THE BOOK--Playing The Percentages In Baseball

Carmona: reported as 4/15, 1.169 service.

He’s already got a contract for 2008, so this is really a 3/14.5 deal, starting in 2009.

Starting in 2009, he would have been a Super2, and eligible for arbitration.  So, this deal covers the first 3 of his arb years.

This deal pays him for 3.0 WAR this year, with 0.4 WAR of aging.

What does Marcel think?  He’s a .570 pitcher, with only 162 IP.  That’s a little low, since he was mostly a reliever 2 years ago.  Let’s bump him up to 180 IP (20 games).

Replacement level is .370 in the AL, so he is +.20 wins for 20 games, or a 4.0 WAR pitcher.

Whoah.  Did the Indians just pull a fast one here?

For him to be a 3 WAR pitcher today, he’d have to have a combination of:
.570 pitcher, 135 IP
.550 pitcher, 150 IP
.530 pitcher, 169 IP
.510 pitcher, 193 IP

He should have signed a 4/22 deal (including 2008).  At least a 4/18 deal.

I see also that the Indians have option years, which further makes this a bad deal for Carmona.

Wow.  This must be the new market.  Shields (possibly a super 2 next year), signed a 4/11 deal (including 2008).

MLB teams will be signing these hot young pitchers as fast as possible after this signing.  This is really incredible.

And just 7 or 8 years ago, Mark Prior got this kind of deal.... out of college!

How the tables have turned on MLBPA.  I cannot believe that payroll is less than half of revenue (NHL and NBA, with their caps are at 55% or so).  If the players were smart, they’d INSIST on a payroll cap, and giving up arbitration for earlier free agency (4 years or age 27).

The entire strategy of MLBPA seems to have been predicated on the owners being foolish enough to behave as if they had a gun to their heads. 

This seems to be gone now. 

Welcome to Moneyball.

I’m beside myself here. 

Mark Mulder, going into the 2008 season, has 7.167 years of service.  That means, going into the 2002 season, he had around 1.167 years of service.  That’s what Carmona has.

http://mlbcontracts.blogspot.com/2004/12/st-louis-cardinals_111971260115041890.html

According to Cot’s, he signed a 4/14.2 deal, for years covering 2002-2005.

This is virtually identical to Carmona!

And that was 6 years go. 

What did Marcel think of Mulder back then?

http://tangotiger.net/marcel/

Here’s Carmona, 2008 and Mulder, 2002 according to Marcel, with Carmona’s numbers first:
K/BB: 114/52, 128/55
ERA: 3.78, 3.93

Carmona was born Dec 1983.  Mulder was born Aug 1977.  That’s 6 years apart.  They were the same age.

This is really incredible.  It’s as if the free agent market has moved at an accelerated pace, while the pre-arb market has not moved at all.

And not a single person in the press has said word one?

I’m hoping deals arrive soon for Joe Saunders, Ervin Santana, and Jered Weaver. 

I doubt Weaver will sign something like this though, he’s a Boras guy.

According to Rotoworld: “The Indians have options for $7 million in 2012, $9 million in 2013 and $14 million in 2014. The price of the first option could escalate by $1 million and the price on the other two by $2 million each”

Even with the options escalated, he seems underpaid. And you have to wonder what it takes to get the maximum value.

One thing about Carmona: his FIP last year was (by my numbers) 3.83, which may indicate that he got a bit lucky with his 3.03 ERA.  BABIP wasn’t overly bad at 0.284. 

But still...If Marcel understood FIP, he might forecast Carmona a bit more pessimistically.  But without running the numbers, I imagine this still works out as, at best, a break-even for Carmona.  The Indians know what they’re doing…
-j

I think I’ve seen this talked about before, but I’m not really sure. When evaluating a contract, is it best to look at total years/total money?

What I mean is, say with a big, long contract like Zito or Soriano (maybe bad examples since I think it’s obvious, especially with Zito, that the deal is awful), is it best to look at the contract at the end of X years and say “Ok, that turned out to be good/bad” or do you say “that was a good deal for the first Y years, but in those last two he was below replacement as a 36-37 year old and making $18 million dollars.”

