Larkin's gonna get the Raines treatment. Book it.

Larkin is my favorite player, but I do not think he gets in. I would call it the "Trammell Treatment", however.

Larkin's gonna get the Raines treatment. Book it.

No way.

Raines: 7 ASG (with the last occuring in his age 27 season), no GG, 1 SS, his highest finish in MVP voting was 5th, and he's getting compared to LFs.

Larkin: 12 ASG (with the last in his final season), 3 GG, 8 SS, an MVP, and he's getting compared against the shortstops of his era.

I'd be shocked if Larkin started with any lower than 50% of the vote on his first ballot.

I think Larkin gets in. It was looking bad for a while with the explosion of SS offense (ARod, Garciaparra, Jeter, Tejada).

But Garciapparra implosion, and all the steroid stuff with ARod and Tejada, probably helps Larkin significantly.

Derek Jeter's most similar batter is Barry Larkin according to BB-Ref. That certainly makes Larkin a Hall of Famer!

Additionally, when you're the best in your league at your position for a decade (demonstrated by the All-Star selections & Silver Sluggers, as well as a few Gold Gloves) you're in. May take 3 or 4 ballots because the counting numbers are little light due to the missed games. Wouldn't be a problem if the voters made a sufficient positional adjustment but it seems they are a little slow on that.

Somebody remember to ping Red Reporter when the Hall of Merit inducts him unanimously this year.

He won't get "Trammelled" because Trammell is still on ballot and everyone voting for him will vote for Larkin. I don't think the HoF elects him this year but I bet he gets 50%.

Larkin is well qualified. His SS cohort is Yount, Ozzie, Appling, Cronin and Negro Leaguers Willie Wells and Grant Johnson. I have him ranked between Appling and Ozzie as the 8th best SS ever.

I like this thread. I hope Barry does not pull any Steve Phillips stuff as part of his TV gigs. That wouldn't help him.

Thanks for the positive remarks. I'm glad to see favorable sentiment about Larkin outside of a Reds blog. I'm obviously biased (he was my favorite player growing up), but I think the data are as solid for him as they could be.
-j

Without a doubt, Larkin's my favorite non-Cardinal of my lifetime. He's a no-brainer first-ballot guy, but I'll bet it takes him three years to get in. 91 out of 99 in steals from '93-'95 ...

Larkin is a clear in the line hofer to me, and I'm one of the guys who regularly harp on missed playing time as an underrated but major issue. I think he does better in the voting than Trammel did.

I think he's a clear HOFer, but I'm not as confident as #3 about his chances.

Larkin should be a no-brainer but nowadays who the hell knows. Raines and Trammell are no-brainers, too. So I have no clue. I guess responding more to fear than anything, I think the hyper-exclusivity of recent times is his biggest obstacle, this utterly and completely phony notion of what the HoF is. It is a place where historically guys who were half of Raines and Trammell and Larkin got to go, but the writers at least are in total denial of what the HoF already is, and so they want to keep it something that it hasn't been since 1946.

So anyway, I am afraid Larkin may have a very long wait. As for me, first ballot.

So anyway, I am afraid Larkin may have a very long wait. As for me, first ballot.

I'd guess 3-5 ballots. "Average" HoFers don't generally go first ballot. If you make the distinction he's probably not a "first ballot guy", but should get it soon thereafter.

I think the hyper-exclusivity of recent times is his biggest obstacle, this utterly and completely phony notion of what the HoF is. It is a place where historically guys who were half of Raines and Trammell and Larkin got to go, but the writers at least are in total denial of what the HoF already is, and so they want to keep it something that it hasn't been since 1946.

I don't know. I think it's more simply a matter of not properly valuing guys. It's hard to argue that the problem is that the BBWAA is tightening its standards when they're electing Jim Rice and Bruce Sutter and electing Kirby Puckett on the first ballot. They just don't do a very good job of figuring out whether Raines or Rice was a better player.

