Jason Rennie's Baseball Blog

4/19/06 - Damon

The Red Sox let their star center fielder go and he landed on "the dark side." Many have said that he will be missed sorely. Is this so? Damon batted lead-off and was often regarded as one of the best lead-off hitters in baseball. The best lead-off hitters: (1) get on base frequently, and (2) have the ability to steal a base. Damon is a very good base stealer. 2003-05, he stole 67 bases with a 82% success rate. Did he do a good job of getting on base? Except for SS and 2B, the Red Sox' position players didn't change much 2003-05. Here's how they compare in terms of on-base percentage (OBP):

  1. Manny .404
  2. Ortiz .383
  3. Nixon .378
  4. Mueller .378
  5. Varitek .369
  6. Damon .364
  7. Millar .362
Damon was 2nd-to-last, and not by much. Granted, the Sox had the top offense in baseball '03-'05, but a lead-off hitter should hardly be considered spectacular if he can barely manage the 6th best OBP on his own team. Damon was probably the best choice for lead-off since no one else in the list has base-stealing abilities like he does. But, he won't be missed. The Sox have a better lead-off hitter in Youkilis (who's OBP abilities more than make up for his lack of speed), and a good, young CF in Coco Crisp.

4/7/06 - Cleveland Indians

The Indians look like an exciting team, no? Lots of young, talented players. Thought I'd do the same analysis as I did for the Sox & Yankees, see how they stack up. Again, I'm completely relying on the Indians depth chart pulled from Sportsline.


That's a powerful lineup. And, I was somewhat conservative in estimating OPS+; some players have already shown that they are capable of better. If a few things go their way, the Indians could end up with the most potent offense in all of baseball. Though, one thing to note is that the park factor for Cleveland was particularly low last year, so that could be skewing the numbers a bit.


Cleveland's pitching looks more subject to criticism, but, what a bullpen! I avoided using career ERA+ for some of the young relievers (e.g. Cabrera, Miller) 'cause I suspect they may not be able to repeat their past performances. So, again, that team average ERA+ is a bit conservative. It doesn't reach the highs of the Yankees and Sox, mainly due to that weak starting rotation. Always a chance that one of the starters will have a break-out season, but I wouldn't bet on it.

Cleveland has a great lineup and an awesome bullpen, but the starting rotation will pull them down and could easily be their doom. It's very hard to win in the playoffs without a few top-notch starters.

4/3/06 - American League East

Once again, the AL East is most likely going to come down to a race between the Red Sox and Yankees. David Pinto thinks that once again the Yankees have the edge. I wonder, is that really the case? I used OPS+ and ERA+ to compare the two teams, making use of SportsLine's Depth Charts to determine probable positions.

Pos.YankeesOPS+Red SoxOPS+

Most of the OPS+ estimates are career averages. I made exceptions for those who have been consistently performing better (Crisp, Varitek, Ortiz) or worse (Williams) than their career averages.

Yankees have the edge on offense. Break-out performances by Crisp, Youkilis or Cano could substantially tilt the balance, as could a Yankee DH other than Bernie.

WeightPos.YankeesERA+Red SoxERA+

All of the estimates are career ERA+ (except Small/Dotel and Papelbon). Each ERA+ is weighted by an estimate of innings pitched ("weight"). The Yankees have Johnson and Rivera, but the quality falls off sharply after that. The Sox have the better overall staff, with a slightly better rotation and a much better non-closer bullpen.

Well, the Sox and Yankees look extremely closely matched. Yankees are 4 points better in OPS+, Sox are 4 points better in ERA+. Could defense be the tie-breaker?

WeightPos.YankeesZRRed SoxZR

I used career average position zone ratings for everyone. I weighted each ZR by a typical range factor for that position (except 1B, for which I estimated an effective weight between middle infield and 3B). The Red Sox come out ahead with substantially better defense at four positions (1B, 2B, SS, RF) and substantially worse defense at two positions (LF, CF). 'course, it's well known that Manny's LF ZR is much lower than it should be due to the green monster. Chances are very good that Manny is no worse than Matsui and the Yankees' only defensive advantage is in CF. For the record, if we bump Manny's ZR to .833, the Sox average ZR is .844. The Yankees allow 1.4% more BIP to turn into hits/errors. That's not a trivial number... Add in the fact that the Red Sox also have the better farm system (with 4 in Baseball America's top 100 prospects vs. 2 for the Yankees), and I think the Sox have the edge this year. 'course, anything can happen!

