard of before 2005 which showed that Rodriguez put up all his numbers in blowouts, but the fact is that he hit .327/.426/.664 with 20 HRs and played Gold Glove caliber defense over the last 2 months of the season when the Yankees were fighting for their playoff lives. That's an MVP to me, and apparently to the BBWAA, as he was an easy MVP selection over David Ortiz. The WARP formula I'm using agreed, pegging Rodriguez for 9.0 WARP, compared to Ortiz's 6.7.

Rodriguez had a disappointing ALDS, but as far as I know, he didn't pitch Game 3 or Game 5. He stranded 6 runners, while Gary Sheffield and Hideki Matsui stranded 31 between them. The point is not that Rodriguez did well in the ALDS, it's that his teammates didn't do well either.

Rodriguez has a career line of .305/.393/.534 with 6 HRs in 118 AB in the playoffs. That's not a "choker" to me.

Which brings us to 2006, and what Rodriguez projects to do.

In this table, RARP are the offensive runs above a replacement third baseman using a linear weights -based formula. FRARP is the defensive fielding runs above a replacement player.

ZiPS says Rodriguez will hit .301/.400/.559 in 2006, with 43 HRs and would be worth 61 runs over a replacement player.

Marcel is much harsher, predicting a line of .296/.386/.543, with 35 HRs. While this is a huge dropoff from 2005, it's very close to what he did in 2004 and would make him 43 RARP. Marcel also tends to regress expected playing time, even for someone who's been as durable historically as Rodriguez has, which further depresses his projected value.

PECOTA is pretty close to ZiPS, predicting Rodriguez being worth 60 runs over a replacement shortstop.

Like I said earlier, in 2005 Rodriguez was worth 9.0 WARP.

My first thought when looking at this was that it appears that Rodriguez is going to be almost two wins worse if these projections hold up. However, the good news is that his defense was uncharacteristically bad last year. He was a +10 defender at shortstop in 2003 and a +11 defender at third base in 2004. That means that last year, Rodriguez's defense was 20 runs worse than his previously established level. And if you go by UZR, he was actually an average defender anyway.

I have no issues with expecting some offensive dropoff this year, although I think Marcel goes overboard. ZiPS and PECOTA seem more likely to me, something in the .300/.400/.575 area. If he can bring back his missing defense, from -11 back to plus 10, he'd actually be more valuable than last year. My guess is the truth lies somewhere in the middle. I think it's reasonable to expect Rodriguez to be at worst an average defender next year, so if he matches his ZiPS/PECOTA, he's an 8.5 WARP player and still a legitimate MVP candidate. I'll still mark third base down for a .5 win downgrade relative to 2005. It's tough to expect Rodriguez to match his career-high tying OPS+ of 167 in 2006.

He's a great player with some flaws. To me, that makes him human, not a pariah.

With infield and catcher complete, I have the Yankees as 1 win better than last year. So far, so good.

February 23, 2006

Looking Ahead to 2006 - Shortstop
by SG

Derek Jeter is one of the most polarizing players in baseball. Scouts rave about his defense while defensive metrics ridicule it. Some Yankee fans and Tim McCarver laud him as clutch, others see that his performance in the postseason is not that dissimilar to his regular season performance. However, there's no doubt that he's a very good player and likely a first ballot Hall of Famer on his current trajectory, whether he is overrated or not.

It's tough to assess just how valuable Jeter is because of the fluctuations in his defensive ratings. While all the metrics showed him as bad from 2000-2003, lately he comes out as average or above in every metric but UZR. Since I trust UZR more than any other metric, I'm not sure what to think.

The one thing that is not in question is Jeter's value offensively as a shortstop. Combining an above average ability to get on base (career OBP of .389 vs. LgOBP of .335) with above average SLG ability (Career SLG of .450 vs. lgSLG of .431) and superior baserunning, in both stealing at an exceptional percentage and advancing on hits, Jeter is typically anywhere from 40 to 60 runs better than a replacement-level shortstop offensively.

According to the WARP formula I'm using here developed by Kyle at OBP for you, Jeter was worth 6.7 wins above a replacement shortstop last season.

So what's up in 2006?

In this table, RARP are the offensive runs above a replacement shortstop using a linear weights -based formula. FRARP is the defensive fielding runs above a replacement player.

ZiPS says Jeter will hit .299/.368/.442, with 18 HRs, and which would make him worth 45 runs more than a replacement shortstop.

Tango Tiger's Marcels predict a line of .298/.363/.449, with 17 HRs. A little lower in the OBP and a bit higher in power, but with a little less playing time, which would make him 40 runs above a replacement shortstop.

PECOTA is pretty close to Marcel, to the point where it also predicts Jeter being worth 40 runs over a replacement shortstop.

I don't have any unique insights to Jeter. He's so consistent he's almost boring. I hope he does not get sacrifice bunt happy with Johnny Damon in front of him, and also hope he can combine the spike in walk rate that he had in 2005 with the isolated power spike he had in 2004. Assuming he's at least close to his 2005 fielding rating in 2006, he should be worth anywhere from 5-7 WARP. The projection systems all seem to think he's going to be a little less valuable offensively next year, and I don't see any reason to doubt them. He'll still be good, but projects to be about 1.5 wins worse. I still hold out hope for another '99 season though.

So far, through catcher, 1B, 2B, and SS, I have the Yankee lineup as 1.5 wins better relative to last year.

I guess I should also mention some of the news coming out of spring training. From

Torre hasn't decided how to line up his sluggers in the middle of the batting order, but he does know that Giambi will join Alex Rodriguez, Gary Sheffield and Hideki Matsui in the 3-6 spots.

Way to go out on a limb Joe.
Tanyon Sturtze, who is a week or so behind the rest of the pitching staff after rehabbing his tired right shoulder all winter, threw off a mound for the first time Wednesday.

This really shouldn't be a concern, but if Sturtze can't go and Dotel's not ready, the Yankee setup corps is going to be a little thin. I hope Kyle Farnsworth can handle the workload.

To me, the most exciting story in camp so far has been Phil Hughes's first batting practice.

-"He throws hard, and he's just a baby -- 19 years old," said manager Joe Torre. "The thing that's unusual for a kid as young as he is, his curveball is really impressive. His stuff is very real."

-"That kid is going to be good; he reminds me of Rocket," Giambi said, making a comparison between Hughes and Roger Clemens. "He's young, but that fastball, it's late. I don't care what the radar gun says, it seems like it's on top of you. He's got good stuff."

-"He has the best arm in camp, no doubt about it. Better than all these guys," said Posada, pointing to a row of lockers which included Randy Johnson and Mariano Rivera. "I don't care how old he is. He's unbelievable. It's effortless the way the ball comes out of his hand at 95-96. He's that impressive. He's the best prospect we've got. It's fun to see."

Maybe they should stop messing around with Carl Pavano and Jaret Wright and bring Hughes to The Show?

February 22, 2006

Looking Ahead to 2006 - Second Base
by SG

Looking Ahead to 2006 - Second Base

The Yankees started 2005 with a second baseman that many of us figured was going to be a disappointment. Coming off a career year that still saw him only hit for an OPS+ of 94, the Yankees signed Tony Womack for 2 years and $4 million dollars.

I figured Womack would be bad. I did not know just how bad. Through the first 23 games of the season, Womack hit .282/.330/.329, which while not great, was about what any reasonable expectation would have had him doing. Fortunately for the Yankees and unfortunately for Womack, the Yankees themselves were losing more than they wanted to. This led to a big shakeup in Tampa Bay that had the Yankees call up Robinson Cano and move Hideki Matsui to center field and shift Womack to left field.

Cano started out slowly, hitting for just a Womackian OPS of .695 over his first 17 games. Thankfully, Joe Torre showed patience, something that I've often accused him of not doing with young players. From May 23 to June 28, Cano hit .281/.312/.504, almost never walking, but showing good power for a second baseman. He hit for a better average but less power in July, .319/.331/.466, and continuing his troubling trend of swinging at almost any pitch he saw.

August sucked, as the league and/or luck apparently caught up to Cano. Although he only K'd 16 times in 111 AB, compared to 14 in 116 AB in July, he hit just .207/.252/.261.