That said, I'm cautiously optimistic on Barry Larkin. I don't think his value is all that hidden. BBWAA voters like to look back at All-Star Games and MVP votes as key measures of how good a player somebody was. And as #3 points out, Larkin does very well there. I know he looks like he just misses on some key stats - .295 career batting average, 198 career home runs - but I think that even most BBWAA voters understand that a .295 hitter is just a little worse than a .300 hitter (and perhaps also that Larkin was pretty obviously a true ".300" hitter for most of his career).

Aside from the curious case of Alan Trammell, the BBWAA also seems to generally like shortstops (Ozzie was 1st ballot, they elected Luis Aparicio, there's already buzz about Omar Vizquel). They do tend to like their shortstops to be a little better fielders, and Larkin could fall into the gap of not as good defensively as Ozzie, not as good offensively as Ripken that has trapped Trammell. But, on the other hand, Larkin won 3 Gold Gloves, and I think most voters are going to understand that he could have won more except that his career overlapped with the greatest defensive shortstop ever.

I like all this optimism about my man Barry, but I am skeptical. I think for many BBWAA voters, he will be seen as "a great guy, a very good player, but not enough of an 'impact' guy to be in the Hall." I foresee arguments like, "He wasn't the hitter Ripken was or the fielder Ozzie was."

Aside from the curious case of Alan Trammell

I think this is the key. They are very similar types of players, from the same basic period of time.

Aside from the curious case of Alan Trammell, the BBWAA also seems to generally like shortstops (Ozzie was 1st ballot, they elected Luis Aparicio, there's already buzz about Omar Vizquel).

They like LONG CAREER shortstops. And Ernie Banks.

Put me in the "don't know if he's getting in" camp. It's a hard case to read the BBWAA on. There's plenty here for them to like -- the MVP most of all but the ASG and GG give them plenty of cover. But they aren't good at positional adjustments, they can easily ding him on missed playing time and what is essentially a peak argument, and they haven't been good at assessing guys who are good all-around players without being great at something. But yes, he'll do better than Trammell or Raines which I guess make his chances for eventual election pretty good.

I think this is the key. They are very similar types of players, from the same basic period of time.

Honestly, (and I don't say this to speak well of HOF voters), I think Larkin's huge advantage over Trammell is that Larkin won the MVP in 1995 while Trammell lost one that he obviously deserved in 1987. I also think that Trammell is much more of a direct comparison with Cal Ripken and Robin Yount, who played in the same league at pretty much the same time, and both of whom won 2 MVP awards.

Larkin was the best SS in the National League from whenever you think he passed Ozzie (I'd say 1990 or so, but YMMV) probably to the day he retired. Now, a lot of that speaks to the fact that all of the good shortstops were in the American League during Larkin's career (Cal, A-Rod, Jeter, Nomar, Tejada), but I think that taking all of this together, Larkin has enough of these odd things that I think BBWAA voters like (winning the MVP, best in his league, lots of All-Star appearances, a handful of Gold Gloves) to make his BBWAA case not THAT much worse than his sabermetric case.

But, as I said above, I'm only "cautiously" optimistic. I'm sure that you're right that we'll see "arguments like, 'He wasn't the hitter Ripken was or the fielder Ozzie was.'" I just think there may be less of those than we see with Trammell.

like all this optimism about my man Barry, but I am skeptical. I think for many BBWAA voters, he will be seen as "a great guy, a very good player, but not enough of an 'impact' guy to be in the Hall." I foresee arguments like, "He wasn't the hitter Ripken was or the fielder Ozzie was."

I'm optimistic about Larkin. He was a slightly better hitter than Trammell and I think his baserunning plus MVP makes him a more "dynamic" candidate than Trammell, who, while deserving, lacks sizzle.