2/25/06 - Lineups

David Pinto has posted a Lineup Analysis page. Enter players from your favorite team, along with your best estimate of their OBA and SLG for next year, and David's script will return to you the optimal lineup. I did this for the Red Sox. Here's what it returns:

1Kevin Youkilis.380.410
2Manny Ramirez.400.600
3Jason Varitek.360.480
4David Ortiz.380.600
5Trot Nixon.365.490
6Mike Lowell.340.460
7Coco Crisp.340.450
8Alex Gonzalez.290.390
9Mark Loretta.370.420

This lineup will supposively produce 5.566 runs per game. You get a nearly identical result by swapping Crisp/Lowell and/or Loretta/Youkilis. There's also little loss in swapping Varitek and Nixon. The worst lineup (with Gonzalez batting lead-off, Crisp #2 and Manny and Ortiz batting #8 and #9) yields 5.074 runs/game. As David mentions, the script tends to put the worst hitter at #8 and will often put a high-OBP hitter at #9 (to improve the chances of having someone on base as the sluggers approach the plate). Makes a lot of sense to me. Here's hoping that Tito finds this page! :)

The lineup I had chosen in my 1/28/06 entry (with Gonzalez replacing Graffanino) rates 5.410 runs/game. I yield. I have to say I like the lineup returned by the script is better than the one I had originally chosen. I now think it's quite valuable to have a table-setter at #9.

1/28/06 - Crisp Fenway

Welp, the Sox have filled their CF hole. Marte is gone to the Indians and Coco Crisp is now a Red Sock. The trade turned out to involve quite a few other players, though besides Marte and Crisp, it seems like a trade of equals---Mota and Shoppach for Josh Bard and David Riske. The good news is that the Sox now have a full line-up. The bad news is that they might be tempted to slot Crisp as their lead-off batter. I sure hope they don't do that---Crisp is relatively poor at getting on base. Here's my revised batting order, with my estimate of OBP and ISO (SLG-AVG):

1Mark Loretta.370.100
2Kevin Youkilis.380.125
3Manny Ramirez.400.300
4David Ortiz.380.300
5Trot Nixon.370.200
6Jason Varitek.360.180
7Mike Lowell.340.180
8Coco Crisp.340.150
9Tony Graffanino.335.130

My batting order philosophy is that the top two positions should be high OBP, moderate-to-low ISO guys. #3 should be high OBP and moderate-to-high ISO. #4 should be high ISO and moderate-to-high OBP. Slots #5-9 should be filled-in roughly according to estimated OPS. If there is a high OBP, low ISO player (that doesn't fit elsewhere) he should take the #9 slot (in the hopes of providing an extra baserunner for the top of the lineup). I'm tempted to move Youkilis to #1, Nixon to #2, and drop Loretta to #9... I'd have no problems swapping Lowell and Crisp. I'd also buy an argument to swap Ramirez and Ortiz to avoid having two lefties back-to-back (Nixon & Ortiz).

As far as positions go, Riske is the new reliever and I think Bard will serve as the backup catcher:

BCDoug MirabelliJosh Bard

It's likely we'll see Flaherty traded away or released...

1/19/06 - 2006 Red Sox

Phew! Another exciting off-season for the Sox. They've done nothing less than uproot a significant portion of their roster. Here's how things looked last year vs. how I think/hope they will look this year. Players who are new or who are taking on a substantially larger role are in bold.

CJason VaritekJason Varitek
BCDoug MirabelliShoppach/Flaherty
2BBellhorn/GraffaninoMark Loretta
SSEdgar RenteriaTony Graffanino
BMIFVazquez/CoraAlex Cora
1BKevin MillarKevin Youkilis
3BBill MuellerMike Lowell
BCIFYoukilis/OlerudJ.T. Snow
LFManny RamirezManny Ramirez
CFJohnny Damon?
RFTrot NixonTrot Nixon
BOFPayton/KaplerAdam Stern
DHDavid OrtizDavid Ortiz
SPMiller/SchillingCurt Schilling
SPDavid WellsJosh Beckett
SPTim WakefieldTim Wakefield
SPMatt ClementMatt Clement
SPBronson ArroyoBronson Arroyo
SP/RPHalama/GonzalezJohnathan Papelbon
CLFoulke/SchillingKeith Foulke
STMike TimlinMike Timlin
RPMike MyersLenny DiNardo
RPEmbree/PapelbonRudy Seanez
RPMantei/BradfordJulian Tavarez