Cano made his adjustments and finished the season on an absolute tear, hitting .384/.395/.634. Combining this with roughly average defense, Cano was a 1.8 WARP player in 2005. However, since Womack did not do so well, the Yankee second basemen probably contributed about 1.0 WARP over the course of 2005. So what's ahead for Cano in 2006? To the projected WARPS!

In this table, RARP are the offensive runs above a replacement second baseman using a linear weights-based formula. FRARP is the defensive fielding runs above a replacement player based on last year.

Again, ZiPS projections are from Dan Szymborski at Baseball Thinkfactory. ZiPS says Cano will hit .286/.317/.438, with 16 HRs, which would be 15 runs better than a replacement second baseman. Combined with last year's defense, he's a 2.0 WARP player. I think Cano has the ability to be a better defender than that, if he can concentrate more and stop making careless mistakes. He has above average range and a very good arm for a second baseman. I read that he called Larry Bowa as soon as Bowa's hiring was announced to ask him what he should work on, which is a good sign that he wants to get better. Whether he actually will is anyone's guess.

Tango Tiger's Marcels predict a line of .298/.330/.463, with 13 HRs and 28 2B 443 AB. Far better rate stats, but with a fairly big reduction in his playing time, which ends up making him still 15 runs better than a replacement player. This adds up to 2.0 WARP player.

PECOTA is harsher than the first two, predicting Cano to be 13 runs better than a replacement player on offense. That would make him a 1.7 WARP player.

I love watching Cano, because of his potential. However, with his lack of plate discipline, his performance will always be dependent on his batting average. He can probably improve a little in this area, but if he doesn't walk that much, so what? A second baseman who can hit for good power and play average defense is still valuable. I'd like to see him hit the ball in the air a little more so he doesn't hit into as many double plays, and I'd like to see him develop a little more patience at the plate, but what's great about watching someone like Cano, and something that we as Yankee fans just don't get to see enough of, is that we have no idea how he will develop. The bulk of players the Yankees run out there have already made their mark and can be expected to perform in a certain way. When Mariano Rivera broke onto the scene as an unremarkable starting pitcher, did anyone think there was anything special about him? All of a sudden, he started blowing his fastball past the Mariners in the 1995 ALDS and developed into a dominant setup man in 1996. I'll just try to enjoy what Cano can do, not focus on what he can't do, and imagine the possiblities if things break just right.

Backing up Cano will be Miguel Cairo, who is a mediocre player who I still find enjoyable to watch. He plays decent defense around the infield and while he's not a good hitter, he is tough to strike out and will usually make a pitcher work. Cairo was -.3 WARP last year for the Mets, but I think if the Yankees are smarter about his usage this year and use him mainly to spot Cano against lefties, who he's hit for a .297/.338/.408 line in his career, he can be an asset. He was really bad against lefties last year, .191/.234/.270, but given his past history and the historic trend of the platoon advantage in baseball, this is very probably just a blip. Cano hit .270/.304/.358 against lefties last season, but struck out almost 50% more often than he did against righties. Interesting fact, all seven of Cano's successful bunts last year were against southpaws.

Anyway, Assuming Cano is worth about 2.0 WARP this season if he does what the projections say, he's about a one win upgrade at second base over last year's combination of him and Womack, assuming Cairo is right around replacement level in reasonably limited playing time. If Cano matches Marcel's rate stat predictions with the playing time predicted by ZiPS, he's a 3.5 WARP player. I like his chances to at least match his projections this year, and if stops pressing at Yankee Stadium(.252/.274/.378 at home, .335/.358/.525 on the road), he could beat them by a good amount.

February 21, 2006

Looking Ahead to 2006 - First Base
by SG

One of the biggest question marks going into to 2005 was what the Yankees could expect from Jason Giambi. Giambi really struggled to start the year, looking completely lost at the plate, drawing walks and doing little else, and drawing complaints from people who felt Ruben Sierra should be starting over him, and mocking derision from the anti-steroids crowd. "This proves that Giambi would have been nothing without steroids!"

Through May 9, Giambi went to plate 101 times. He had 15 hits, 3 HR, and was batting .195/.386/.325. On May 11, Giambi met with Brian Cashman and Joe Torre, where the subject of him going to the minors to get his stroke back was broached. This never materialized, but it certainly seemed to spark something in Giambi.

Now back in Oakland for a 3 game series where he was booed mercilessly, Giambi went 3 for 12, However, in the 7th inning on May 15th came what I still feel is the most important hit that Giambi had all season for himself, and possibly for the team. Giambi doubled down the right field line with two outs to give the Yankees a 5-4 lead. He slowly regained his hitting stroke, although not his power, hitting .311/.436/.422 over the next 30 games with just 2 HRs.

Then came July, and a monstrous month. .355/.524/.974, 14 HRs and 24 RBI as the Yankees kept afloat during a rough stretch of the schedule with 21 games against Cleveland, Boston, Texas, Los Angeles of Anaheim, and Minnesota.

Giambi fell off a bit in August, hitting .250/.448/.500 with 6 HRs, still very respectable. Giambi finished off the season hitting .244/420/.512 over September and October, and ended the year with a solid line of .271/.440/.535, with 32 HRs.

Did Giambi cheat earlier in his career? Without question. Did he cheat last year? I think he didn't, others may feel differently, and it is certainly their right to think so. He's done it before, which makes him more likely to do it again. While I think it's pretty silly to think that someone who had a pituitary tumor would take Human Growth Hormone and risk the potential return of a growth in their brain that could kill him, some feel it is the only way that Giambi could have recovered. While I think it's stupid that Giambi's "undetectable steroids" didn't work until July, took August off, and then returned to efficacy again in September, some may disagree. Those who feel it's possible that Giambi did what he did last year cleanly will not convince those who don't, or vice versa, so I'd rather not talk about it here. I will say that I hope he's clean, because it's a good story about second chances and a return from adversity, even if his difficulties were self-inflicted.

Getting off my soapbox, all that really matters for the Yankees on the field in 2006 is what Giambi will give them offensively and defensively. I'll look at the same set of projections that I did for Posada, which can be seen in the table below.

In this table, RARP are the offensive runs above a replacement catcher using a linear weights -based formula. FRARP is the defensive fielding runs above a replacement player.

Again, ZiPS projections are from Dan Szymborski at Baseball Thinkfactory. ZiPS says Giambi will hit .250/.401/.488, with 28 HRs, which would be 23 runs better than a replacement first baseman. Combined with last year's defense, that's a 3.0 WARP player.

Tango Tiger's Marcels predict a line of .248/.388/.464, with 24 HRs and 24 2B 407 AB, which would be 18 runs better than a replacement player. This adds up to 2.3 WARP player.

No PECOTA specifics again, but it's more favorable than the first two, predicting Giambi to be 38 runs better than a replacement player on offense. That would make Giambi a 4.7 WARP player.

The last row is a little something I'm calling 'What If?' What if Giambi hit how he did from May 15 through the end of the season over a full season? That would be a .288/.454/.584 line, with 46 HRs and would make him 73 runs better than a replacement first baseman offensively. That's pretty unlikely, although I'd love to see it.

Giambi gave back 4 runs of his value on defense last year. If he's playing more often in the field this season, that may go up, but his improved offense when he plays first probably makes up for it, so I'll leave it as a -4.

I think that the projections for Giambi are low based on his lost 2004. Giambi was a 5.5 WARP player last year. I'm not sure what the plan is for how much time the Yankees are going to use him at first, but I'd guess they would hope to get 120 games out of him at first, and another 30 at DH.

Tino Martinez has announced his retirement after 16 seasons to take an announcing job at ESPN. Tino will always be remembered as one of the faces of the Yankees 1996-2000 run. Overrated by many Yankee fans and underrated by many statheads, he was a solid player whose time has come. I'm glad he got to retire as a Yankee since it seemed to be important to him. If not for his HR streak early last year, the Yankees probably miss the playoffs. Outside that streak he was pretty bad though, so I think he picked the right time to go. Tino provided .5 WARP last year over 131 games and 303 AB.

I'm going to take a guess at 6.0 WARP for Giambi this year, which would keep the Yankees around the same level as they were in 2005. Despite Giambi's hot hitting from May to October, he's going to be 35 and has had enough injuries that getting certain production out of him is going to be an open question.

