On Tuesday, Justin Inaz wrote an article for The Hardball Times laying out who he would like to see on this year’s All-Star team, using projection data rather than seasonal data to select the players. I found his article interesting, but disagreed with the premise, so we argued about what the All-Star Game should be. That conversation is below.
Dave: Over at THT, you relayed a sentiment about the All-Star Game, and in particular, your dislike for the way it is treated as a reward of small sample flukes. You even called the game itself a “frustrating experience” because of how much luck can influence decisions over which players are named to the team each year. In the article, you suggest that we’d be better off if the teams were chosen by systems designed to estimate a player’s current true talent levels, with 2011 year-to-date stats playing a small role in who is selected to represent their teams at the event.
First off, am I re-stating your position accurately? Is there anything you’d like to add to that summary, or an aspect of your view that I left out?
Justin: My position is that All Star Games should be about showcasing the best players, not the players who have gotten lucky. Whether we use projections or not is a matter of personal choice (maybe the alternative is scouting? Or some combination of both?), but projections are a good way to estimate player talent level.
I actually think that a lot of fans (a majority?) agree with the notion that All Star Games should be a showcase of the best talent. That is how they seem to be advertised. It’s just that we (fans, media industry, etc) put far too much weight on a player’s current-season statistics when judging talent. 2011 statistics do matter, and we absolutely should use them as part of the evaluation process. But we shouldn’t ignore what players have done, for example, in the past calendar year, or in prior years. As it is, second-half performances get almost no consideration in All Star selection, which I think is absurd.
Dave: See, that’s a notion that I don’t actually agree with. To me, the best players in baseball get showcased every night on Sportscenter, MLB Tonight, and every website in existence that covers baseball. I don’t know that anyone needs more Alex Rodriguez in their life. By going to an estimate of true talent level, we’re essentially guaranteeing that the rosters will be pretty similar from year to year, and that just seems pretty boring to me.
The All-Star Game is supposed to be a showcase for the sport itself. I don’t really care to see the same group of players every year, even if they are the best. It’s an exhibition game that doesn’t mean anything (sorry Bud), so getting the most talented roster possible isn’t something I’m overly concerned with.
Justin: The rosters would be similar, but not identical. There are players who are clearly the best at their positions in their league, and they’d be annual mainstays. But there are also positions at which there are 2-4 players who are all within spitting distance of each other in true talent, and they’ll swap out from year to year. Injuries will also liven things up a bit. This year compared to last year, I have Tulo overtaking Hanley, Zobrist over Pedroia, Yunel over Jeter, Hamilton over Crawford, and Bautista over Cruz. And that’s just the position players.
Also: Alex Rodriguez didn’t make my All Star team, at least not as a starter. :)
Dave: Okay, so, you don’t like rewarding luck. That’s an understandable sentiment, so let me ask you this – do the playoffs frustrate you as well? If something as relatively harmless as All-Star roster selections being influenced by two month samples bother you, how about handing out the title of World Champion to a team that plays well for one month? Baseball has been rewarding small sample performances with their most important prize for over 100 years, but while people will talk about how the winner of the World Series isn’t necessarily the best team, I’ve never heard anyone suggest we should get rid of the post-season entirely. If you see that as different, how?
Justin: Yes, I think there is a difference there. In the post-season, we’re “choosing” eight good/great teams and letting them battle it out in a tournament that will be decided by a combination of luck and skill. It is meant to be a showcase of the best teams in baseball, and you have to earn your chance to participate in the post-season via a long, difficult post-season. And while teams can get a bit lucky during the regular season via success in one-run ballgames or playing in an unusually weak division, most of the time, the best teams in baseball will get to play in the post-season. We know that the team that ultimately wins the World Series will be influenced by luck, but at least we can be assured that all of the teams who are participating are quality, deserving teams (with the occasional exception like the 2006 Cardinals, but even they had Pujols and Carpenter!).
The way I see the All-Star Game, at least as it “should be,” is as showcase of the best players in baseball. The best players, once chosen, get to battle it out in a single game that will (like the post-season) be decided on luck and skill. Therefore, my preference is that the selection process be one that generally will select the best players in baseball, as opposed to players who had a hot month or two and played above their talent level. Ty Wigginton is a decent little player, but he shouldn’t be the NL representative at third base, even if his 2011 numbers suggest that perhaps he should be.