I hope the Sox consider Graffanino for the SS job; he's the better hitter and I think he has the fielding skills to play there. Though I wouldn't be surprised to see Cora and Graffanino swap roles above. I hope the Sox don't waste talent or money on a new SS as I think Pedroia will be ready to fill that role come 2007. I hope the Sox use Shoppach as their primary backup catcher---he's had enough time at AAA, he's supposed to be a top defensive catcher and I don't think he'll hit any worse than Flaherty's career .667 OPS. And, it'd be good for the Sox to have some young blood at the catcher position. In fact, if Flaherty's minor league record is any indication (drop of ~100 points of OPS from AAA to MLB), Shoppach should be able to hit .700+ OPS in the big leagues (last full season at AAA: .859 OPS), which is more than plenty for a backup catcher.

Pitching should be improved. Last year, we largely saw a rotation of Wakefield/Arroyo/Clement/Wells/Miller. This year, I think we'll mostly see Schilling/Beckett/Wakefield/Arroyo/Clement. Pretty nice, IMHO. I think it's pretty likely we'll see Wells traded/released. It almost goes without saying that the 'pen will be better. Foulke can't do much worse than last year and Mota/Tavarez/Seanez/Papelbon will be an improvement over Embree/Mantei/Gonzalez/Halama. I think Papelbon is very close to being ready for Boston, though I certainly wouldn't mind seeing him pitch an extra few months at Pawtucket. It's likely we'll see Hansen as a regular contributor by mid-season. I just hope the Sox give him some experience in Pawtucket before bringing him back to Boston. I bet we'll see Mota released/traded to make room for Hansen by mid-season.

The Sox still have a glaring hole in CF. But, decent CF's don't seem too hard to come by; three of them went through the Red Sox organization (Dave Roberts, Jay Payton and Joze Cruz Jr.) in a matter of 8 months, 12/20/04-8/9/05. And, Durbin/Murphy may be ready for prime-time next year, so I hope they don't waste $$$ on a long-term contract for a slightly-above-average CF (like they did w/ Renteria/SS last year).

Looks like the Sox will have another fantastic draft, with no lost picks and three bonus 1st round picks (that makes for four of the first 43 picks!). Prepare for more complaining from the Brewers' and Devil Rays' brass (who's teams get no bonus picks).

Here's how I think/hope the batting order will look:

  1. Mark Loretta
  2. Kevin Youkilis
  3. Manny Ramirez
  4. David Ortiz
  5. Jason Varitek
  6. Trot Nixon
  7. Mike Lowell
  8. Tony Graffanino
  9. CF

I'd also consider putting Youkilis at #1, Varitek at #2, moving up Nixon and Lowell and dropping Loretta to #7. And, of course, this depends on who the Sox get for CF. I'm not hoping for much. Even without a "superstar" CF, the Sox look well-positioned for another run at the Series.

10/11/05 - The Atlanta Braves

The Atlanta Braves have made it to the playoffs as the NL East division winner for 14 straight years (1991-93,1995-2005) under the management of Bobby Cox. But, they've only won one World Series during that reign. David Pinto asks: where are the calls for Cox to be fired? Shouldn't he have won more World Series with all of those chances? What is the chance that such a team would win only one World Series?

Let's start with this question: what is the chance that the Atlanta Braves would have won X World Series given these 14 opportunities, 11 years with the 3-level playoffs, 3 years with the 2-level playoffs? To be able to say anything specific, we need an assumption to simplify the problem. Here's the assumption I'm going to use: every team that made it to the playoffs had an equal chance of winning the World Series. Thus, each year the Braves made it into the playoffs 1995-2005, they had a .125 chance (1-in-8) of winning it all; 1991-1993, they had a .25 chance (1-in-4) of winning it all.

With this assumption, we can now calculate the chance of having won a certain number of World Series (WS). Let's start with the end points. There's only one way they could have won zero WS---if they had lost every time! I.e. (1-.125)^11*(1-.25)^3 or 9.7%. So, there's a non-trivial chance of playing in 14 playoffs and not winning a single Series. Now for the other end---the chance that they would have won every Series. That's .125^11*.25^3 or .0000000002%. I.e. winning the last 14 WS would have been practially impossible.