Andy Phillips will back up Giambi at first base, but it's tough to guess how much playing time he will see as a 1B. From what I've read from Joe Torre, he won't play much DH at all. He's probably right around replacement level, maybe a bit better, so I wouldn't expect him to skew the numbers either way.

February 19, 2006

Looking Ahead to 2006 - Yankee Catchers
by SG

With pitchers and catchers now in camp, I figured I should do some kind of previews for the Yankees. I'll start with the catchers.

Back in November, I took a look at Jorge Posada and his contract. While Posada showed signs of decline last year and has a ton of money due to him over the next two years, I still have some decent hopes for a slight rebound next year.

Let's look at some projections for what Posada may do in 2006.

In this table, RARP are the offensive runs above a replacement catcher using a linear weights -based formula. FRARP is the defensive fielding runs above a replacement player.

ZiPS says Posada will hit .254/.358/.424, with 18 HR and 25 2B in 465 AB, and which would make him a 3.3 WARP player.

Tango Tiger's Marcel predicts a line of .264/.366/.443, with 19 HRs and 24 2B in 451 AB. This would be an improvement of about 3 runs over last year's performance, or about 1/3 of a win and woud make him a 3.6 WARP player.

I won't post PECOTA's specifics since it is subscriber-only, but it predicts Posada to be a 4.0 WARP(wins above replacement) player.

These WARP numbers are all factoring the offensive projections above and then using last year's defensive numbers. While Posada scored fairly high last year in defense and should be expected to give some of that back if you look at his past defensive performance, the Yankees have brought in Tony Pena to work with their catchers this season and he was one of the best defensive catchers of his time, so that may mitigate an expected return to mediocrity.

Gone, and not missed, is John "Bad Flash" Flaherty. How bad was Flaherty last year? -1.1 WARP. His replacement is Kelly Stinnett, who's biggest selling point is that he's not Flaherty. Stinnett was .6 WARP last year, so if he can repeat a similar performance, his signing has the potential to be close to a 2 win upgrade. Given his age (36) and his career line of .239/.320/.390, this is far from a given. I can't find any scouting reports on Stinnett, but he rated as an average defender last year. Stinnett is a much better hitter against lefties, so even though he has caught Randy Johnson before it is probably best for the Yankees to avoid falling into the personal catcher trap and let him see most of his playing time against southpaws. Early indications from RJ and Posada are that they intend to work together this year, so that's good news on that front.

It's very thin after this, as the only other catcher on the 40 man roster is Wil Nieves, who was acquired for Bret Prinz last season and put up a line of .289/.312/.395 in AAA Columbus last year, and will be 29.

One signing of note was Ben Davis, who was invited to spring training on an NRI. Davis was a very highly touted prospect at one point, breaking into the majors at 22 off a .308/.386/.512 season in AAA Las Vegas in the hitter happy PCL. He's probably best-known for his bunt single with one out in the eighth inning that denied Curt Schilling a perfect game back in 2001. Davis is 29, and has to this point put up the awful career line of .237/.306/.366, but he may be a better option than Nieves if catching depth is ever needed.

Anyway, assuming a slight rebound by Posada as per Marcel and PECOTA, and with the move from Flaherty to Stinnett, it would appear that the Yankee catchers are about a two win upgrade relative to last season.

And now for something completely different. Potch posted a link to a lineup toy that lets you find the RC/27 for any group of 9 players. I contacted the developer and got the code and tweaked it to work with ZiPS and Marcel projections for 2006. It chooses the order for you, but you can play around with various combinations of players and see what group of 9 players the Yankees should be running out there most days. You can even use it to see how much of an upgrade a real DH would be over Phillips/Bernie.

Here's the link for anyone who's interested.

February 17, 2006

Spring Training Notes - 2/17
by SG

Well, pitchers and catchers have reported, which means no shortage of puff pieces in the New York virtual fish wraps.

Newsday has the breaking news that Yankees signed Scott Erickson, who was last decent in 1999. There's no problem with accruing depth, but I count about 15 pitchers I'd rather see starting in pinstripes in 2006.

1) RJ
2) Moose
3) Wang
4) Chacon
5) Pavano
6) Wright
7) Small
8) Villone
9) Matt Desalvo
10) Sean Henn
11) Jorge De Paula
12) Tyler Clippard
13) Stephen White
14) Darrell Rasner
15) Scott Proctor

If 11 of these guys get injured, I hope Erickson gets a chance.

Speaking of Randy Johnson, he's looking for a better year in 2006.

"Seventeen wins is good," he said, "but I expect more this year."

So do we Randy. So do we.

Mike Mussina is looking for a contract extension. I have no idea what Moose will do this year, but maybe looking for a contract will be a motivating factor for him. If he stays healthy and pitches reasonably well, I'd consider bringing him back at a reduced rate for a year or two unless it looks like the farm is ready to graduate a few starters and he'd be blocking anyone.

I'm still not buying the hype for World Baseball Classic, but maybe as it gets closer I'll feel differently. And Ozzie Guillen should just shut his mouth.

February 16, 2006

Sports Bloggers Live Interview
by SG

I did an interview on Sportsbloggers live which can be listened to here for anyone who wants to make fun of my voice. Nothing too earth-shattering here, it's about 4 minutes in.

Update: Here's the Direct Link to my piece.

February 15, 2006

Newsday: Pavano hurting again
by SG

TAMPA, Fla. -- The Yankees officially open spring training when pitchers and catchers report today, yet they already have their first injury concern.

Yesterday, the team told Carl Pavano, coming off a disappointing, injury-plagued first season in pinstripes, that a back specialist has recommended he should not throw off a mound for 10 to 14 days, putting his status for Opening Day in question.

Like him or not, Pavano is a better option than either Jaret Wright or Aaron Small, so the Yankees better hope this is just a minor setback. - Verducci: Jeter, not Damon, should bat first in Yankees' lineup
by SG

Slice it any way you want, but Derek Jeter is a better leadoff batter than Damon. The problem with the 2005 Yankees wasn't the top man in the batting order -- they led all of baseball with a .379 OBP in the leadoff spot; Damon's Red Sox were sixth at .363 -- it was that GM Brian Cashman foisted Tony Womack on manager Joe Torre, who was foolish enough to give Womack 146 plate appearances in the two hole. Womack returned a .529 OPS in that slot, ranking 73rd among the 74 players with 100 plate appearances hitting second. (Ruben Gotay of Kansas City was, insult of insults, worse than Womack.)

Interesting article by Tom Verducci. It will never happen, but I'd love to see the Yankees not get pressured into batting Damon leadoff just because of his reputation.

My ideal lineup would be:


Then, the Yankees could satisfy Michael Kay's desire to have a "Circular Lineup" and "Two Leadoff Hitters."

February 12, 2006

Diamond Mind Projections part two, PECOTA
by SG

I hope most of you enjoyed MGL's interview and subsequent discussion in the last entry. Thanks again to MGL for his time, and I heartily recommend picking up The Book. You'll learn a bit about baseball, and help support Retrosheet.

Last week, I posted some WARP (Wins above Replacement Player) and Diamond Mind simulation results based on ZiPS. For those who don't like ZiPS, thankfully there are other projection systems out there. I ran through another 100 Diamond Mind simulations using Baseball Prospectus's PECOTA projections, and the results were far more encouraging for Yankee fans.