Can a deal be good and bad (from the perspective of the team getting appropriate performance for the money)? And at what point is giving an extra year or two worth those first years? Are the final years on long term deals of FAs actually more important than the dollar amount because a guy like Zito will most likely be wasting a roster spot at the end of that deal. The Giants have the money, but they’ll still only be able to field 25 guys.

Justin, actually according to the THT stats, Carmona’s FIP last year was 4.05 meaning his ERA benefited from a whole whopping run of luck.  He might not have an incredible run like he did last year but he will definitely regress much less than say Kyle Kendrick will this year.

Eric,

I tend to prefer my own numbers to those someone else has calculated...and in the case of FIP, that’s particularly the case because I haven’t been able to reproduce THT’s FIP using their stated equations.  One should be able to back-calculate their additive constant and have it be the same across players (at least within a team), but the times I’ve tried it hasn’t worked out.  So I’m not honestly sure how they get to their FIP numbers, though mine are usually close to theirs (as they are in this case).

My numbers on Carmona are from my piece on the Indians last year:
http://jinaz-reds.blogspot.com/2007/09/postseason-profile-cleveland-indians.html
I used the FIP equation published on THT’s website, and I calculated the coefficient to force FIP=ERA (probably MLB ERA in this case, ‘cause I get sloppy by league differences at times).  I sometimes also do a park factor correction on the HR totals, but I don’t think I did that in this case.

As for Carmona, I think “definitely” is a bit too strong.  But in general I agree--as I said, if Marcel knew about DIPS, he would probably be a bit more pessimistic about Carmona than he is.  Still, there’s only so much you can teach a monkey, eh?
-j

The FIP fangraphs has for Carmona is 3.90, so basically right down the middle… might want to just use that. His xFIP, according to THT, was 3.99 in 2007.

I tend to prefer my own numbers to those someone else has calculated...and in the case of FIP, that’s particularly the case because I haven’t been able to reproduce THT’s FIP using their stated equations.  One should be able to back-calculate their additive constant and have it be the same across players (at least within a team), but the times I’ve tried it hasn’t worked out.  So I’m not honestly sure how they get to their FIP numbers, though mine are usually close to theirs (as they are in this case).

***

Justin,

Could it be that you’re not including HBP? THT does.

JD, there are a ton of variables that go into a contract being bad or good, but the primary factor is whether a player was overpaid for, given his talent/performance projection which essentially tells us how many wins a player will add to a team as compared to some baseline.

If a contract is way high, then a team probably (but not definitely) did not know how to properly evaluate that player, which is an egregious mistake since all of these projection systems are as good as each other and most of them are freely available on the internet.  Even if a team overpays for a player, that does not necessarily mean that it was a bad signing or that it is not “good” for the team or even profitable for them.  For example, it might have been a profitable signing and the reason they overpaid is that there were no cheaper alternatives.  But generally when a team overpays a lot for talent, it means that they are bad at evaluating that talent, and if they underpay, it means that they got lucky or smart, who knows, but either way, it was a “good buy” at least.

As far as long-term, short term, etc., we can figure out the value of a contract no matter how long and what the status of a player is in each year, and then compare that to how much is paid.  To do that, we extend the projection assuming that a player will play less amount per every subsequent year, due to injury and age, and his rate performance will decline after about age 27 or 28 and increase prior to around 27, etc.  We also assume FA salary inflation of 5-10% per year or whatever.  So we can put an “exact” dollar amount on a contract, whether it is long or short, based on the talent of the player (wins above replacement), his age, his status (FA, arb-eligible, etc.), salary inflation, etc.

Obviously, and as you point out, there are many considerations for each individual team in terms of whether or not a short or long-term contract is particularly good for that team.  We usually don’t get into that here - we use a one-sized fits all valuation process.  Obviously that is not correct, but at least it can usually tell us how much the average team would have paid for this player, given his age, status, and true talent level, if all teams knew exactly what that true talent was, given the current market value of a marginal win.

Could it be that you’re not including HBP? THT does.