The middle points (1,2,3,...,13) are a little more complicated because there are multiple ways in which you can win. Consider what really happened---the Braves won one WS. What's the chance of that happening? Well, they could have won in any of the 14 years. The chance of winning one 1995-2005, but losing the rest is (1-.125)^10*(1-.25)^3*.125; the chance of winning one 1991-93, but losing the rest is (1-.125)^11*(1-.25)^2*.25. There are 11 ways to win one 1995-2005, 3 ways to win one 1991-93, so the total probability of winning one is 11*(1-.125)^10*(1-.25)^3*.125+3*(1-.125)^11*(1-.25)^2*.25 or 25.0%. Anyone familiar with the binomial theorem should see where this is going. Here's the full table (may not sum to 1 due to rounding):

Number of
World Series Wins

Here's the code to generate the table. As you can see, most of the mass of the distribution is between 1 and 3 wins. The expected value is 2.11. So, the Braves aren't far from where you'd expect them to be. Considering that 34.7% of teams with the same record of getting into the playoffs would have won one or zero WS, I don't think you can blame any member of the Braves for a "poor" playoff track-record.

7/29/05 - Jose Cruz Jr.

The Arizona Diamondback just designated Jose Cruz Jr. for assignment (DFA). A quick glance at his batting average this year (.213) and you might conclude that his bat is washed up. But batting average is a somewhat poor measure of a player's offensive contribution to a team. OPS (on-base percentage plus slugging percentage) is much better in that respect. And, Cruz's OPS this year (.783) has been just fine. He's hitting slightly better than he has any of the past three years and he's very close to his career OPS. In fact, his .783 OPS is quite good for a center fielder. Even "superstar" Carlos Beltran isn't hitting that well this year. But, if Cruz is hitting so well, why would Arizona DFA him? Could they really be that obsessive about a low batting average? And, well, since he must now be traded or let go, aren't there a lot of teams that would love to have a center fielder with that kind of offense?

Here's the dig: the guy can no longer field. Or at least, so far this year, he's been absolutely terrible in center field. Over his career, he's caught 87% of the balls hit to the defensive regions assigned to the center fielder. This year, over 415 innings, he's only getting to 77% of the balls he's supposed to catch. Let me tell you: that is a huge drop-off. No center fielder with more than 100 innings this year is catching balls at such a low rate! If this is representative of his real fielding ability, it's unlikely he'll see the likes of center field again. And, without the ability to play center field, his .783 OPS looks pretty mediocre.

This drop-off in fielding ability is probably due to a pinched-nerve back injury he suffered earlier in the year. It's likely that Cruz returned to playing before he was fully healed and his fielding suffered as a result...

6/15/05 - A Slow Bat?

There has been much discussion on the soxprospects.com message boards about the idea that Kevin Youkilis has a "slow bat", i.e. that he doesn't have a quick-enough swing to hit a good fastball. I have to say, I was extremely skeptical of this notion. To this point, Kevin has hit .807 OPS in his first 306 major league plate appearences (PAs). Few players start their careers hitting that well. Granted, Kevin is a bit older than your average prospect---he was 25 for his first year of major league ball. But, even if he merely manages to keep hitting at a .800 OPS clip, he'll almost certainly have a place in the majors as a starting 3B. But, it's hard to survive in the majors without the ability to hit a good fastball. Does Youkilis really have a "slow bat," or is this just heresay?

Toward the end of the 2004 season, the Boston Globe posted scouting charts for all the Red Sox hitters. Here's Youkilis' chart. Here's how he hit vs. various types of pitches:

Pitch TypeBatting Average
Youkilis appears to be poor vs. the curve, mediocre vs. the fastball, and good vs. sliders & changeups. If Youkilis did have a "slow bat" during the '04 season, we'd expect that his fastball average would be much lower than that of other major league players. In fact, this does appear to be the case. Here's a list of '04 Sox players and their fastball batting averages, compiled from the same scouting data:
PlayerFastball Batting Average
*Note: to calculate overall average for switch hitters, I used 35% average vs. LHP and 65% average vs. RHP. All of the regulars had a substantially higher fastball batting average than Youkilis. And, many of these players draw walks at a rate similar to that of Youkilis. So, it does seem that Youkilis is not a very good fastball hitter; and, a likely cause of not being able to hit fastballs well is a somewhat slow swing.