Team W L RF RA DIV WC DIV% WC% Made% High W High L
American League
New York(A) 91 71 879 785 53 19 52.5% 18.5% 71.0% 106 80
Boston 87 75 852 767 33 14 33.0% 13.8% 46.8% 107 73
Toronto 83 79 767 749 13 6 12.5% 5.5% 18.0% 100 68
Baltimore 75 87 747 804 2 2 2.0% 2.0% 4.0% 92 58
Tampa Bay 70 92 698 826 0 1 0.0% 1.0% 1.0% 90 53

Minnesota 84 78 742 711 44 2 44.0% 2.0% 46.0% 104 72
Cleveland 83 79 787 762 38 3 37.5% 3.0% 40.5% 98 62
Chicago(A) 79 83 743 764 11 3 11.0% 2.5% 13.5% 99 65
Detroit 78 84 757 785 8 0 7.5% 0.0% 7.5% 92 62
Kansas City 68 94 692 818 0 0 0.0% 0.0% 0.0% 86 52

Oakland 94 68 781 655 75 11 75.0% 11.0% 86.0% 109 77
Los Angeles(A) 84 78 729 710 9 17 8.5% 16.8% 25.3% 97 71
Seattle 84 78 758 741 13 18 12.5% 17.5% 30.0% 98 60
Texas 79 83 830 845 4 6 4.0% 6.3% 10.3% 96 60

National League
Philadelphia 89 73 790 705 46 18 45.5% 17.8% 63.3% 101 70
New York(N) 89 73 794 741 44 16 43.5% 16.0% 59.5% 107 68
Atlanta 81 81 763 759 11 8 10.5% 7.5% 18.0% 98 61
Washington 76 86 695 744 1 1 0.5% 1.0% 1.5% 88 62
Florida 70 92 689 783 0 0 0.0% 0.0% 0.0% 86 55

St. Louis 93 69 759 645 74 13 74.0% 12.5% 86.5% 109 79
Milwaukee 83 79 745 727 11 10 10.5% 10.3% 20.8% 104 66
Chicago(N) 82 80 741 730 7 13 6.5% 12.5% 19.0% 95 69
Houston 79 83 725 754 4 5 4.0% 4.5% 8.5% 101 65
Pittsburgh 78 84 716 740 5 4 5.0% 3.8% 8.8% 94 63
Cincinnati 74 88 774 849 0 2 0.0% 1.5% 1.5% 91 60

Los Angeles(N) 93 69 808 695 87 3 86.5% 2.5% 89.0% 106 77
San Francisco 82 80 758 738 10 8 9.5% 8.0% 17.5% 100 64
San Diego 79 83 740 744 3 1 3.0% 1.0% 4.0% 95 57
Arizona 76 86 763 810 1 1 1.0% 1.0% 2.0% 89 67
Colorado 67 95 775 912 0 0 0.0% 0.0% 0.0% 81 57

The Yankees improve by 6 wins and score about 50 more runs a season. I'll take that.

I won't do any more of these until towards the end of spring training, when rosters are more set and we can sort out the lineups and pitching staffs more accurately.

In other news, the Yankees and Shawn Chacon avoided arbitration, which I was happy to hear. While I have a lot of concerns about Chacon's performance next year, I think risking acrimony by going to arbitration doesn't make much sense. Let's hope as his memory of pitching in Colorado gets further away, Chacon can continue to succeed. MGL discussed Chacon on a thread on Baseball Think Factory, and said the following:

I have Chacon projected to be a good (better than average) pitcher. I think that Pecota's projection for him is flat out not even close, but I could be wrong. I have a lot of respect for their projections overall. While it is indeed tricky to project players coming from the Rockies, even using proper park adjustments, there is one thing that no forecasters do that I do, which will make a big difference in the projections. That is to adjust for the hangover effect for Rockies players. IOW, their road stats while playing for the Rockies need to be adjusted quite a bit as well as their home stats. When Rockies players switch teams (or go from another team TO the Rockies), their road stats (not including their road stats in Coors Field) go up substantially (or down substantially if they are GOING to the Rox). If you don't adjust for that, you will substantially under-project players like Chacon.

So basically I think that Chacon is a pretty good pitcher (and should post around a 4.50 ERA for the Yankees) and that Small sucks. Wright is not half the pitcher that Chacon is. I think that he (Wright) is terrible as well.

The Yankees made a minor move that I think was very smart, picking up Darrell Rasner on waivers from Washington. Rasner was rated as the National's 8th best prospect last year, but he profiles as a fifth starter/middle reliever. He is not a great prospect by any means, but he adds some pitching dept to the organization, and is only 25 years old.

Pitchers and catchers report in 3 days. How cool is that?

February 8, 2006

An Interview with MGL
by SG

Mitchel Lichtman, better known in the on-line baseball community as MGL, is one of the foremost experts in the field of sabermetrics today. He is best known as the inventor of UZR (Ultimate Zone Rating), the most advanced form of defensive statistical analysis currently available. Although he now consults for a major league baseball team and UZR has become proprietary, he is still willing to share some of the results every year as well as his other methods of player evaluation. MGL has collaborated with Tom Tango and Andy Dolphin to write a new book about baseball strategy called The Book, about which more information can be found here. MGL agreed to do a collaborative interview with Replacement Level Yankee Weblog and to discuss the upcoming book as well as some general Yankee questions.

1) Assuming no more roster moves of any significance, how do you see the AL East playing out?

MGL: Despite some major roster moves and some serious spending by Toronto, unfortunately for them and for their fans, I expect the final standings to be similar to last year and to recent years.

With Boston and the Yankees spending money as they can afford to do and still having a core group of excellent players, they will be the powerhouses of the division again, barring the usual, such as major injuries, major bad luck for them, or good luck for one of the other teams.

Toronto will have excellent pitching on the strength of the best one, two combination in baseball, and one of the best closers in baseball, assuming they all remain reasonably healthy. Their offense, however, and infield defense, leave a lot to be desired, for a contending team at least. They got rid of their only offensive superstar in Delgado, and while Glaus can hit a bunch, he appears to be a defensive liability. Adams is a terrible defender and moving Hill from third to second will not help either. Overall, I project them at 87 wins and third place, which is not bad actually. Maybe it is considering how much money they spent – I don’t know. If any of their pitchers breaks down, especially Halladay or Burnett, or even Ryan, they could easily struggle to post 80 wins.

Tampa, despite some people’s optimism, is still not a good team. We should, however, see major improvement in their w/l record, although much of that will come from the “plexiglass” principle (essentially regression toward the mean) alone. Their offense and defense will be sub-par again, though not as bad as last year. Their pitching will be pretty bad again. The only remotely decent pitcher on the staff is Kazmir, who is and was the real deal, for all you Met fans (while Zambrano never was). Baez was a below-average closer and rest of the pen is still pretty weak. Look for them to win 76 games at best.

Baltimore could be the surprise team of the division. After a hot start last year, they basically were unlucky for the season and should have won 80 games or so based on their underlying statistics, as well as their player projections going into the season. They have a core group of outstanding players (only one of whom is a household name) in Tejada, Roberts, and Mora, and Bedard and Cabrera are well-above average pitchers. Look for the O’s to win 86 games, good for fourth place.

As I said, Boston and the Yankees are by far the cream of the crop in the division. Boston has incredible depth in their starting pitching. How many teams would love to have Clement and Arroyo batting for the 5th slot in the rotation? Heck, on a few teams, either of those guys would be the ace of the staff. Of course if Beckett does not remain healthy and Schilling does not regain at least some of his pre-2005 form, their pitching will not be nearly as good as I expect it to be. Also, having Timlin rather than a healthy Foulke as their closer, is not going to help their pen. I expect them to win around 94 games, assuming everyone stays reasonably healthy, good enough for second place in the division and an excellent shot at the Wild Card again.

The Yankees have a powerhouse of a lineup again this year. The only weakness is at DH, and even that is not too bad. With a little luck, they could score 1000 runs in 2006. I think that 900 is a slam dunk. Their starting pitching, one through four, is pretty good, if RJ stays healthy and keep his velocity consistently in the mid 90’s. Mussina, while well on the downslope of his great career, should still be able to post good numbers, and while Wang and Chacon are not as good as they pitched for the Yankees last year (especially Chacon), they are still pretty decent pitchers. And of course, Mo remains one of the best closers in baseball. Pavano and Wright are the stepchildren in the back of the rotation this year. I much prefer Pavano. I think Wright is a terrible pitcher who had a good, flukey year for Atlanta in 04, perhaps helped by the tutelage of Mazzone. I expect the Yankees to once again lead the division, this time with 96 or 97 wins.

Of course, the better you are, the more hurt you are by injuries and other unforeseen circumstances, so you never know. The plexiglass principle also applies when projecting player and team performance.

2) What do you see going forward for Robinson Cano and Chien-Ming Wang? Andy Phillips looks like he may be used in a fairly significant role in 2006. How do you think he can do based on his minor league performance? What do you think about the Yankees farm? Which home grown players do you see benefiting the team the most this next year? 2 years down the line? Is Phillip Hughes too far away for you to give an opinion on?

