On Regression and Appreciation
Josh Hamilton admires his lone World Series home run on Saturday, October 30th.One of the oft-used criticisms of those who prefer a more analytical approach to their baseball enjoyment is that we lose track of the beauty of the game by immersing ourselves in numbers (in our mother’s basement). Glossing over the caricatures, I want to instead spend my time discussing how the more analytical approach can, and should, result in a greater appreciation of the game.
The 2010 Texas Rangers reached a heretofore unseen level of success in the history of the Texas Rangers. In my time following them, I have encountered fans of every sort; from the retiree who has attended home games almost every year since they came to Texas to the proud new parents taking their kid to their first game; from the baseball fan a few states away who picked this team years ago to the diligent bedside visitor spending quality time with her ailing father-in-law. Baseball brings us together in strange ways; the 2010 Rangers especially so. The success after so much waiting is what many of us will appreciate when we look back. Our history can’t help but remind us that these moments should be savored.
As the shortened off-season gains momentum, we will all undoubtedly get pulled into the Chase for Cliff Lee, free agency and trade rumors, and that interminable period in late January when it seems that spring training will never arrive. As we get excited for next season, I want to remind us all of an important, oft-forgotten, point of baseball analysis:
Regression happens.
It is very easy to look at the current Rangers team, retain Cliff Lee, add in free agent or trade target du jour and assume that since the Rangers won 90 games and the division last year, that they are likely to win 90-plus games this year. While fun and easy to do, this is actually a quite poor approach. Veteran baseball fans will tell you that it’s because every year is a new year and regale you with stories of the Hope of Opening Day.
In contrast, the more statistically inclined baseball fan will tell you that it’s because every year is a new year. Hmm, perhaps we’re not quite so contrary after all? Our reasoning for this strongly involves the concept of regression toward the mean. Since the use of statistics means selecting out certain numbers to make your point, it’s important to realize how likely it is that the statistics you’ve chosen will project going forward. In most cases, the answer is less than you expect.
Let’s take an example. Let’s say P. Francis, an unknown Rangers AAA player comes up to the major leagues in September and gets a hit in his only at-bat. Should we expect this (probably handsome) player to continue to bat 1.000 for the rest of his career? Of course not. Intuitively, we all understand regression to the mean in its most extreme forms. The unsaid point here is that there is (likely) a gap between P. Francis’ true talent level and his current batting average.
This happens with most every major league player as one projects forward based on their current statistics. The truth is that the most elite performances from one season are not repeated the following season. This is because though their true talent level may be very good, there is a certain amount of luck involved to achieve MVP or Cy Young production; that luck is unlikely to continue the next year.
The 2010 Texas Rangers had some truly elite performers. Josh Hamilton’s 8.0 WAR was worth almost an entire win over the second-place AL MVP candidate Adrian Beltre (7.1 WAR), and he did that in 23 fewer games. Whether you want to regress this to his career mean or the league mean, the point is that it is unwise to expect such an amazing season in 2011. Cliff Lee had a 7.1 WAR season, significantly better than Justin Verlander’s 6.3 WAR. This is not to say that Josh Hamilton or Cliff Lee will be an average player next year. However, they are likely to be closer to it than our fan instincts would first have us believe.
How much to regress these numbers and what to regress them to is a topic for another time. The point that I wanted to make here is that as we all get swept up in the fervor of the Hot Stove, don’t forget to take time to appreciate how truly great some of the 2010 individual performances were and how they all came together to form such a memorable team. For veteran fans and statheads agree on this: April brings a new team, a new season.
It’s just that this time, it has an entirely different meaning for Rangers fans.
[Editor's note: If you're unsure as to exactly who Prashanth Francis is, click here for the explanatory website announcement. Prashanth is currently pursuing his MD/PhD at Texas A&M, doing biochemistry research in Houston, TX, but today stops down to make his BBTiA debut, and I'm excited to welcome him to the active writing roster. -Joey]
Analysis, Hot Stove Season, The Off-Season
Reader Comments (34)
And now for something completely different: Buck signed with the Marlins.