Does this mean that Youkilis is doomed? Hardly. A lot of major league hitters have serious deficiencies (e.g. Millar vs. the curve; Bellhorn & Ramirez vs. the slider); many major league pitchers don't have a great fastball; and, he's somewhat young and may still be able to improve his swing. However, it does provide some insight as to his ultimate celing (don't expect him to follow in A-Rod's footsteps!); it also means that his numbers to this point may be somewhat inflated---if he doesn't improve his ability to hit fastballs, pitchers will take advantage of this as he plays more. On the other hand, since he's fairly good with off-speed stuff, it's not unrealistic to see him hitting .900+ OPS if he is able to substantially improve his ability to hit fastballs!

6/14/05 - The Draft Game

Last week, Major League Baseball conducted its annual amateur draft. Of course, a big event like that couldn't go over without some controvery. A big talking point was the fact that the Red Sox and Cardinals both had a large number of top picks. The Sox had 6 of the first 60 picks (the average team had only 2 of the first 60 picks); the Cardinals had 6 of the first 80 picks (the average team had only 2.7 of the first 80 picks). The Red Sox and Cardinals played in the World Series last year. How did they end up with so many good picks? Some general managers cried foul, as can be read in this Milwaukee Journal Sentinel article. The Brewers' GM, Doug Melvin had this to say:

There's something wrong when Boston and St. Louis have all those early picks and teams like us have only two in the first three rounds.
Chuck LaMar, GM of the Devil Rays, was similarly annoyed.
It's asinine. For the world champion Boston Red Sox to have five or six picks between our eighth pick and 58th, something is not right. And that's not blaming the Boston Red Sox. It has nothing to do with them. It has to do with the system that I think has needed to be changed for years.

One of the tenents of the draft is that it's supposed to improve the prospects of poor teams. In-line with this philosophy, each team gets a pick each round in their reverse regular season finishing order. The team with the worst regular season record gets the first pick each round, the second worst team get the second pick, etc. But there are also rules to compensate teams for players lost to free agency---the idea being that it is the small-market teams losing good players to free agency. The compensation is based on the player's "ranking," which is heavily based on his last two seasons' "triple crown stats" (i.e. AVG, HR, RBI for batters, W/SV, ERA, SO for pitchers). Players are "type A" if they're in the top 30% of comparable players; type B players rank in the next 20%; type C players rank in the next 10%. No compensation is offered for the bottom 40% of players. Baseball America's draft order details which picks serve as compensation for a lost free agent. Let's review how the Sox and Cardinals got so many draft picks.

The Red Sox lost three players to free agency and gained six picks in return:

Free AgentCompensation
Pedro Martinez (type A) to Mets1 2nd, 1 supp. 1st
Derek Lowe (type A) to Dodgers1 1st, 1 supp. 1st
Orlando Cabrera (type A) to Angels1 1st, 1 supp. 1st
The Dodgers and Angels were in the top half of the '04 regular season winning percentage; the Mets were in the bottom half of the '04 regular season winning percentage. This is the reason the Sox took a 2nd round pick from the Mets and 1st round picks from the Dodgers and Angels. The Sox also signed a number of premier players: Renteria (type A), Wells (type A), Clement (type B) and Wade Miller. In principle, the Sox should have lost three first round draft picks due to these signings! But, compensation picks can't be lost (the idea being that poor teams are the ones who are supposed to be losing top players to free agency), so the Sox instead lost their normally assigned 1st, 2nd and 3rd round picks.

The Cardinals lost two players to free agency and gained four picks in return:

Free AgentCompensation
Edgar Renteria (type A) to Red Sox1 1st, 1 supp. 1st
Mike Matheny (type A) to Giants1 2nd, 1 supp. 1st
The Cardinals would have gotten a 1st round pick for Matheny had the Giants not signed a higher-ranking FA in Armando Benitez. The Cardinals didn't sign any high-ranking FAs, so they kept their regular picks. They signed Eckstein to replace Renteria and used one of their own recent draft picks (Yadier Molina) to replace Matheny.