MGL: I know that they are popular discussion topics for fans, but the minor leagues and prospects are not my forte. That being said, I think that Cano is the real deal and should help the Yankees for years to come. His hitting was surprisingly good last year, and should regress a little, given his modest MLE’s in the minors. There are some differences of opinion as far as his defense goes. UZR has him as slightly below average in 05. I have heard others say that he is not that bad. If his defense ends up being above average, he is quite a valuable commodity. If it ends up being sub-par, then he won’t be so valuable.

As I said, I think that Wang is the real deal too, and ought to remain an above-average starter, which is a very valuable commodity, given that league-average FA starters command 8 mil a year or so. He was very good (his MLE) in the minors so it was not surprising that he would have pitched so well in 05.

Andy Phillips ought to be playing somewhere in the majors. He is a good hitter. I don’t know much about his defense and I don’t know whether his natural position is second or third. He is much more valuable if he can play second (adequately) of course. If he is slated to play second, the Yankees would need to trade either him or Cano, both of whom ought to be powerful trade bait I would think. If he is touted for third base, then they need to move A-Rod back to SS (where he belongs) and Jeter to somewhere else (where he belongs). I don’t see that happening any time soon.

Just scanning my MLE’s, the Yankees appear to have a fine crop of hitters at Columbus but not at Trenton.

3) On the major league side, there is obviously concern with the ages of several Yankee players. Are there any warning signs with Jason Giambi and Gary Sheffield that we should be wary of? How do you see Johnny Damon holding up over the next four years, and how much did the Yankees overpay for him?

MGL: Because of the steroid situation, Giambi will always be somewhat of a question mark. He works out at a gym a few minutes from where I live, and from what I understand he is still (sans PED’s) in great shape (and a heck of a nice guy they say).

Sheff’s defense is not good at all, but he can still hit the heck out of the ball. There really is no such thing as a “warning sign” as far as projecting a player to precipitously or suddenly decline in performance. Players over the age of 26 generally (and without chemical enhancements) start to decline overall offensively. Over the age of 30, they decline quite a bit each year, and over the age of 35, they decline even more each year.

BTW, players generally decline defensively starting from a very early age (early 20’s), much like the aging curve for their triples rate. First base defense might be an exception to that rule. Anyway, we should still see Sheff and Giambi hitting well-above average for their positions, although not nearly as well as they hit 3 or 4 years ago.

I don’t know how well Damon will or will not “hold up.” Fast players tend to age better than slow ones, which is a good sign for him, although almost all CF’er s are fast, right? His defense appears to be not nearly as good as it once was (and of course his arm is a rag, although the Yankees should be used to that in CF), which was one reason why Boston let him go I think. He can still hit well for a CF’er and runs the bases well. As with most long-term contracts, you pay for current expected performance and hope that a player’s decline with age (assuming he is past his prime), and chance of major injury, is balanced by salary inflation.

How much did the Yankees overpay for Damon? That’s a leading question! Given that he projects to be about 3 wins above replacement this year, that is worth about 8-10 mil on the open (FA) market for a CF’er, I suppose. Personally, I would not have paid him more than 7 or 8 mil a year (my personal limit per marginal win is around 2 mil). And as I said, although long-term contracts tend to get balanced by inflation, 4 years is a little long for a 32 year-old. By Yankee standards though, 13 mil a year is not that bad, maybe about 40-50% more than he is actually worth in the open market.

4) The Yankee defense has been a sore spot for quite a while. How do they project defensively in 2006? Can you share any 2005 UZRs for any of the Yankees? Is there any potential beneficial effect on the LF and RF due to the upgrade from Bernie Williams to Damon?

MGL: First, I don’t think that one player’s defense really impacts another.

As I said, although Bernie was terrible in CF, Damon may not be that good himself. I don’t expect the Yankee defense to be much better than last year, especially with Giambi now full-time at first (which makes up for the Bernie to Damon upgrade in center). But who cares about defense (other than the pitchers’ agents) when you have that kind of offense?

Pound for pound, a run saved on defense (or pitching) is a smidgen better than a run gained on offense, but for all practical purposes, they are one and the same. Defense and offense combined, the Yankees have by far the best starting lineup in baseball. Although they don’t have the best pitching, they are probably the best overall team in baseball as well.

Here are some 2005 UZR’s for the Yankees. Keep in mind that one year UZR’s don’t tell you a whole lot about a player’s true defensive value (or his defensive projection). For that, you want multi-year UZR’s. For example, Giambi has generally been quite minus in the past. A-Rod is usually much better than 2005 (historically he is a GG caliber SS), Damon has been better, and Sheff is usually not that bad. Posada is probably not as good as his 05 defensive numbers either.

Giambi 0
Cano -1
Jeter -8
A-Rod +1
Matsui -2
Bernie/Damon -12/-6
Sheff -20
Posada +4

5) The subject matter of your new book, The Book, looks very intriguing to me, particularly the sections about bunting and optimizing a lineup. Have you analyzed any managers' tendencies in this? If so, how would you rate Joe Torre in this regard? I also wonder if bullpen usage is something you tackle in The Book?

MGL: Yes, we have a nice chapter or two on bullpen usage. We mostly talk about the best times to bring in your best reliever (your closer), as well as the use of LOOGY’s, ROOGY’s, and “platooning” pitchers in general. We talk about other things regarding bullpen (and starter) usage as well. Keep in mind that we go into quite a bit of statistical detail throughout the book. It is not “light bedtime reading.”

From time to time we mention a particular manager or two, but we don’t discuss manager tendencies per se.

I used to think that Torre was one of the better strategic managers in baseball, but I have since changed my mind for various reasons (not the least of which was when he started Enrique Wilson twice in the ALCS versus Pedro because Wilson had “done well” against him in the past). He is probably not a whole better or worse than any other manager, which would give him a C- in my class. One thing I do like that he tends to do in the playoffs is use Rivera for two-inning saves. He doesn’t do that enough in the regular season though. No managers do.

6) Torre is obsessed with pitcher/batter matchups when making out lineups. Most of us think this is folly based on the small sample size. Is there any predictive value to a particular type of batter/pitcher matchup? If so, what kind of predictive value do past results give to future performances?

MGL: See my comments in the last section. Funny, I didn’t know you were going to ask that when I mentioned the Pedro/Wilson thing! I won’t answer your question about batter/pitcher matchups in general. I’ll let your readers find out by reading The Book. The analysis we do is fascinating, if I may say so myself.

By the way, if any readers are uncomfortable with me touting The Book, please keep in mind that 100% of my proceeds from book sales are being donated to Retrosheet. So if you are on the fence with regard to buying the book, if you want to support Retrosheet (, a wonderful volunteer organization that compiles data for research, get off the fence!

7) What makes you believe UZR is a better defensive metric than other's metrics like standard Zone Rating (and all the variations of it), David Gassko's range, David Pinto's PMR, and Baseball Prospectus's Davenport fielding numbers? Is there such a thing as one, 'best', defensive metric?

MGL: While all of those you mentioned are good (among others not mentioned), some are better than others, by virtue of the data used and the breadth and depth (essentially the rigor) of the methodology used to “crunch” that data. For example, BP’s DFT’s do not use PBP data, nor does Gassko’s “Range.” Or Bill James’ Defensive Win Shares. Although the “Range” methodology is excellent (DFT’s are a “black box” so I can’t comment on their methodology), the value of a non-PBP based defensive metric is limited. Pinto’s PMR (and a couple of other PBP-based metrics) suffer from some methodological problems (or weaknesses at least) in my opinion.

The current incarnation of UZR (which is different from the “published” version), is extremely sound and uses very granular data (from STATS Inc.). It is very good in the “adjustments” it makes (the handedness of the batter, the park, the speed of the batted ball, the base/outs state, etc.).

STATS Zone Rating is fine. It just doesn’t make very good use of the data. For example, a soft ground ball in one location is treated the same as a hard one, and a fly ball in the outfield is treated the same as a line drive (I think). As well, all balls in a fielder’s zone are treated the same regardless of their actual location (5 feet from the fielder or 15 feet). There are no park adjustments either. That makes a big difference in parks like Colorado, and in left field at Fenway, for example.