I think 3-5 years and he's in.

retired players, Highest OPS+, 1500+ G at SS,

I'm feeling better about his chances since reading this thread, but initially I was in the "ain't gonna happen" camp. The backhand they gave to Trammell (and the less deserving but still way better than the HOF voters seem to think Lou Whitaker) makes me worry that they just have no idea what the numbers of a valuable shortstop from the '80s-early '90s look like.

They like LONG CAREER shortstops

Larkin played 19 seasons. Now, eyeballing it, only about 8 of those were full seasons (and 2 of those were strike years), so he didn't play a ton of games. But I wonder if this isn't another thing, like the MVP and the All-Star games, that perhaps unfairly helps his BBWAA case, where the BBWAA is going to under-penalize Larkin for the missed playing time and just give him credit for hanging around for almost two decades.

(and the less deserving but still way better than the HOF voters seem to think Lou Whitaker)

Despite no evidence whatsoever, I still believe that Whitaker got screwed by a BBWAA-wide belief that "He's a Hall of Famer, but not a first ballot one, so I'll vote for him next year." Otherwise, I just can't come up with a decent case as to why he didn't even garner the minimum 5% needed to stay on the ballot.

I'm with the 3-5 crowd. But if he doesn't start well, he could face the same problem that will make a Raines induction more problematic. With so many potential HoFers joining the ballot in the coming years, those Rice or Gossage or (I suspect) Blyleven-type slow climbs to induction could be tough to build.

With so many potential HoFers joining the ballot in the coming years, those Rice or Gossage or (I suspect) Blyleven-type slow climbs to induction could be tough to build.

Yeah, but steroid taint will slow a lot of those new guys down. A lot of reporters may want to make at least a first ballot protest vote against Bonds, Clemens, Sosa, etc., leaving some extra votes for Raines, Larkin, etc.

I still believe that Whitaker got screwed by a BBWAA-wide belief that "He's a Hall of Famer, but not a first ballot one, so I'll vote for him next year."

If I remember correctly, Trammell came onto the ballot the year after Whitaker. I think it's plausible to think that some folks thought it'd be neat to see them inducted together, so they decided to not start voting for Whitaker before they had a chance to vote for Trammell too. Then again, given Trammell's vote totals, how many such people could there realistically have been?

like all this optimism about my man Barry, but I am skeptical. I think for many BBWAA voters, he will be seen as "a great guy, a very good player, but not enough of an 'impact' guy to be in the Hall." I foresee arguments like, "He wasn't the hitter Ripken was or the fielder Ozzie was."

That's my worry. I see Larkin as a great player who should go in easily, but some of the decisions by the HOF voters frankly make me question their mental capacity.

Larkin's got a full career in one city, a world championship in a small market, and MVP award. These things should work in his favor, but Trammell had 2 of the 3 things going for him. So did Kirby Puckett, those bonuses worked for him. Larkin is kind of like Kirby Puckett but playing at a more difficult defensive position.

Larkin was a better player, per game played, than Cal Ripken. Ripken was greater simply because he was able to play so much more.

Yeah, but steroid taint will slow a lot of those new guys down. A lot of reporters may want to make at least a first ballot protest vote against Bonds, Clemens, Sosa, etc., leaving some extra votes for Raines, Larkin, etc.

But that's not necessarily going to help those backlog guys. If Bonds, Clemens, etc. are languishing on the ballot when players of their stature would have ordinarily gone in first ballot, it's only going to add to the clutter.

But that's not necessarily going to help those backlog guys. If Bonds, Clemens, etc. are languishing on the ballot when players of their stature would have ordinarily gone in first ballot, it's only going to add to the clutter.

I was thinking more of a "backlash" in favor of those considered to be "pre-steroids". I guess it can cut both ways.

Otherwise, I just can't come up with a decent case as to why he didn't even garner the minimum 5% needed to stay on the ballot.