Is the system flawed? Well, I don't think you can look to the Cardinals to make such a claim. They acted like a small-market team, replacing Renteria and Matheny with (respectively) a low-budget FA and a home-grown player. If anything, the Cardinals seem to have been rewarded for intelligent management. What about the Red Sox? I think there's more of a case here, but you have to congratulate the Red Sox for taking maximum advantage of the existing rules. In terms of player "types," the Sox' net change was the loss of a type A player and the gain of a type B player. And yet, their net change in terms of draft picks was to gain one 1st round pick, and three (1st) supplemental picks, and to lose one 3rd round pick; they also greatly improved their 2nd round positioning. That seems like a major gain in draft picks for hardly any loss of player "type." I think there are two problems here. First is the rewarding of two picks for a lost type A free agent. This encourages turnover. The Angels, Cardinals and Red Sox effectively rotated SS over the 04-05 offseason, yet they netted two supplemental 1st round picks! Another issue brought to light by the Red Sox' picks is the fact that a team can't lose compensation picks. If anyone should be crying foul over this draft, it should be the Padres and the Cubs, who, (respecively) got 2nd and 3rd round picks instead of 1st round picks as compensation for losing Wells and Clement. If, say, the Marlins had signed Clement instead of the Red Sox, the Cubs would have received the 16th pick of the draft! Instead, they got the Sox' 3rd round pick (#106) as compensation for Clement.

What could be done to make the draft more "fair?" First of all, teams shouldn't get two picks for losing a type A player. A team should gain one pick as compensation for losing a premier free agent; a team should lose one pick as a penalty for signing a premier free agent. Second, a team's compensation should not depend on the team who signs the free agent. Here's how I think the compensation system should work:

In contrast to today's rules, which encourage player turnover, these rules would slightly discourage turnover. They would also be more fair to teams that lose FAs, since the compensation would be minimally dependent on outside factors. One flaw of this system is that a team that nets more than one FA signing of a particular type per year may never incur a penalty for some of their signings. But this is a mute point since the current system encourages player turnover, yet only one team (Yankees) regularly signs at least one more FA of a particular type (A) than it releases. In this proposed system, even the Yankees may change their ways. And, in contrast to the current system---where a team can achieve a net gain in player types and a net gain in draft picks---a team is always punished for a net increase of a "type" of player. Another benefit of this system is that it would limit the dilution of later round picks.

5/26/05 - The Physics of Baseball

The American Scientist has an excellent article on the physics of pitching. There is also good analysis of the thought process of hitters and what it takes to recognize and hit major league pitches. In particular, the authors discuss why 2-seam fastballs and 4-seam sliders are easier to hit than 4-seam fastballs and 2-seam sliders.

5/23/05 - Win Probability

Fans, announcers, and even baseball statisticians all feel the need to attribute the winning of a game to an individual player. There is even a formal set of rules for determining whether a pitcher is assigned responsibility for the outcome of the game. Announcers tend to select players who hit home runs at semi-pivotal moments because they look good on TV. No one cheers the player who draws a lot of walks but strikes out a lot.

Often times, I find these proclaimations highly subjective or even arbitrary. I think there is a better way, and it's called win probability. Win probability tells you the chance that the a team will win a game given the current inning, # of outs and baserunner positions. We can calculate a player's contribution to a win by looking at how the win probability changes every time he makes a plate appearance, faces a batter or steals a base. The sum of these differences represents how much the player contributed to the win. The player who draws four walks could easily provide more value than the player who hits a solo home run, but strikes out with men on three times. It also provides a level "playing field" for hitters and pitchers, and considers a player's total contribution rather than his single best contribution.

5/17/05 - Should Manny Ramirez go to the Hall of Fame?

Ryne Sandberg posed this question in one of his articles: "Is Manny Ramirez a Hall of Famer?" He went on to compare Manny to Jim Rice and Andre Dawson and said that Manny shouldn't be considered for the HOF until he hits 500 HR. I wondered whether the comparisons to Rice and Dawson were reasonable. After doing some research on baseball-reference.com, using OPS+ as my guide, I decided that Manny is in a league well above that of Jim and Andrew. So, I wrote to Ryne with my thoughts.

Hi Ryne, I wanted to chime in on the "will Manny Ramirez get to the HOF" debate. Jim Rice and Andre Dawson were both great players, but I don't think their hitting abilities compare to Manny. A good way to compare hitters from different eras is OPS+. It tells you how a player's OPS compares to other players in the same year/league and adjusts for park effects; a 100 OPS+ is average; larger than 100 OPS+ is above average.