Nope, I include HBP.
-j

Justin,

Are you excluding IBB from the BB totals?

studes,
That might be it!  I don’t remember seeing that in your equation--though I do see it there now.
-j

Arizona signed Chris Young for 5/28, which could turn into 6/37.5 (link on name). Cot’s has Young with 1 year and 45 days service time.

I don’t remember seeing that in your equation--though I do see it there now.

Pretty sneaky of me, eh?

Just curious, do you rate contracts replacing arbitration and their value considering the difference from what they’d get through standard arbitration and injury chance?

If a free agent signed this deal, it would be an entirely different story, but a 24 year-old pitcher with a single good season and another 5 years until Free Agency would be justified in cutting a deal to guarantee salary, especially with a significant innings increase like he had in 2007, no?

(This isn’t to say that the team options are a good deal, but buying out his arbitration years at a discount seems to be rather standard, doesn’t it?)

Evan Longoria: 6/17.5 (plus a one year option followed by a two year option; 9/44 possible)

Service time… uh, what does six days come out to?  0.03?  Jeez.

Brilliant (for the Rays).  This deal is through 2013, and the club holds an option for 2014.  The ONLY reason they hold that option is because they kept him in the minors (otherwise, he’d have been a free agent). It’s a cold-hearted business, but Rays management made alot of money here by holding that option.

He is being valued at just 2.0 WAR, with 0.2 wins in aging.  Talk about Longoria selling low.  Sheesh.

***

This is ENTIRELY the fault of the MLBPA.  The minimum salary in the NHL is HIGHER than in MLB, despite the fact that the NHL revenue and payrolls are half that of MLB.

The minimum salary in MLB should be set at close to 1 million$, if not 1.25 million$.  While the minimum salary in the NHL is almost 0.5 million, the average salary of first year players is probably closer to 0.6 million. 

The MLB teams are preying on the players’ incredible lack of pay in their first 2 or 3 years (400-500K per year) to sell them on the idea of guaranteed payments to take them through their arb years as multi-millionaires.

The players have only themselves to blame for this fiasco.

On the other hand, this huge imbalance in pay allows teams like the Rays to compete for talent.

MLBPA is acting like a Mickey Mouse organization, and MLB teams are tripping all over themselves trying to lock in prices that have barely changed since Nomar and Vlad started the trend TEN years ago.

And still, the press is only concerned about birth certificates.  What a truly horrible and disgusting state of affairs this is.

(Yeah, yeah… poor millionaires… boo hoo hoo.  Consider it said.  No need to repeat it.)

Underpaying young stars will become a much bigger issue in the next couple years as more teams take advantage of long-term deals like this one.  Remember Jonathan Papelbon’s mini-tirade earlier this spring?  There will be more of those.

I do hope we don’t see MLB have an NBA or NFL-style system.  There needs to be a way to pay young players, but still keep them with their original team longer.  Players developing with one team is a great baseball tradition.

I completely agree with #19, but have an even better example than the Nomar and Vlad contracts from 10 years ago.

When Hamels complained about his deal the Phillies were set to play the Braves in an exhibition game.  A local repoerter had some qotes with Smoltz as an elder statesman of the game and Smoltz talked about how he was renewed as a young player and he let it bother him and got off to a slow start.  That was the focus of the story.  But in telling his story Smoltz mentioned that he was renewed in ~1991 at 355k.  Hamels was renewed with the same service time at ~500k in 2008. 

I don’t know the exact revnue figures in the early 1990s, but overall MLB revenues have gone up something like 400% while these low service time players like Hamels are getting renewed at a 50% premium over what similar players were getting 18 years ago!  I’m not even sure if that increase is keeping up with inflation never mind the massive revenue growth in baseball.

For the last couple of decades the MLBPA has been fighting for top end salaries and arguing that the rising tide would lift all boats.  But now that player payroll as a percentage of revenues has fallen so dramatically it’s hard not to accept that that philosophy stopped working some time ago.  It’s time for the younger player to take a more active role in the union and start to fight for higher wages from the bottom up.

I’m with you all the way Philly.