I agree. I am one of those who speaks heretically and opposes paying the money for Mr Lee. He is a wonderful pitcher but I feel it would gut our club and it's chances for the future. I would much rather we addressed older aging players and their position play that in several cases was sub-par, or the position is still unsettled. Pitching wins championships but if you don't field and hit, you don't win championships, either. To that end, Mr Francis, who is coming up as a rent-a-pitcher next year that we could fill in a playoff spot with?
Diogenes - I also noticed this morning that John Buck signed with the Marlines for $18m/3 years.
I am not a Vmart fan because of what I have read about his poor defense. People talk about him playing at DH and even some 1B, but then who is the catcher? I guess they resign CJ's personal catcher again. I am beginning to think trade instead of FA signing perhaps.
Prashanth Francis - Good article and BTW Gig'em Aggies!
An interesting aspect of your discussion is that this would be a very bad year to attempt to extend Hamilton as some have suggested. Undoubtably he will regress back towards his mean next year. So a better look at his injury track record will yield a better projection of how many years one might optimally wish to buy. Also, it will be a lot cheaper to extend him when he is not coming off an AL MVP year. What are the odds he would repeat.
It also points out the problem with signing Lee. How many years can a club really afford (other then the damn Yankees) as he slowly regresses? Think The Face for an example. By afford I mean both paying and playing because if they are paying that much he will be playing.
You know, regression happens the other way, right? Not everybody performed above the mean. Many were below it and should naturally regress back to it in the same way.
It could be argued that the Rangers did not perform to their regular season expectations given this division's overall quality. Injuries could explain that, but I think it has something to do with the fact that you had very little production from a lot of key positions this year.
I would expect this TEAM to regress more to mean...meaning they could actually regress and yet improve upon their 2010 record and results. Better overall health (Cruz? Kinsler? Hamilton?), "prospecty" type players coming more of age (Scheppers?), young players taking the next step (Holland? Moreland?); pitchers improving to their means (Feldman?), and an entire season of capable players actually providing some production at those 2010 organizational holes (Catcher? 1B?).
Even if they don't resign Cliff Lee, there's more than enough reason to believe this team will improve in 2011, especially if they make even a modest move (or two) during the hot-stove season.
Welcome, P-Fran.
Great stuff, P -- congrats.
Bats and Balls makes a good point. I still contend that this team underachieved in the regular season. This was a massively talented team. With that said, you have to build a contending team with the playoffs in mind, and that means (in this new era) at least three very talented starters, one of whom must be of the shut-down variety. It is not a stupid suggestion made by Andy, above, that the better bet might be going with the young rotation, minus Lee, and looking for the rental pitcher available at the July trade deadline to top the rotation. The problems with that approach are (1) you may not be in contention in July without Lee, and (2) you can't keep trading away top prospects for short-term rentals (unless you're awfully good at drafting with the compensation picks).
Glad to have you hear Prashanth. Any preferences on nicknames/abbreviations?
Good discussion, guys - and welcome PF. I'm mainly responding to jd21's comments here... What you bring up is exactly why I'm much more anxious over the idea of trading away a bunch of guys to get Greinke than I am about paying a huge contract for Lee. The latter case only takes $, even if it could potentially hamper the team when big payouts are due other players. But you can't really play the what if game like that - you have to just plan year by year and make smart decisions. But if getting Greinke (and he's close to his FA season mind you) costs us Holland, Perez, Profar and another prospect, as Newberg suggested this morning, man I have to seriously think about that.
I keep reminding myself that Texas got to where they are by being on the other end of the 4 or 5 for 1 deal, and it doesn't take long to completely empty out one's farm system of talent. Just take a look at Houston, for instance. That franchise is where we were about 8 years ago. Man, I sure as hell don't want to see Texas go back to that kind of business.
Oh damn. I meant it's good to have you here Prashanth.
You know, regression happens the other way, right? Not everybody performed above the mean. Many were below it and should naturally regress back to it in the same way.
Okay, but the question is, would the regression of the "many" back towards average be strong enough to offset the perceived overperformance? In a vacuum where no signings/trades were made, I don't know that I would expect the catching position to regress very much, given the on-hand talent. If the Rangers sign a new everyday catcher, that's not really the "good" form of regression so much as it is an outright upgrade.