All of that makes ZR sound bad, but it really is quite good, especially when it is expressed as runs saved or earned so that a person can combine it with offensive RC or linear weights. All of those things that ZR does not account for (which UZR does) tend to even out in the long run, so the larger the sample, the more ZR will agree with UZR and other good PBP-based metrics. There is still much work to be done with defensive metrics, but most of the work is in recording more granular data (like the exact speed, trajectory and “hang time” of batted balls, and the location of the fielders before the play emerges) and making the best use of that data. I think that we are 90% of the way there though, at least as far as UZR and other good PBP metrics are concerned. Some people would disagree with me.

8) What flaws, if any, can you point out in your own metric as well as in the other aforementioned metrics?

MGL: I think I addressed some of that already. The accuracy of the data being recorded is always a problem. Also, not knowing the exact position of the fielders is a problem. Discretionary plays, like a fly ball that more than one fielder can easily catch is problematic as well.

Most of the metrics can make better use of subjective information, which is not usually included in the data. For example, knowing that a certain play was “easy and routine” in the opinion of the observer/data recorder, or “spectacular,” or “could not have been gotten by anyone,” is extremely useful information above and beyond the location of the batted ball, and even the location of the fielder as well. None of the metrics mentioned herein utilize such information as far as I am aware (although there are proprietary systems that do). I am currently working on adding some subjective data to UZR (STATS provides a “judgment” on all plays made).

The ideal system or “holy grail” of defense metrics really is a combination of objective “data crunching” (using very granular data of course) and subjective observation (on a play by play basis – I am not taking about traditional “scouting”). That is true for offense and even pitching, BTW.

9) For offensive statistics, which publicly available statistic do you think encompasses the best combination of value-adding features of batters? I know you prefer your version of linear weights(SLWTS) which incorporate baserunning and defense. Is there the publicly available data to construct a reasonable facsimile of those? Also, what is your preferred method of evaluating pitchers, and do you treat relievers and starters separately? If you had to use publicly available stats without calculation for evaluating hitters and pitchers, what would you choose?

MGL: Let me start by saying this. I don’t know what you mean by “value-adding features.” Whenever I talk about a metric for evaluating batters or pitchers (or for defense), I am talking about a player’s “true talent.” IOW, it answers the question, “Who would you rather have on your team (a generic team) in the future if you were interested in winning the most games you could (not counting playing time or chances of injury)?” As far as value in the past or in a certain context, as in evaluating past performance for an MVP-type award, I don’t know what the best statistics are, and that doesn’t interest me anyway. So keep that in mind as I answer these questions.

Let me start with the last one. Of course, these days when you say “publicly available,” that could mean on Baseball Prospectus or it could mean on ESPN or Fox Sports. If you mean “mainstream publicly available,” which I assume you do, then ERA and OPS are the best stats to use. Anything else (mainstream) adds nothing to the argument. Keep in mind a few things if you use raw OPS or ERA. One, one year does not a projection (or an estimate of “true talent”) make. IOW, sample size beware, and learn how to do some rudimentary “regression toward the mean.”

Two, always place raw numbers into context with league averages. For example, in 2001, the strike zone was changed and league average OPS and ERA went down. The players’ overall talent did not change.

Three, even without any numbers, mentally do some park adjusting if you can. A player who plays half in games in Colorado will not have nearly the same raw numbers as an equivalently talented player who plays half his games in Dodger Stadium. We all know that though, right?

Four, you really should compare players at the same defensive position or mentally make an adjustment for that position. For example, the average first baseman hits 20 runs per year or around 100 points better in OPS. So if you are comparing one player to another and they play different positions make sure you at least mentally make some kind of positional adjustment. As a rule of thumb, first basemen and corner OF’ers are 10 runs better per year at hitting than the league average batter, CF’ers and third baseman are average, second baseman are 5 runs worse, SS are 10 runs worse, and catchers are 15 runs worse.

When comparing pitchers, you must also compare apples to apples. That means that you should not use any metric to compare a reliever to a starter, “straight up.” For example, the same pitcher when relieving (one inning or so), is likely to post an ERA almost one run better than when starting, with some caveats. We talk about that in The Book, BTW. Also, when evaluating relievers, forget about “strand rates” and all that nonsense (again, assuming you are evaluating for projection purposes). Adjusted (for park and league) ERA is still the way to go for relievers as well as starters. Of course, I am still talking about “publicly available” stats. There are better ways to evaluate and project pitchers.

For example, the best way to evaluate pitchers is to use ERC which is “component ERA,” or translating a pitcher’s component stats (s,d,t,hr,bb+hp,so) into an ERA looking stat, using a linear weights formula or better yet, something called Base Runs (linear weights works best when a player does not impact himself; for teams and pitchers, where the player or team impacts itself, Base Runs works better).

That takes away the “luck” associated with ERA. There is no evidence that one pitcher is able to “control” his ERA better than another – only that one pitcher is better able to “control” his K, BB, HR, and to a lesser extent, his singles, doubles, and triples rates. That is why ERC works better for estimating true talent or projecting future performance than ERA. In the long run, ERA will approach ERC as the “luck” associated with ERA tends to even out.

Since regression to the mean is an integral and important part of estimating a player’s true talent and projecting his future performance (we also discuss that in The Book), one way to do that is to use a pitcher’s DIPS ERA or FIP rather than actual ERA or ERC. The less data you have, the more you should consider using DIPS or FIP to project and compare pitchers, since DIPS and FIP are essentially an extreme way of regressing portions of a pitcher’s component stats. If you want to know more about DIPS and FIP, do a search on the web. It is too complicated for me to explain herein. For batters and offensive metrics, you will have to do the “regressing to the mean” yourself (in order to take sample stats and “convert” them to a projection or an estimate of true talent).

For batters, and for offense, if you have access to more than OPS, you want to use a linear weights type formula. In my Superlwts, I simply add defense and baserunning (and a few other things) to an offensive linear weights formula. Also, my linear weights formula is a little more rigorous than the traditional ones (for example, it gives different weights to ground outs and fly outs depending on the handedness of the batter and includes RBOE’s).

If you only have access to OPS and say ZR for defense, it is tough to “add” the two together as they are not of the same scale. That is why I prefer all metrics to be in “runs” above or below average. That way you can add them all up to come up with a total value for a player. As a rule of thumb, keep in mind that good hitters (say an OPS of over .800) are usually around 10 to 20 runs above average per season, great hitters are 20-30 runs above average (OPS of .850 to .900) and superstars (OPS of over .900) are 30 to 60 runs above average in offense. (Bonds in his heyday is in a class by himself.) The superstar class is the likes of Pujols and A-Rod. On defense, the best players are 20 runs above average (and the worst are 20 runs below average) per season. And good ones are 10 runs better or worse than average. So for example, Rolen is around 30 runs in offense better than average and another 15 to 20 on defense. Jeter is 10 to 15 runs above average in offense and -10 runs in defense.

As you can see, the spread in talent in offense is about twice that of defense. BTW, baserunning adds or subtracts another 3 or 4 runs a year to a player’s overall talent for the best and worst baserunners. Outfield arms (and turning the GDP for IF’ers) are worth another 4 or 5 runs at best (or worst).

10) The Yankees are talking about batting Johnny Damon leadoff even though he projects to have the 7th lowest OBP on the Yankees by ZiPS. Is there such a thing as an optimal lineup construction, a formula or procedure for ordering players, that is universally applicable to all teams in MLB? Does staggering lefties and righties in the lineup help nullify platoon advantages, and is that something that is analyzed in The Book?

MGL: Yes, we have a great chapter on optimizing lineups. Since the exact order of the lineup rarely matters that much, you always want to avoid having two lefties batting next to each other, if at all possible.

As far as optimizing a lineup based on the exact composition of the players, yes there is an exact “formula” or algorithm you can use to do that. It balances leveraging each player’s strengths and weaknesses (power, OBP, etc.) with the fact that each lineup slot gets about .1 more PA per game than the one below it. If a manager (or fan) does not want to use the exact “formulas” we also give some “rules of thumb” for optimizing a lineup.

11) In a strict sabermetric sense, if you're the manager, and Barry Bonds comes up bat, do you ever walk him? What situational characteristics must be present for your decision to be more or less absolute?