***

It seems that it’s always been this way though. 
Here’s what these columns means:
year: duh
Minimum: minimum salary as best as I could infer
800th: the salary of the 800th highest paid player
avg800: average salary of the top 800 paid players (*)
800thIndex: the value in “800th” divided by the value in “avg800”

year Minimum 800thPlayer avg800 800thIndex
1993 $109,000 $109,000 $1,110,698 9.8%
1994 $109,000 $109,000 $1,148,424 9.5%
1995 $109,000 $109,000 $1,163,994 9.4%
1996 $109,000 $112,000 $1,178,347 9.5%
1997 $150,000 $150,000 $1,385,877 10.8%
1998 $170,000 $175,000 $1,555,544 11.3%
1999 $200,000 $205,000 $1,816,423 11.3%
2000 $200,000 $200,000 $2,073,669 9.6%
2001 $200,000 $203,500 $2,435,787 8.4%
2002 $200,000 $202,000 $2,518,588 8.0%
2003 $300,000 $300,000 $2,650,453 11.3%
2004 $300,000 $300,000 $2,576,707 11.6%
2005 $316,000 $316,000 $2,723,727 11.6%
2006 $327,000 $327,000 $2,894,291 11.3%
2007 $380,000 $380,000 $3,075,911 12.4%

(*) It bothers me to no end when average salaries are reported, and you don’t know if that’s for 750 players or 832 players or whatnot.  The more players you count, the lower the average, since all those extra players undoubtedly earn the minimum.

As we can see, it seems that MLBPA and MLB have a basic agreement that the minimum salary is around 10% of the league average.  And in fact, the minimum salary of 2007 is the highest it’s ever been, proportionately.

***

Here is how the 500th highest salary compares to the average of the top 800 players:

year 500th 500thIndex
1993 $245,000 22.1%
1994 $226,000 19.7%
1995 $220,000 18.9%
1996 $225,000 19.1%
1997 $275,000 19.8%
1998 $325,000 20.9%
1999 $450,000 24.8%
2000 $475,000 22.9%
2001 $550,000 22.6%
2002 $500,000 19.9%
2003 $500,000 18.9%
2004 $500,000 19.4%
2005 $500,000 18.4%
2006 $600,000 20.7%
2007 $575,000 18.7%

1999 was the peak for the “middle class”, but otherwise, it stayed at roughly 20% the whole time.

***

Here it is for the 250th highest paid player every year:
year 250th 250thIndex
1993 $1,100,000 99%
1994 $1,100,000 96%
1995 $900,000 77%
1996 $1,000,000 85%
1997 $1,420,000 102%
1998 $1,600,000 103%
1999 $2,000,000 110%
2000 $2,300,000 111%
2001 $2,750,000 113%
2002 $2,800,000 111%
2003 $2,900,000 109%
2004 $2,500,000 97%
2005 $2,750,000 101%
2006 $3,000,000 104%
2007 $3,450,000 112%

The overall average is 102%.  So, historically, the 250th highest paid player is roughly the same as the average of the top 800 paid players.

Lots of ebbs and flows, but we can see that after 1996, there was a takeoff in the “upper middle class” salaries.

***

Finally, the top 100 salaries:
year 100th 100thIndex
1993 $3,000,000 270%
1994 $3,250,000 283%
1995 $3,450,000 296%
1996 $3,250,000 276%
1997 $3,610,000 260%
1998 $3,950,000 254%
1999 $4,750,000 262%
2000 $5,375,000 259%
2001 $6,125,000 251%
2002 $6,367,542 253%
2003 $6,850,000 258%
2004 $6,500,000 252%
2005 $7,250,000 266%
2006 $7,666,667 265%
2007 $7,758,503 252%

The average is 264%.  We can see that it’s really the 1993-1996 time period where things were really imbalanced.

More study needed…

Very interesting data.  I would have thought more of an imbalance.  Several years ago I saw a report comparing the average salary to the mean salary and as I recall the average was over 2M and the mean was in the 800k range.  Unfortunately, most stories about salary growth do no include that comparison, but I’ve always assumed that the average salary (now over 3M I beleive) has gorwn much faster than the mean.  As a result pre-Arb players who mosly make up the bottom half were falling further behind.