There's first base, of course, where you would expect Mitch Moreland to give you a decent-sized boost given a full season of production, but I don't know that the internal improvement there and any positive regression by Elvis/Borbon/Kinsler is sufficient to offset the other kind.
It would be foolish to expect any player to ever put up an 8.0 WAR type season. I wouldn't expect it of Babe Ruth even. I might not be surprised if it happened, but I certainly wouldn't expect it. It's just that hard to achieve, even if the talent is there.
Thanks for the welcome everybody!
@Andy: It's hard to gauge whether Cliff Lee will be paid too much as we haven't seen the contract offers yet. I do feel that the 2011 Rangers, like the 2010 Rangers, are a bit different from most teams in Rangers history in that they have very good No. 2 and No.3 pitchers in their rotation and have a relatively unique need for a true ace. As far as a name for the back end who might give some depth, I'm much more interested in taking a flyer on somebody like Jeff Francis than I am ponying up for Carl Pavano.
@Jon: I didn't want to get into it in depth with the article, but there is a difference between the downward trend of numbers we see as a player ages and the statistical concept of regression to the mean. Regression to the mean is trying to account for luck while age-related drops in production are a loss of true talent.
Tigers give Benoit $16M... which is absolutely crazy! i think Benoit was a good pitcher when he was the Rangers and he obviously had a great year... but c'mon, $16m for a bullpen arm that's not even a closer... nuts!
I haven't heard anything about Carlos Pena. Wouldn't he be a nice consolation prize if we can't sign someone like Victor Martinez? I know he sucked this year but maybe a move to a great place to hit would bump that up... not to mention him mashing to the tune of 40 HR/100 RBI.
He would have to be affordable... and some idiot GM is bound to give him a fat contract... but maybe, just maybe he'd be willing to sign a 1 yr deal to boost his value and hit the market again next year. And I'm not suggesting Moreland isn't the long term answer at 1B... I'm merely pointing out that Pena may be a nice addition to the Rangers lineup and sort of bridge the gap until Moreland is a proven commodity.
The more I read about Cliff Lee, the more I think he's not going to be someone the Rangers can afford. It sounds like he's going to dig in his heels and wait for the top bid, be it in Texas, NY, or Siberia.
I like the idea of trying out Feliz at SP, moving Ogando to closer, bring up Scheppers, and sign an impact bat or 2.
We do need a TORP... and maybe Grienke becomes a reality... but if we went into next year with CJ, Colby, Hunter, Holland, and Feliz... I'd be ok with it. If the Feliz expirement doesn't pan out, JD tracks down a good pitcher on a bad team and offers a reasonable package of prospects.
@BatsandBalls: Yes, regression can and will happen in a positive direction. It is, however, unlikely that the Rangers will have players regress 'upwards' enough to completely account for the downward regression, in my opinion. Predicting injury/health is tricky. After one injury, you aren't likely to consider a player injury-prone. But at this point, don't you have to factor in a certain about of time on the DL for players like Hamilton, Cruz and Kinsler? Unfortunately, these are our best players. Fortunately, if any of them are ever all healthy for an entire season, they are capable of putting up an MVP-type season. Also, don't confuse 'upward' regression with progression, which is what I would call the expected improvement from the still very young Andrus next year.
@Dave: No thoughts on nicknames, I find they have to happen naturally.
@tball: You're right of course...though it certainly seems like Pujols is trying to prove us all wrong. He can't really be a true talent 8.0 WAR guy....can he?
"@Jon: I didn't want to get into it in depth with the article, but there is a difference between the downward trend of numbers we see as a player ages and the statistical concept of regression to the mean. Regression to the mean is trying to account for luck while age-related drops in production are a loss of true talent."
Yes there is. I didn't elaborate, but in Hamilton's case both forms of regression are in play.
I seriously doubt that he will win the AL MVP (regression to the mean) even if one was to disregard the age factor.
I also really wonder about his overall health and durability. He missed time and was trying to space out his shots. (age realated drop in production)
In summary I think arbing Hamilton this year is the way to go; not extending him. Love to extend him, but one would be definitely buying at a high point.