Dave: I wonder if going to this kind of system would actually make the rosters better, though. Would getting rid of the “fluke” guys not be outweighed by also getting rid of the “real but short track record” guys? Should Jose Bautista have not been an all-star last year? How about Albert Pujols as a rookie? Losing these guys, who do get elected now and wouldn’t under your system, might be more harmful to having the best player’s possible on the rosters than tossing out the few fluke guys who sneak onto the roster due to a few decent months in April, May, and June.
Justin: Well, recognizing outstanding rookie players is something that projection systems can do, if the player’s minor league track record is strong enough. But yes, many of those guys would likely be left off that first year. My guess is that we get more short-term fluky guys than Jose Bautistas. Bautista, after all, has been the ultimate exception to just about every rule as far as projecting future performance goes. And Bautista’s the guy in RF this year regardless, just as Pujols would likely have been by his second year.
The question really is how much real information can we get out of 350 PA’s or thereabouts. Pizza Cutter gave us one evaluation, and it’s pretty conservative. You’re saying we can add a few diamonds to our All Star Games using first-half performances. And that’s probably true. But we’ll also get a lot of coal. I’m not at all confident in our ability to identify which is which in small samples.
Dave: Would your opinion be the same if the event was held after the season like the NFL’s pro bowl? I realize that no one cares about the Pro Bowl, but I also never hear anyone object to players being elected to that game based on solely how they did in the current season. Sure, they’ve played the entire season at that point, but there’s still luck in a 162 game sample. Are you mostly against rewarding players for flukes because they’re so much more prevalent in half-season samples, or would have you have the same objection regardless of when the game was played?
Justin: I would have far less of a problem with the 162-game sample being the deciding factor. First, as you acknowledged, you would have doubled your sample size. Fluky performances would be less prevalent, and you’ll more often get agreements between projections and full-season performances. Second, and perhaps more importantly, every pitch a player sees or throws would be considered when determining All Star representation, rather than ignoring half of a player’s work each season.
If people want to use the calendar year splits here at FanGraphs to determine who our All-Stars are, I have no real problem with this. I can absolutely see the argument to using the All Star Game to recognize the best annual performances, period. But the best 2.5-month performances? No, I just don’t get that.
Dave: What if you viewed the game not as a showcase of the game’s best players, but as a reward for performance during a specific timeframe? Would you be frustrated if some parent bought their kid a puppy as a gift for getting a 4.0 GPA during their first semester of school? Would you tell them to wait for their kid to regress to the mean in the spring?
At nearly every level of the sport, all-star teams are named after the season is over and are a reward to the player’s who had the best season. It’s only at the professional level that these games are moved to mid-season for marketing reasons. I don’t know that the timing of the game should be a good enough reason to change what they were invented for in the first place.
Justin: If it’s a reward for performance–and I can be happy with that being our definition of who deserves to go–I’d like the performance to be over the past calendar year. As I said above, I find it completely absurd that second-half performances have little to no bearing on All Star team selection.
Here’s an example of what I’m talking about. Ryan Zimmerman is probably the best third baseman in the National League. The projections have him as the pretty clear choice for that slot. But he’s missed a lot of time this year with an abdominal injury. Does he not deserve to go? Well, he ranks 18th in WAR in 2011 among NL 3B’s, so if 2011 stats are your criterion, the answer is no. In contrast, he is second among NL 3B’s in WAR over the past calendar year behind Chase Headley (4.9 WAR vs. 4.5 WAR), but has done that in 233 fewer PA’s. I think Zimmerman should be the guy, and that the All Star Game is more fun with him at 3B than with Chase Headley or Ty Wigginton.
Dave: And finally, do you really not find joy in watching the career journeyman walk around with his kid and his video camera, smiling from ear to ear, realizing that he’s a Major League All-Star for the first time in his career? Do you not think of how proud he is to call his parents and tell them that their boy is going to the ASG? Do you really prefer watching a guy who has been there 11 times beg out of the game because he hurt himself sitting on his overstuffed wallet? Why do you hate happiness?