Jim Rice achieved a 158 OPS+ in 1978 (his best year); in his 10 best years, Jim averaged an OPS+ of 136. Andre Dawson's best OPS+ was 157 in 1981; in his 10 best years, Andre averaged an OPS+ of 132. Manny, on the other hand, achieved a 190 OPS+ in 2002 and has averaged a 160 OPS+ over the past 10 years. i.e. Manny is 25-30 points of OPS+ better than Jim and Andre. Manny has also lead the league in OPS three times; Jim and Andre combined to lead the league once. Frank Robinson (career OPS+ of 154, led the league in OPS three times, hit for power and drew lots of walks) is probably a better comparison.

Frank Robinson actually lead the league in OPS four times (I missed the time with Baltimore). Being a Red Sox fan, I'm a bit biased, but considering that (1) he's the now youngest active player with 400 HR (though A-Rod will soon claim this title), and (2) he has the third highest career OPS+ among active players I think Manny's chances of getting into the HOF are very good.

5/10/05 - Steroids and Home Runs

ESPN has a nice article on
steroids and home runs. I think the most interesting aspect of the article is the list of HR/G stats over the first five weeks of each of the past 10 MLB seasons. So far in 2005, we've seen 0.19 fewer home runs per game than to this point in 2004. But, this isn't the worst drop-off that's been observed over the past 10 years. In 1997, there were 0.48 fewer HR/G than in 1996! In fact, if you look at consecutive differences in HR/G and rank them, you find that the current drop-off is hardly unusual:
YearHR/G difference from previous year
It's the sixth largest difference, only one step from the median (+0.09 in 2004). Three years had shown larger drop-offs through five weeks. In other words, what we're seeing is well within the range of normal differences---we can't really say anything yet.

Btw, the Hardball Times has a nice article on closers and the use of bullpens, and an excellent article on why the White Sox are winning so many games (hint: it has nothing to do with "small ball").

5/8/05 - Rosters

With the Red Sox seemingly aggressively promoting some of their players, there has been talk on soxprospects.com of the constraints these moves might have on their 40-man roster. Brad del Barba has a nice discussion of one aspect of the 40-man roster: the Rule 5 Draft. Basically, there's a draft during the winter meetings where teams can "draft" players from other teams. However, players on the 40-man are protected. Also, high school players are protected for three years from signing; college players are protected for two years. And, to "draft" a player, a team must have an empty spot on their 40-man roster. And, a drafted player must remain on the major league roster the entire next year. Also, teams may draft "up" minor leaguers (AA->AAA or A->AA).

Baseball America also has a discussion of the Rule 5 Draft.

The MSN Baseball forum has a nice discussion of minor and major league roster rules.

ESPN has a nice article on baseball transactions.

4/30/05 - Attendance

With offense clearly on the decline for 2005, the question of what draws fans recently came up on Baseball Musings. Andrew pointed out that it shouldn't be hard to draw some conclusions from the numbers. I followed up his comments by running a regression on 2004 winning percentage, runs and attendance numbers. I took winning pct, runs and total attendance for each 2004 team from baseball-reference.com, normalized winning pct and runs to zero mean and unit standard deviation, then ran regression with a bias/intercept. Winning percentage had the larger weight, but the weight on runs was also positive, so maybe people will come out to see offense, but only come out in droves for winning teams. Of course, these results must be taken with a large helping of salt---there are so many other factors in play that one year's worth of data surely isn't enough to get a true picture. Here are the weights:

WEIGHTS: pct=351874 runs=130625 bias=2416629
Since I "standardized" the input variables, there's not some clear, intuitive explanation for what the weights mean, except that winning is about 2.7 times more important than offense for drawing people to the games.

4/29/05 - OBP vs. SLG

A few days ago, I made a series of posts on Baseball Musings in response to Michael Lewis' article on power in Baseball. It's an excellent article; Lewis is great fun to read. When I got to page five of the article, I read this:

On-base percentage is between two and three times more valuable than slugging percentage, depending on whom you ask within baseball's statistical community.
I also recalled reading this in Lewis' Moneyball book. "I don't buy that," I thought. Two-to-three times? A walk more than half as valuable as a home run? If the typical on-base-percentage (OBP) were .600, not around .340, I'd be willing to believe this, but this claim seemed a bit far-fetched.