Joey:
I don't disagree...not at all...but I think a conversation about regression using WAR should probably be balanced against the fact that, collectively, Hamilton, Kins, and Cruz will probably play more games in 2011 than they did in 2010.
I think there are internal improvements to be made, but if you measure regression of the PLAYER, it's certainly hard to quantify. But if you measure regression of the POSITION, then the position can INCREASE to the mean through a good acquisition in the off-season. Call it an "upgrade" if you want, but it's still regressing the POSITION by external factors. And really, when you have a couple of positions that didn't even perform to replacement level (I'm guessing there, but not probably far off), then it's not too difficult to anticipate that an organization in this position (close to winning it all), with these resources, will probably fix that for the 2011 season.
On my morning drive, I got a little fed up with the Musers talking about Cashman is selling Lee on the Yankees being championship-ready over the next 5 to 6 years while the Rangers have a smaller window of opportunity. I think we'd all disagree that's BS, but it's irritating that the Musers didn't disagree with that assessment, as if it's a legitimate selling point for the Yankees. I think that perception comes from people saying that team's regression will be negative, which P.F.'s posting seems to support.
In total, I think this team underperformed during the regular season for a variety of reasons, primarily due to injury (and I didn't even mention Lee being ineffective for a month due to injury). So, with "upgrades," I feel any regression next season will be positive.
@Prashanth - I was more interesting in determining if there were any nicknames/abreviations that annoy you. I'm already familiar with your commenting as GhettoBear04. If anything is game, or if there's just one name that would cause you to hunt me down and kill me in my sleep, that's all I wanted to know.
@Bats and Balls (RE: Musers) - Let me preface this by saying I like the Musers for the most part. As far as the Ticket goes, I'd say the Musers are easily the least baseball savvy show that 1310 has to offer. The Hardline is pretty baseball obsessed. Bob, Dan, and Norm are all in a very deep fantasy baseball leagues. Danny and Sean Bass actually frequent this site. What can we actually cite as proof that the Musers have a clue when they're talking about baseball? It seems like every non-Musers on air personality at the very least reads the Newberg Report. I've stopped worrying about any opinions the Musers have about baseball.
How does far does regression work against the Rangers as it pertains to the A's/Angels? Will upward regression from them and downward regression from the Rangers bridge that ten game gap? I wouldn't think so...actual upgrades and downgrades will still be they key to '11.
The Musers are the best, funniest, most enjoyable radio program I have ever heard in my entire life of 44 years, hands down, no contest. But you don't go to them expecting hard-core sports insight. That's not what they do - and thank God for that.
Oh, the Musers being baseball-idiots is a given...it was just an illustration as to how people get the wrong perceptions regarding how the Rangers just HAVE to go backwards after such a successful season.
My point concerning that was that people confuse "regression" in statistics with the Webster's definition of the verb, to "regress." The latter term means to go backwards (opposite of progress). The former term means to normalize. They are two different things. Nobody ever talks about regression UP to the mean, but they should.
Plus, you can regress anything to the mean. Doesn't mean that a.) the regression is correctly done - it's hard to do a regression if you don't know the "mean" - or that b.) the regression is relevant to anything - overall team improvement should be the debate, not whether Hamilton is MVP again or we have another ROY candidate. In other words, it can be strongly argued that by forming other regressions (on performance of the batting order or by position), that this team can improve. Statistics do not have to be applied solely on internal factors. Simply put, I think next year will be more special than this year as there's likely to be more "progress" than "regress."
But putting that aside, according to P.F., we are supposed to appreciate this year's "elite performances" because regression is inevitable next year, correct? But where's the regress (dictionary term) going to occur?
- In the regular season, Hamilton played in 133 games in 2010. He didn't get his average above .300 until early June and he accumulated his WAR, for the most part, in 3 months time. It was an MVP season, but that's pretty much what Hamilton is. I think his "mean" is pretty substantial, at least more than given credit for.
- Lee earned the majority of his WAR before coming to Arlington. The only "elite" we really saw from him was in the play-offs. If we land Lee, I'd hope that any regression to of the GOOD type; or I should say, at $25 million per year, it better be!