Justin: Coincidentally, my 5-year old accused me of hating (her) happiness just last night, so it must be true!
I think these things you mention are nice side-effects of the current system. But to me, they are not a good enough reason to keep things as they are.
Here’s the thing: we often act like the All Star game is meaningless. But it isn’t. The outcome of the game is (or should be–I hate the “This time it counts!” schtick), but getting selected as a player does matter. It’s one of the major ways that we recognize outstanding players. Being selected to the All Star Game stays with a player as part of his narrative for at least a year afterward, if not more. You hear about it on broadcasts all the time. And All Star Game appearances are always–always!–brought up as supporting (or dissenting) evidence when discussing merits for Hall of Fame induction. They might even influence contract negotiations or arbitration hearings. Does it influence “real” games or seasons? Aside from the home-team advantage in the World Series thing, no, and it shouldn’t. But baseball as an industry (and baseball fandom) is about more than just what happens on the field, and that’s where the All Star game has relevance.
There is a compromise position here, and I’ll mention it because if I don’t then someone else will: All Star Games are, and should be, popularity contests. If we agree to that, then it really does (and should) just come down to who gets the most votes. But judging by the number of All Star “snub” articles that are written every year once the teams are announced, I don’t think most people really believe that. I think people want the best players to be in that game. Using projections is the way that I prefer decide upon who those best players are.
Dave: Can you name a single deserving player that has ever been kept out of the Hall Of Fame due to a lack of All-Star appearances? I know that voters look at things like All-Star appearances in their decision making process, but I’ve never seen a case where a guy who is a strong candidate in every other aspect was denied because his spot got taken by a worse player in some Midsummer Classic.
Justin: I said it was a factor, not the factor. I am sure that a player’s numbers are the cornerstone of any Hall of Fame case for most voters. But celebrity, as at least partially evidenced by All Star Games, does matter. I don’t have a linkable example to provide right now, but I seem to remember seeing that as one argument made against Gene Tenace, for example, who appeared in just one All Star Game. His is not a slam dunk case, but he ranks 17th in WAR and has the highest era- and park-adjusted wOBA of any of the top-25 catchers. Again, I’m not saying it’s a major determining factor, but it is part of a player’s portfolio of accomplishments, and probably does influence voters. If you want me to quantify that, I can’t.
Dave: I think we could even argue this the other way. By going with multi-year track records, you would miss out on guys who broke into the league and were stars right away. We’d lose out on having guys make the All-Star team in the first X number of seasons of their career, and for some truly great players, we might even reduce the number of times they make the roster.
Justin: Perhaps. I think that some great players already make the roster more than they should. Although that’s for popularity reasons, not small sample performance reasons.
To turn that argument back at you, doing it your way would miss out on guys who broke in, struggled over the first month or two, but then went on to have a brilliant last 4 months of the season…especially if they got a slow start again their second year! Had he not gotten injured, Buster Posey might well fit this mold, and he’s right there with Brian McCann in my projections spreadsheet.
Dave: I’d argue that the joy of rewarding a guy who might otherwise never get that kind of attention is also something that I put a pretty high value on. James Shields will probably make the All-Star team this year, and I doubt he would under a system that valued multiple years of performance. I’d imagine Shields will be pretty thrilled to play in his first ever All-Star game, and given that he’s a pitcher and an inconsistent one at that, this might be the only chance he ever gets to go. What if Shields blows out his elbow in September? Are we really okay that he never got to experience the All-Star game despite being one of the league’s best pitchers so far in 2011?
I’m not. I want to see players rewarded for their performance, even if it is a fluke.
Justin: I’m thrilled for Shields if he makes it. Any person who fancies themselves a sabermetrician has to be thrilled to see his luck reverse after what happened to him last year. It’s become one of the most dramatic examples of DIPS in action!
But is he one of the top 5-7 AL pitchers, in true talent, right now? He might be, but he also may be a bit outside of that tier (my spreadsheet has him in the low 20′s, but I’d personally place him in the 10-20 range). If he’s not “deserving” (by the criterion of true talent) and still makes it, that means that someone who *is* deserving does not make it. And what if that player, whoever he is, goes on to blow out his elbow and doesn’t get his day in the sun? That would seem to me to be the greater loss.