I can see that, and I have tried not to bash the BBWAA too much, but reading some of their Jim Rice arguments has changed my view of them a bit. It isn't so much that I care that Rice is in the Hall per se, but more that his selection and the reasons given show the ways a lot of these guys seem to think. I remember Larkin's career very well; I followed him and what people said about him very closely. He is, IMO, exactly the kind of guy who, like Grich, Whitaker and Trammell, "doesn't feel like a Hall of Famer." Yes, he won an MVP, (so did Zoilo Versalles) and was on a lot of All-Star teams (so was Dave Concepcion). But Larkin was never "the guy" from a national MSM standpoint, even on his own team, except for maybe in 1995 and who, other than hard-core Reds fans, remembers the 1995 Reds? The Reds revolved around Rose, then around Davis and the Nasty Boys, and then they got Griffey. Larkin wasn't Yount; he wasn't Ripken, he wasn't a flashy defensive guy. He was "just a heck of a ballplayer" and that is not really what the BBWAA seems to be about.

I hope I am wrong, and it seems a lot of guys here think I will be. I am not saying "No way he gets in" but I think if he does, it is going to take a long time.

Certainly, comparing Larkin to other SS's in or out of the HOF is very important, but consider who else will be on the ballot in 2010 (and beyond). Here's pretty much the entire 2010 ballot:

Dawson
Blyleven
Smith
Morris
Trammell
Parker
Mattingly
Murphy
Baines
Raines
McGwire

And the newbies to the ballot:

Alomar
Larkin
E. Martinez
McGriff
Gallaraga
Burks

Who from this list is likely to get lots of support? Well, Dawson and Blyleven are the only returnees remotely close to induction. Lee Smith and Jack Morris are the only other returnees above 22%. Alomar, Larkin, and Edgar Martinez seem like the only newbies likely to get significant support in year one.

- It seems to me that Dawson will probably get in this year, along with Alomar.
- Blyleven, who is almost out of years on the ballot, will probably get in on the 2011 ballot. If Larkin can get 50% or more this year, I bet he gets in on the 2011 ballot, too. Bagwell is the only no-brainer from the 2011 class on the ballot, and he'll probably get in on the first try. (Larry Walker and Rafael Palmeiro are new on the 2011 ballot, too.)
- 2012 has, with the arguable exception of Bernie Williams, nobody who will even make the 5% minimum to stay on the ballot. It is my opinion that anybody appearing on the 2010 ballot - whether returning or newbie candidate - that can't get it in by 2012 is unlikely to get in at any point. Why? The 2013 ballot (Bonds, Clemens, Biggio, Piazza, Schilling, Sosa) and 2014 ballot (Maddux, Glavine, Frank Thomas, Mussina) will be stocking the next generation of HOF candidates. Also, many of the candidates currently on the ballot will be simply running out of eligibility (Mattingly, Murphy, Parker, Blyleven) by 2013.

Bottom line: I think Larkin doesn't get in on year one, but is in no later than 2012.

I think Alomar's the obvious comparison to Larkin, and he could affect Larkin either way. Alomar's got more of the "feels like a Hall-of-Famer" vibe to him, I think (he was compared to Joe Morgan and talked about as possibly the best 2B ever at his peak), which could hurt Larkin by comparison. On the other hand, if you compare their stats, even with a BBWAA bias, it's really hard to see that Alomar is much, if any, better than Larkin. So it's possible that some voters may take a look at Larkin, realize that he was as good as Alomar - who they were planning to vote for because he "feels like a Hall-of-Famer" - and decide to vote for Larkin, too.

But really, who the hell knows how some of these guys figure out who to vote for?

But really, who the hell knows how some of these guys figure out who to vote for?

Before he was elected, a Toronto writer indicated that he would never vote for Gary Carter because, essentially, Carter had once been mean to him - it was about something stupid like Carter blowing off one of his questions. Based on that alone, I think it's safe to say that there are some pretty strange reasons used by individuals to judge whether or not a player is a Hall of Famer.