- Cruz was pretty good (4 walk-off homers is pretty much appreciated). But he played 108 games...coming off 3 DL stints that had to hurt his stats each time (getting the timing back). Cruz still had a 5.1 WAR...which projects to 7.6 WAR had he played 160 games.
- Kinsler earned 3.1 WAR in only 103 games this year...and he definitely suffered timing issues having missed all of April and August. Even so, if he played 160 games he would have earned 4.8 WAR.
Assuming we sign Lee, we'll likely see overall improvements in 3 of the 4 players listed above next year...and one, Hamilton, it could be argued that he might actually improve upon his WAR numbers next year.
My argument here is that the only player P.F. mentions that could likely go backwards next year is Hamilton...and that's subject to debate. Lewis and Wilson would be good debate candidates as well, but they weren't mentioned by P.F - and I believe we have reason to believe they'll improve next year anyway. And even Hamilton does NOT retain his form, I doubt that any of us, stat-geek or otherwise, will ever NOT appreciate what we saw this year.
BTW, pertaining to the thought that Josh Hamilton can't possibly repeat his 8 WAR season...
Witness the following contemporary players:
- Alex Rodriguez - 9 seasons of 7.8 WAR or better
- Albert Pujols - 8 seasons of 7.7 WAR or better
- Barry Bonds - 13 seasons of 7.7 WAR or better
- Chase Utley - 4 seasons of 7.3 WAR or better
- Ken Griffey Jr. - 6 seasons of 7.1 WAR or better
- Chipper Jones - 4 seasons of 7.4 WAR or better
- Jeff Bagwell - 5 seasons of 7.0 WAR or better
- Frank Thomas - 4 seasons of 7.3 WAR or better
And these interesting ones:
- Cal Ripken - 4 seasons of 7.4 WAR or better
- Rickey Henderson - 6 seasons 7.0 WAR or better
I think Hamilton compares pretty well to most of these guys, and he's certainly as toolsy as you need to be to put up high WAR numbers. If he can stay a little healthier, I don't see how he's NOT a 7.0 or WAR better type of player.
I have two questions for the lot of you:
1) Are there any comparable players that can help to predict how Josh' body will hold up into his 30s? Off the top of my head, I can not think of anyone that was a substance abuser in his early 20s or a similar precedent that could help predict.. I am too optimistic and/or naive, but I have never seen any player as good as Josh Hamilton. I mean the talent is overwhelming. Bonds in his early '90s seasons was similar.... When old baseball fans tell me that he is the best they've seen since Mantle, I believe them. How does such a wonderful talent age to league average? He strikes me as the type of player that will always be high WAR, even his aged talent is above league average. Injuries are becoming a frustrating habit though. Hamilton will be another of the what if careers. Joe Jackson-like, with what numbers his career might have produced if it had been a full healthy 20 years.
2) Cheap Carlos Peña or Moreland at 1B, all things considered, which way would you go. I think most agree Peña is above average at manning the position, and he is a power threat. Going on the regression toward the mean, you would have to believe Peña just had his worst hitting season and can be expected to hit .240 or so next year.
Tre - As one of the old fans I immediately connect Josh and Mickey.
They also both had serious substance abuse problems. Mantle late in life sobered up and felt great remorse over the problems his alcoholism had caused family and friends. His career was marked by serious injuries as well.
The two players are not that far apart in many ways.
The Ticket doesn't have a single real baseball guy. Norm's probably the closest. Rhyner is a joke. I love the Musers, but they don't know anything about baseball at all.
Rhyner said he was surprised that Buck didn't net more money. Thats all I needed to hear to know how much he knows haha.
I can't imagine anyone who isn't a completely oblivious Yankee's fans stating that the Yankees are a more serious contender over the next 4 to 5 years than the Rangers.
If I'm proved wrong on that then so will every sabrematicin because the Yankees will win division after division and championship after championship with a team that averages more 40+ years old and that has NEVER been done. If they do that then I think every member of the Yankees needs to be tested everyday for steroids of somekind or umps need to examined to see if they are truly impartial.