MGL: Again, we have an excellent chapter on when and when not to issue the IBB. With Bonds there are times to issue the walk and times to not. In general, he is walked too much, at least over the last few years. Most managers are about using risk-averse (or “getting fired or lambasted in the media-averse”) strategies rather than the “correct” (the ones that give your team the best chance of winning the game without sacrificing future wins) ones. Of course, who knows the correct ones in the first place?

Like most anything else, all you can do is ask the experts. You can’t expect a manager to know what the optimal strategies and decisions are off the top of his head. That is impossible unless he is some kind of a savant. In order to figure them out, it usually takes some heavy statistical analysis and research. The problem of course, is that managers generally think they know the correct strategies, which is ludicrous of course, and scoff when someone suggests that they can find out if they really want to (like by reading our book).

It is a good thing for the consumer that most corporations are not run like baseball teams are run. Can you imagine the corporate CEO or manager of SONY trying to figure out the best way to manufacture a computer or television by intuition?

12) After all your years of research, what do you believe regarding the notion of 'clutch?' And in either case, what situations do you use to come up with that decision?

MGL: Another great chapter in The Book (are you sure you haven’t stolen a copy of the manuscript?). Andy Dolphin has done the best study on the existence of and predictive value of clutch hitting I’ve ever seen. His results are different than what most prior researchers arrived at.

13) Based on in-game managerial decisions, do you feel there are any managers in major league (or minor league) baseball that constitute a greater value added in terms of runs scored/allowed? Maybe a top 5, or bottom 5 list? Where would Torre rank? Is there even a way to quantify all managerial decisions based on runs added over average, etc?

MGL: Sure, you could quantify the value of managerial decisions, like the IBB, sac bunt, stolen base, use of the bullpen, etc. I’ve never done it and have never seen it done. It would take a lot of work and you would have to know the value of the various alternatives in the first place.

Most of that would show up in the runs scored and runs allowed so this notion that you can somehow evaluate a manager by the difference between his team’s actual and Pythagorean w/l percentage is nonsense. A few things like properly leveraging your relievers (like bringing in your best relievers when it counts the most and your worst relievers when it counts the least, regardless of the inning and score, although inning and score are related to leverage of course) will not show up (as much) in a team’s runs scored/allowed, but will impact their w/l record. But there is still too much noise in both a team’s actual and Pythagorean record to make any sense of it in terms of evaluating a manager. Maybe after 10 or 20 years you might find something.

As far as Torre is concerned, as I said, I used to think he was better than most, but now I am of the opinion (loosely based on fact and evidence) that he is middle of the pack, maybe a tad better than the average manager. I think part of the illusion that he is a good tactician comes from the fact that he has had such a good team over the last 10 years that he has had fewer opportunities than most managers to do stupid things.

In any case, I don’t think that the actual spread of manager talent amounts to more than a few runs a year. Of course I am talking about tactical things, not about teaching and motivating players, which I am sure are a big part (maybe even the most important part) of a manager’s skill set. On the other hand, if I were able to instruct a manager in the fine points of all of these things we are talking about and more, and they would listen, or they would simply read a book like ours, I firmly believe I could add 2-3 wins to a team’s w/l record at the drop of a hat. And you know what 2-3 wins is worth in today’s player market!

14) Of all the good-old 'traditional' baseball practices out there, like batting a speedy contact hitter first, etc., is there any one practice that truly makes you cringe when you see it happen?

MGL: There are lots. Without giving away too many of the book’s secrets, batting a poor but speedy (and good bunter or contact hitter) batter in the “two hole” is one. Sacrifice bunting a pitcher with first and third and one out is another. Sacrifice bunting a good-hitting (or even average-hitting) pitcher with one out and a runner on first only is a terrible mistake as well. There are lots more.

15) You have been working in an advisory capacity for a major league team for some time. What surprised you most about the inner workings of a major league front office? To what extent do you feel statistics are used by major league teams now? Besides offense and pitching, are they using defensive metrics? Are there inroads being made into new, under-appreciated skills in the current marketplace as the price of OBP has gone up?

MGL: Many, if not most teams, are using statistical analysis to some degree or another. Only a few are using advanced techniques and employ competent analysts and sabermetricians. Of course I don’t know for sure. This is conjecture on my part. More and more teams will of course do so in the future and some teams will be reluctant for a long time, for one reason or another.

There are some teams using advanced defensive metrics as far as I know. Not many though. The concept outlined in Moneyball and parroted ad nauseam in the media, that “OPB was the skill undervalued by teams,” and the thing that Beane and the A’s were able to take advantage of, is silly. As long as there are some teams that are better than others in doing anything, be it scouting, drafting, or analyzing player performance and strategies using advanced statistical techniques, there will always be players and situations that are undervalued and overvalued, and there will always be teams (the ones who are better than the other teams) that can take advantage of these things, and the market in general.

The fact that statistical evaluation of players and rigorous player projections are so valuable and so important simply means that if you can do those things well (which only a small minority of teams presently can) you will have lots of opportunities to take advantage of teams that don’t do those things well or at all. It has nothing to do with OBP or defense per se being over or under-valued in the market. That was an oversimplification and somewhat of a fiction made popular by a book and by an author who knew nothing about baseball or sabermetrics, as far as I know.

16) Can you tell us a little more about The Book, and your co-authors? How long has this project been in the works, and did any of the things you researched really surprise you?

MGL: I’ve talked a lot about the book already. I think it is groundbreaking. A little gory in the math department, but not so much that the average fan won’t appreciate it I don’t think. There are excerpts from every chapter on our web site ( and a brief description of the book itself.

My co-authors are Tom Tango and Andy Dolphin. They both have their own web sites devoted to baseball and other sports. They are brilliant sabermetricians and Andy is a brilliant mathematician/statistician as well. We have been doing research and working on writing the book for almost three years now I think. Writing a book like this is not as easy as some might think, especially when you can’t just sit down and “write away” like I am doing now. Practically every word in the book is based on hours and hours of sometimes painful research and analysis.

17) When and where can fans like myself, and anyone else who has enjoyed what they have read here, get your book, and how much will it cost?

MGL: Again, the web site is All the info is there including a link to purchase the book. The pre-order price is $16.95 plus $4.95 in S&H. The regular price once we start delivering is $18.95. We have free shipping on orders of 5 or more. I think the price is pretty reasonable.

The book has been a labor of love for us, and, as I said, 100% of my share of the proceeds is being donated to Retrosheet. Thanks for the opportunity to waggle my fingers while talking baseball, and good luck to the Yankees and all their fans.

February 6, 2006

Yankees vs. Red Sox - Positional Comparison
by SG

Chofo asked me to do a positional comparison of the Red Sox and Yankees. While I try to ignore the Red Sox as much as possible, and the last thing I want to do is attract Red Sox fans here, I'm suffering from blogger's block anyway, so here it is.

I wondered what would be a good way to compare the two teams. I wanted to try and choose a combination of stats that includes offense, defense, and baserunning, to give a more complete picture of how the two teams' positional players shake out. While I could have used Baseball Prospectus's WARP (Wins above replacement player) stat, I have a few problems with it. First of all, I'm not enamored with B Pro's defensive statistics. Secondly, I think that the fact that they set a replacement level player as both a replacement level hitter and a replacement level fielder is unrealistic. If a player can neither hit passably or field passably, they don't have any chance of being on a major league team for any amount of time. Lastly, I wanted to look forward to what may happen in 2006, not past history.

The idea of a wins above replacement stat is still intriguing to me, and thankfully Kyle, from the new blog OBP for you has sent me a spreadsheet with his own calculated WARP. He expounds on his system here, for those who are interested. I'm a big fan of linear weights as a method of player evaluation as I think it is the most thorough way to analyze a player's contributions to offense, and it incorporates more data than any other system (GIDP, HBP, SB, CS, etc.,).

Ergo, I took Kyle's spreadsheet and used the infamous ZiPS projections from Dan Szymborski at Baseball Think Factory to calculate rough projected WARP for 2006. The biggest flaw here is that I have no concept of how to project defense, so I'm using last year's defensive numbers (as calculated by Kyle using Zone Rating) as a proxy. Since Crisp is moving to CF I will adjust his numbers accordingly (move him from well above average in LF to average in CF). Also, since ZiPS does not predict playing time, I'm adjusting the PT allocation by my own guesses. Anyway, here's the 2006 Yankees vs. the 2006 Red Sox position players, by this convoluted, projected WARP.