Guys like Trammell and Larkin get screwed because they don't have one obvious talent or moment they're really famous for and, perhaps even more importantly, they don't have the seasonable benchmarks. Larkin scored over 100 runs twice, never drove in 100, had 30 HRs once, and stole over 30 bases 3-4 times. Those numbers don't jump out at the average voter. Same for Trammell -- he scored over 100 runs 3 times and drove in over 100 once. Rice, OTOH, had over 100 RBIs 8 times, and was over 120 twice and over 130 twice in an era when few hitters were able to put up those kinds of numbers.

I think Alomar's the obvious comparison to Larkin, and he could affect Larkin either way. Alomar's got more of the "feels like a Hall-of-Famer" vibe to him, I think (he was compared to Joe Morgan and talked about as possibly the best 2B ever at his peak), which could hurt Larkin by comparison. On the other hand, if you compare their stats, even with a BBWAA bias, it's really hard to see that Alomar is much, if any, better than Larkin. So it's possible that some voters may take a look at Larkin, realize that he was as good as Alomar - who they were planning to vote for because he "feels like a Hall-of-Famer" - and decide to vote for Larkin, too.

But Alomar has the defensive reputation that Larkin doesn't.

I think an advatange Larkin has is being on TV. His co-workers can constantly talk up his comps, and he will be seen more as "one of them", so I think it enhances his chances.

I think an advantage Larkin has is being on TV.

This will get Steve Phillips in the executives' wing some day as well.

Man, I would vote for at least eight of those guys on the 2010 ballot. (Blyleven, Trammell, Murphy, Raines, McGwire, Alomar, Larkin, McGriff) You could convince me to vote for Dawson, Smith, or Martinez.

But Alomar has the defensive reputation that Larkin doesn't.

If you believe play-by-play defensive metrics, Alomar doesn't deserve this. Dan Rosenheck, who put together a comprehensive WARP-type measurement, argued that Alomar was below his in/out line for HOF/HOM basically because Alomar wasn't nearly as good defensively as his reputation and because the replacement level for modern 2B is comparable to 3B (Larkin, by comparison, is well over his line - far closer in value to Ripken, I believe, than to the actual borderline of the Hall of Fame). Not that there are going to be many BBWAA voters who will cite Dan Rosenheck's WARP numbers when putting together their ballot.

The Hall of Fame will cease to exist in my mind if Larkin isn't inducted.

I don't think he'll get in on the first try, but I think he'll get in at some point. The MVP Award, and his role in the 1990 Reds World Series Championship will help him with the BBWAA. It's a small sample size, sure, but his career postseason line was outstanding - .338/.397/.465 for an OPS of .862 in 17 games.

Guys like Trammell and Larkin get screwed because they don't have one obvious talent or moment they're really famous for and, perhaps even more importantly, they don't have the seasonable benchmarks. Larkin scored over 100 runs twice, never drove in 100, had 30 HRs once, and stole over 30 bases 3-4 times.

Larkin did have the good fortune to have his 30-HR season line up with one of his 30-SB seasons, so he's a 30/30 guy. That's something, although, again, who the hell knows what that's worth to the BBWAA.

Alomar, Larkin, and Edgar Martinez seem like the only newbies likely to get significant support in year one.

Uh, Edgar? Define significant - unless they start giving HoF votes to USS Mariner posters, I can't see him ever besting 20% or so.

delete - wrong thread

Vortex? I think you meant to post that over here.

Although I suppose it is possible that you think about Barry Larkin in a very different way than the rest of us.

I can't imagine Larkin won't get in. It is (almost) inconceivable. Larkin is similar toa ll those guys who didn't get in but better. He's clearly better than Trammell, better than Whitaker, better than Raines. He'll make it easily.

Vortex? I think you meant to post that over here.

Although I suppose it is possible that you think about Barry Larkin in a very different way than the rest of us.

Damn - that's what happens when you keep two threads open at once. Sorry about that!

Uh, Edgar? Define significant - unless they start giving HoF votes to USS Mariner posters, I can't see him ever besting 20% or so.

I'm with you; I think he'll fall off the ballot in the first or second year.