The Yankees are simply too old for the next 4 to 5 years. Granted many of the players are/were elite players but talk about regressing to the mean? Boston is in the same boat. While Tampa is is in perpetual fire-sale mode Toronto and Baltimore are not. The Yankees in the next two or three years are going to need a new starting 3rd baseman, starting shortstop, starting catcher, closer, at least THREE starters, TWO outfielders as well as replacing large parts of their bench and bullpen. No team can do that. The longer they try to stay contenders the longer it will take to rebuild.
The Yankees may stay contenders next year by adding a couple of FA's but after that give me Toronto or Baltimore
@Bats and Balls - Who said Hamilton "can't possibly" repeat an 8 WAR season? I just said it's foolish to expect it. It's a rare event, with only one player reaching that level this season, and typically just two or three a year. Whether or not a player is capable of reaching that level, it's still special and unique and not to be "expected".
Welcome to BBTiA. Run away as soon as possible.
@Parks - As a fan of yours, I hope your "run away" comment wasn't directed at me.
@T-Ball: It's rare this year; the "year of the pitcher." It's not all that rare on the whole and is pretty common place among the very best, transcendent players in baseball...which I'd argue that Hamilton IS one. A-Rod has come close in almost half his seasons; as has Utley. Pujols has come close in about 80% of his seasons. While I realize you need a track record to perform a regression, that just plays into my point..we simply don't know HOW he will regress. What is normal for Josh Hamilton, a player who's singular talent is just now coinciding with the peak mental and physical aspects of his game?
And if Hamilton did "regress" to a 7 WAR season, I doubt anybody here would lose sleep over it. If regression did occur in that fashion, it'll likely be a 2009-like season where he's injured all year. I doubt it'll be because of performance.
You may say it's foolish to expect it...and perhaps it is. But it's equally foolish to think that he'll be all that far off from it, given his obvious level of talent and abiility.
And in the context of the blog posting, I still haven't heard anybody explain the areas of negative regression this team will experience next year and exactly why this team won't be able to overcome that?
Bats and Balls: Nope. I was just welcoming Pras to the fold. The "run away" comment was a joke.
I don't think it's out of the ordinary to think Hamilton can have another great season. He had the great season where he stole the show at the Home Run Derby. Then last year was a mediocre season, and this year came back strong. I think part of that had to do with his supporting cast this year, with Vlad hitting right behind him.
I go back and forth on how much they should spend on Lee. I do hope that they keep him for a few reasons. First, I saw a stat a while back that showed when pitchers have a dramatic increase in their IP from one season to the next, their performance drops off the following season. I saw this in relation to Chad Billingsley on the Dodgers. CJ Wilson is now in that same group, so there's no guarantee he's going to repeat his season. I'm fine with Lewis, Holland and Hunter in the rotation, and you could fill out the rotation with Harrison, or someone else they can pick up. Feliz is a possibility, although I'd rather see him stay as closer for now.
But more so, is just the continuation. Right before the World Series, I saw the Rangers lead off a sports broadcast for the first time over the Cowboys, which I think is the first time I've seen it happen in the two and a half years I've lived in Dallas. They have the diehard Rangers fans already, but last year's performance captured a lot of fringe and bandwagon fans. So re-signing Lee is a good way to continue to keep those fans in the seats. I still think they'll be successful if they don't sign Lee, but from a momentum standpoint, it would be great to keep him.
Cliff: you touch on a point I have considered the past couple of days, that being: "what are NYY CHANCES of NOT making playoffs the next few years?". Consider BOS: Lackey, Lester, Bucholz, Beckett= pretty salty rotation. No way they're as injured as last season (which they competed well enough in for most of the year); they will be as formidible, and likely more so. TOR: some people said they'd lose 100+ games in 2010 just because of playing in AL EAST; they are coming on no question about it. BAL will also be much improved. TAM...? still will have great starting staff, but how much will they be weakened by a Crawford-less defense. Even so, I can envision NYY struggling over the next few years. Who in the AL WEST will get better fast enough to catch TEX? Just can't see it. ANA will be better, with Crawford; and Morales healthy;, but TEX ought to be able to top them. But WITH Lee, I think they are a juggernaut. NYY even with Lee, no guarantees. Hope Cliffy sees it that way.