Jorge Posada (Projected 2006 WARP: 3.3 )
Jason Varitek (Projected 2006 WARP: 3.7 )

Not much to say here except that they're the same age, and every year except last year Posada has been the better player. While I think Posada has a chance to bounce back a bit next year, I'll begrudgingly give the advantage to the Red Sox.

Advantage: Red Sox, by .4 wins

First Base
Jason Giambi (Projected 2006 WARP: 5.4) and Andy Phillips (Projected 2006 WARP: 3.0)
Kevin Youkilis (Projected 2006 WARP: 3.0) and J.T. Snow (Projected 2006 WARP: 1.4)

Bad defense and a bad 2004 have Giambi's projected WARP at 5.4. Given his age and past history, I wouldn't adjust this upwards at all. Snow and Youkilis's WARP total is misleading as it assumes 224 games. I won't talk about Phillips here, assuming he will DH more often than playing 1B. However, I think he will be a good defensive 1B based on what little I saw of him there and the fact that he moved there from more challenging positions (2B and 3B).

I'd give Youkilis his WARP of 3.0 assuming he'll get his 109 games in as per ZiPS. For Snow, I'll cut his 1.4 WARP in half, assuming he isn't getting more than 60 of the 115 games he's projected for (if the Red Sox are smart) and will probably see most of his action as a defensive replacement. So Sox total WARP at 1B is 3.7.

Advantage: Yankees, by 1.7 wins

Second Base
Robinson Cano (Projected 2006 WARP: 2.6 )
Mark Loretta (Projected 2006 WARP: 4.3 )

Cano's a young player who impressed Yankee fans last year. He's still got a ways to go to catch Mark Loretta. This is one instance where it's possible, as Cano is at an age where continued growth is reasonable. However, while I am very hopeful about Cano's future, I think he's got some growing to do. It could come this year, but it will more probably take a few years.

Loretta had a down year last year, but he was hurt, and very good in 2003 and 2004. He'll be 34 this season, which is a bit older for a 2B, but I think it's pretty likely that he will be more valuable than Cano this year, if he stays healthy. I'd be surprised to see the gap between the two this big though. I think Cano's going to hit for more power than ZiPS projects.

Advantage: Red Sox, by 1.7 wins

Third Base
Alex "MVP" Rodriguez (Projected 2006 WARP: 8.4)
Mike Lowell (Projected 2006 WARP: 4.6 )

All hail the reigning MVP. A disappointing defensive season costs Rodriguez about a win of value. I'd expect some offensive regression next year, but hope for defensive improvement to make up for some of it.

Lowell had a horrible year last year, but he's been a good player in the past and has a chance for a bounce back season moving to a bandbox. For some reason few are giving him much chance of bouncing back, but he's only 32. I think it's pretty reasonable that he hits around his 2002 level (.276/.346/.471).

Still a big edge for the Yanks.

Advantage: Yankees, by 3.8 wins

Derek Jeter (Projected 2006 WARP: 5.6)
Alex Gonzalez (Projected 2006 WARP: 2.8)

Jeter's better. Ignore the dolts at Fenway who tell you otherwise. He's a better hitter and a better base runner. He gives back some ground defensively to Sea Bass, who is a good defender but a pretty bad hitter. ZiPS is on crack with its .275/.326/.451 prediction, as he's bringing a career line of .245/.291/.391. I know he's going to a hitters' park from a pitchers' park, but he's a .695 OPS hitter on the road, compared to .669 at home.

Basically, if Gonzalez hits for a .777 OPS next year, the Sox will have gotten a steal.

Advantage: Yankees, 2.8 wins

Left Field
Hideki Matsui (Projected 2006 WARP: 4.9)
Manny Ramirez (Projected 2006 WARP: 8.1)

Matsui's not in Manny's league right now, although he's a good enough player. I was hoping that the 2004 Matsui was the real Matsui. That version was worth about 7.3 wins.

Ramirez's throwing arm last year helped alleviate some of the damage his poor fielding wrought. I'd guess he'll hit a little worse this year, but not appreciably enough to change this.

Advantage: Red Sox, by 3.2 wins

Center Field
Johnny Damon (Projected 2006 WARP: 3.2)
Coco Crisp (Projected 2006 WARP: 5.2)

One guy cost $52 million. One cost Edgar Renteria and $12 million.

Damon's a big upgrade in CF over what the Yankees had last year, probably over 5 wins. He projects to have the 7th lowest OBP of any of the starting lineup, but will bat leadoff.

Crisp already projects better than Damon, but who knows if he can handle a full gig in CF? He was a stellar defender in LF, but pretty bad in CF from both defensive metrics and from scouting reports.

Neither one of these guys can dent bread with their throwing arms.

Advantage: Red Sox, by 1.9 wins

Right Field
Gary Sheffield (Projected 2006 WARP: 5.8)
Trot Nixon (Projected 2006 WARP: 4.4)

Advantage: Yankees, by 1.4 wins

Sheffield's still a dangerous hitter. ZiPS projects him to fall off a lot more than I do, and I think his projection is low. I won't mess with it though since there's a decent chance of either a decline or some missed injury time.

Nixon stinks against lefties, but since his projection has him playing 113 games I won't adjust it, assuming the remaining games will be filled by someone awful like Gabe Kapler.

Designated Hitter
Bernie Williams (Projected 2006 WARP: 1.8) and Andy Phillips (Projected 2006 WARP: 3.0 )
David "2nd Place" Ortiz (Projected 2006 WARP: 9.2)

ZiPS says Phillips will play 115 games or so. That sounds reasonable assuming no other bats are picked up, splitting time at DH and first base. I would assume Giambi plays DH when Phillips is at 1B, which leaves Bernie with 40 games or so. While my heart hopes Bernie has a last hurrah in pinstripes before riding off into the sunset, my head says to expect very little from him. Hopefully if he is not up to the job, he doesn't keep getting undeserved playing time based on his past performance.

None of the massaging of numbers matters. This is a huge edge for Boston. Ortiz is a monster at the plate, and singlehandedly seems to prove that there is such a thing as clutch hitters.

Advantage: Red Sox, by 5.7 wins

The grand total shows the Red Sox's position players being 3.2 wins better than the Yankees. I now expect 100 comments about how stupid this is.

There's a long list of reasons that these may be way off, so don't go nuts about them. First of all, these are the same ZiPS projections that gave us those well-received Diamond Mind results, so of course they will show the Red Sox being better. Also, I think projecting defense is important. If Rodriguez improves from the -11 runs he played at last year to the + 10 runs he played at in 2004, that's two wins by itself. If Matsui hits closer to the 2004 version, that's 2 more wins. If Cano hits a little better and Loretta doesn't rebound, there's another win or two.

And remember, this doesn't include pitching, or the benches. This is just a comparison of the starting seven position players and DH.

Anyway, those are the numbers based on ZiPS. I may or may not do more of these with different projections depending on how they are received.

Thanks to Kyle at OBP for you and Dan Szymborski at Baseball Think Factory for the data and formulas used in this entry.

by SG

Even though the Yankees have a new pitching coach, their old one will still be joining them for spring training later this month.

Mel Stottlemyre, who retired following the end of last season because he was tired of the second-guessing coming out of Tampa, will be coming to Yankee spring training in a celebrity coaching capacity, The Post has learned.

Stottlemyre, who was replaced by new pitching coach Ron Guidry, will work with the pitchers for all of spring training.


February 2, 2006 - Jays ready to rule the roost
by SG's predicting the Blue Jays to win the AL East according to the article linked above. While I think it's possible, a lot has to break right for them and wrong for the Yankees and Boston for them to actually do it.

There's another article about the Yankees' depth in arms, which I'm a little tired about hearing about. I don't really consider having Jaret Wright and Aaron Small around as depth, since neither is likely to be very good.

Pitcher and catchers report in two weeks for the Yankees.

Here's a list of the Yankees' non-roster invitees:

Included in the list are some what we hope will be key parts of the Yankee future (J.B. Cox, Phil Hughes, Stephen White), some ghosts of rings past(Ramiro Mendoza), and some people who are hopefully just getting a paid vacation in the sun (Al Leiter, Russ Johnson).

For a kick, check out the age difference between Hughes and Leiter.