Alomar, Larkin, and Edgar Martinez seem like the only newbies likely to get significant support in year one.

What do people think about McGriff - both in terms of whether he deserves it or whether he'll get any support? I could see him possibly being in a nice position to pick up some "anti-steroid" support as the last great clean slugger. I have a hard time getting a good feel for whether he "deserves" to be in the Hall.

I don't think he's going to fall off, but I also don't think he's going to have a decent showing. I just expect Edgar to spend years floating around the 10-15% marks, with no notable changes.

vortex

(giggle)

let ME help you out there - barry larkin is indeed a FINE lookin man. definitely NOT no willie mcgee

oh yeah. and of COURSE barry is a HOFer

and it will probably take him a few times to get in. but he certainly is more deserving than jim ed rice, that is fer SHER

Just to put the links here:

Hall of Merit Larkin thread.

Hall of Merit 2010 Ballot Discussion.

The Larkin thread gets into things like the 1995 MVP vote. The 2010 ballot discussion goes on for over 200 posts; most of the Larkin vs. Alomar discussion is on the first page. Plenty of talk about Edgar in that thread as well.

A few years ago, I posted a "Keltner List" summary for McGriff, in which he actually did quite well. I'm not sure if it's still hanging out in the archives, but I'd be happy to redo it for the sake of argument. He'd be on the lower end of first basemen, but if you're a peak value fan, he's got a far better case than Palmeiro.

A few years ago, I posted a "Keltner List" summary for McGriff, in which he actually did quite well. I'm not sure if it's still hanging out in the archives, but I'd be happy to redo it for the sake of argument. He'd be on the lower end of first basemen, but if you're a peak value fan, he's got a far better case than Palmeiro.

vortex

(giggle)

let ME help you out there - barry larkin is indeed a FINE lookin man. definitely NOT no willie mcgee

I'm sure he is, but, um, that's not what I... (crawls under rock...)

Guys like Trammell and Larkin get screwed because they don't have one obvious talent or moment they're really famous for and, perhaps even more importantly, they don't have the seasonable benchmarks. Larkin scored over 100 runs twice, never drove in 100, had 30 HRs once, and stole over 30 bases 3-4 times. Those numbers don't jump out at the average voter. Same for Trammell -- he scored over 100 runs 3 times and drove in over 100 once. Rice, OTOH, had over 100 RBIs 8 times, and was over 120 twice and over 130 twice in an era when few hitters were able to put up those kinds of numbers.

If the HOF voters are so ignorant as to compare the hitting stats of a corner OF/DH with that of a SS, then it's time to blow up the institution. Me, I'm not so convinced it's gotten so bad...yet.

I think the better comp for Larkin is not Trammell, but Sandberg. Sandberg, like Larkin, has the MVP and the pile of gold gloves. And Sandberg didn't get in on the first ballot, but he was inevitable from the first balloting.

EDIT: Damn, Larkin's GG totals are slightly less than Trammell's. Stupid Ozzie playing in the same league. I still think I'm right about Larkin being on the Sandberg path.

Barry's BBREF Comps:

Alan Trammell (921)
Ray Durham (908)
Ryne Sandberg (903) *
Derek Jeter (901)
Lou Whitaker (879)
Roberto Alomar (877)
Johnny Damon (870)
Joe Torre (858)
Joe Cronin (854) *
Julio Franco (853)


He is 118 on the HoF Monitor and 43 on HoF Standards.

Trammell's HoF Monitor also = 118.

Gray Ink: Larkin 66, Trammell 48
HOF Standards: Larkin 43, Trammell 37

I think the better comp for Larkin is not Trammell, but Sandberg. Sandberg, like Larkin, has the MVP and the pile of gold gloves. And Sandberg didn't get in on the first ballot, but he was inevitable from the first balloting.