Then, I realized, such a claim is not difficult to evaluate---take team statistics and run regression on runs with OBP and slugging percentage (SLG) as the input variables. In my initial regressions, I found SLG and multi-base hits to be quite valuable. In fact, the results indicated that a point of SLG was actually more valuable than a point of OBP.

Tangotiger pointed out a problem---I made the mistake of forcing the intercept to zero and/or not including a variable representing outs. In a sense, this forced the model to limit the value of a walk or hit purely to its direct run production and did not allow it to account for the benefit of not producing an out.

I grabbed 2004 team stats from Baseball Reference and ran the regression anew. First, I considered the following question: given a team's OBP and SLG averages, what is the number of runs you'd expect to see them score in a single game? I used OBP, SLG and a constant 1 (also known as "bias" or "intercept") as variables, and runs/games as the target value. The results were a bit closer to the claim in Lewis' writing: OBP was given a weight of 16.85, SLG got a weight of 12.83, and the intercept was -6.31. In other words, in 2004, one point of OBP was worth 0.01685 runs/game or 2.73 runs in a 162 game season. One point of SLG was worth 0.01283 runs/game or 2.08 runs in a 162 game season. Each point of OBP was worth 1.3 times as many runs as a point of SLG. Can we use this to say something about the value of individual events, such as walks and home runs? Not exactly---there are many ways in which the basic events could be affecting runs to result in these numbers.

To further explore, I considered the value of individual events, such as a single, double, walk, home run. What are each of these events worth in terms of runs? Or, at least, what were they worth on average in 2004? I used singles, doubles, triples, home runs, walks+HBP and outs as my variables. I calculated outs as AB-H+CS+GDP+SH+SF. The resulting weights are:

Home Run1.84
Walk or HBP0.31
These numbers are a good reminder that regression simply finds the weights that best explain the data. The regression doesn't "know" that a double is worth more than a single. And, apparently, in 2004, a double wasn't worth more than a single. Maybe it was common for pitchers to walk the next batter; maybe doubles occurred with the bases empty moreso than singles; maybe pitchers tended to give up doubles in isolation and were more likely to give up a string of singles. Whatever the reason, singles correlated slightly better with runs than did doubles. Investigation of other years might find that this was a fluke; or, maybe doubles aren't worth as much as is usually thought. Another surprise is that triples come out nearly as valuable as home runs. A home run certainly drives in all base runners plus the hitter; a triple also scores all the base runners, but the hitter may not make it home. Worth noting is the fact that triples are somewhat rare---there were only triples 898 in 2004, compared to 5451 home runs. Part of the problem may be a small sample size.

Looking at these individual event numbers compared to the OBP/SLG numbers, we can see why a point of OBP would be more valuable than a point of SLG. How many runs is an on-base event worth? That is, what's the average number of runs we should expect when a hitter gets on base (gets a hit, walk or HBP)? That's easy: weight each event by it's run value and divide by the number of events. There were 29,254 singles, 8,919 doubles, 898 triples, 5451 home runs and 18,062 walks+HBP in 2004. We find that an on-base event is worth 0.6758 runs:

(29254*.67 + 8919*.62 + 898*1.71 + 5451*1.84 + 18062*.31)/(29254+8919+898+5451+18062) = 0.6758
SLG is total bases divided by at-bats. What is a "base" worth? To calculate this, we weight each hit event (single, double, triple, home run) by it's run value (as calculated by the regression) and divide by total bases. We find that each "base" is worth 0.5126 runs:
(29254*.67 + 8919*.62 + 898*1.71 + 5451*1.84)/(29254+2*8919+3*898+4*5451) = 0.5126
Note that the ratio of these two values 0.6758/0.5126=1.3 is exactly the value we found when we ran the regression with only OBP and SLG.

Is a point of OBP more valuable than a point of SLG because a walk is nearly as valuable as a home run? No. Quite the opposite, in fact. A home run is worth about six times as many runs as a walk. It looks like the main reason why OBP gets a larger weight than SLG is the fact that singles provide more runs per base than extra-base hits. The high value of triples and home runs make up for the low value of walks/HBP in the calculation of OBP. But, there's no factor in SLG to make up for the fact that extra-base hits aren't worth as many runs per base as singles. So, a point of SLG ends up being worth less than a point of OBP (at least for 2004). Hopefully I'll get a chance to come back to this and look at additional data...

Jason Rennie
Last modified: Tue Nov 14 11:29:00 EST 2006