Another positive about a 1M minimum salary is that it would provide a much bigger financial pay off for the career minor leaguer/fringe MLB player.  Certainly a few months of 400k pay is pretty attractive, but for the guys who don’t get pensions and never sign multi-year deals it won’t last long or really make up for the slave wages paid to minor leaguers.

In the 1987-96 drafts about 1500 players made the majors and ~67% ended their careers with less than 10 WARP usually piecing together partial seasons.  Those kinds of players - a majority - would be the ones that would get the biggest marginal benefit from a large jump in MLB minimum and it’s hard to argue that baseball has done much financially for those kinds of players.

Philly: your use of “mean” is really “median”.

Mean=average.

Yup, that’s what you get for quickie post before heading out the door…

My long post got eaten, so I will just provide the link this time.
vr, Xei

Hanley Ramirez signs for 6/70 (link on name).

Hanley: 2.014 years.days of service, entering 2008.

He’d be arb-eligible in 2009-11, and FA in 2002.

Hitting: .383 wOBA, making him +2.5 WAA

Fielding: yech.  Fans see him as -1.0 wins relative to average at position.  UZR is disgusted by him at -2.0 wins.  Dewan is at -1.5 wins.  Let’s go with an average at -1.5 wins

Position: he gets +0.5 wins

Repl: in the NL he gets +2.0 wins

He plays all the time, so let’s give him 90% playing time.

WAR = 0.90 * (2.5 - 1.5 + 0.5 + 2.0) = 3.15

He’s young, so his aging curve will be fairly modest.  Let’s drop him by 0.25 wins rather than the stand 0.50 wins. 

Presuming he’s got a 6yr deal (not an “extension"), and including his service time, then the fair value would be 6/55 deal for him.

If it was a 6yr extension, meaning the contract starts in 2009, then 6/68 is the fair deal.

Seeing that there were alot of contracts signed at a discount for guys in Ramirez’s salary class, his fielding has not been discounted at all.

Kazmir: 3/28.5 extension, starting in 2009.  He has 3+ service years entering 2008.  So, his 2011 season would have been a free agent year.

Hard to figure out how good he is.  Let’s make him a .575 pitcher, with 20 full games, with a league repl level of .390.  That makes his WAR entering 2008 as 3.7.  He’s very young, so we should give him a modest aging curve.

I’ve got him at a 3/35 deal as a fair value with fairly aggressive aging.

I think Kazmir has sold himself short here.

I forgot to mention: if you make him a 2.7 WAR entering 2009, and drop him by 0.5 each year, that would be worth 3/28.5.

2.7 WAR on 20 full games is a .525 pitcher.

Ryan Braun: 8/45, starting in 2008.  Entering 2008, he had 129 service days, which may have made him eligible for super2.

Hitting: +3.7 wins above average per 162.

Fielding/position: presuming he’s an average fielder in LF, then he’s -0.5 wins.  If he’s worse than average, he’s -1.0 wins.

Repl level: +2 wins per 162

Playing time: let’s give him 85%.

Obviously, very rough calculations.  He was a top draft pick (5th or something, in the same draft group as Zimmerman and Tulo), so we have more information for regression. 

WAR = 0.85 * (3.7 - 0.7 + 2) = 4.25

He’s 24.5 years old, so we give him low aging for the first few years, then the normal 0.50 aging for the rest.  Let’s say 0.40 wins lost per year.

I get: 8/64 (with the service time knowledge).

Looks like the standard discount that non-millionaire ballplayers give (Tulo, Utley, et al).

To get a 8/45 deal, with 0.40 wins in aging, would mean to start him as 3.6 WAR for the 2008 season, and going down from there.

Interesting clause in Braun’s contract (link on name):

“If he qualifies as a “Super 2” arbitration-eligible player following the 2009 season, the total value of the deal will jump from $45 million to $51 million. Braun’s salaries for his arbitration-eligible years would be $3.5 million in 2010, $5.5 million in 2011, $7.5 million in 2012 and $9 million in 2013.”

They seem to value it as a % of 12.75 million.

3.5/12.75 = 27%
5.5/12.75 = 43%
7.5/12.75 = 59%
9.0/12.75 = 71%