Well, what Sandberg has over Larkin is in-season durability. Ryne has 11 seasons over 150 games playes, Larkin only 4, and only 2 more over 140 and one of them was pretty poor (74 OPS+), leading to more and better big seasons for Sandberg.

At what point does it get bad enough that you'd rather hang out at the Hall of NOT Fame more?

Larkin could fall into the gap of not as good defensively as Ozzie, not as good offensively as Ripken that has trapped Trammell

Where is the support for the argument that Larkin was as good defensively as Ripken? People talk about his power at the SS position, but the fact that Ripken was also an excellent fielder seems to get ignored.

I think the voters who saw Larkin play and remember him are likely to vote for him. I think the ones who just look at the stat sheet may be less impressed by the yearly and career totals, especially as compared to what some modern shortstops have done. I don't think he makes it on the first ballot, but I can see him benefiting from strong advocacy by his supporters.

Sandberg:

Black Ink: 14
Gray Ink: 144
HOF Monitor: 158

Edit: Barry won both a Clemente and a Gehrig Memorial Award while he was playing.

The MVP Award is sure to be the huge difference between Larkin and Trammell in the eyes of the voters just as having a Cy Young Award led to Sutter's eventual induction and the not having one saw Quiz get dropped from the ballot his first year.

I sure hope he gets elected. Sean Smith's Wins Above Replacement Rankings have him at 58th all time. Seems like a no brainer.

But what do the voters like? I created two models earlier this year. One is called Predicting Who Makes The Hall Of Fame Using A Logit Model
. It gives him a probability of only about 17% of making it. The model took into account career average, number of 100 RBI seasons, all-star games, PAs, MVP awards, world series performance, getting 3000 hits and being a catcher.

The other model was called What Determines Vote Percentage In The First Year Of Hall Of Fame Eligibility? (Part 2). It said it would be 32% for Larkin. It took into account the same things as above plus getting 500 HRs, getting 500 SBs, gold gloves (but not being a catcher).

I sure hope my models are wrong. But this analysis was based on what the voters did from 1990-2009. Maybe I could add something into the model about being on TV (as Chris Dial mentioned this above and has happened for Rice-but I think Blyleven is a TV commentator and has not made it-and no I am not really Rich Lederer).

It seems like there is a class of players like Raines, Will Clark and Bobby Grich who were very good all around players who lack some big defining stat and don't get in. A shame.

It said it would be 32% for Larkin.

This sounds about right. I just don't think a preponderance of BBWAA guys will see Larkin that differently than they have seen Trammell.

OTOH, my predictions about most sports-related issues usually suck, so Barry has that in his favor.

In the logit model I mention above, all-star games have a big impact. If he had 11 instead of 9, his probability would jump to 76%. Feel free to suggest how I could improve my model. Also, I have a post called The Marginal Impact Statistics Have on Hall of Fame Voting
where you can see how big an impact different variables have on the chance of getting electedc.

In the logit model I mention above, all-star games have a big impact. If he had 11 instead of 9, his probability would jump to 76%.

Barry Larkin had 12 All-Star appearances. I think Bill Freehan is the only eligible player with more not in the Hall.

Retrosheet has him with 9 games played in. That is what I used in the initial analysis. Maybe he got selected a few times and did not play. Retrosheet has Freehan with 8. I sure wish Larkin had 12, then he would get in easily, I think. Probability would jump to 93%. I know that seems like too much, but it is a non-lnear model. The other model would jumpt his first year% to 70.

BTW, when is the vote and when are the results announced?

Maybe he got selected a few times and did not play.

Yeah, I got the 12 and the fewer-than-Freehan things from BB-Ref (although, BB-Ref has been acting weirdly for me lately, so I can't actually confirm that right now). Of course, if you estimated your model based on Retrosheet numbers, then you'd need to re-run your model with everybody else's non-appearance ASG's added in, too, so maybe it wouldn't make as much difference anyway.

But really, I think that the sheer number of All-Star games (even 9 is a pretty big number) and the fact that he won an MVP award are the biggest things he has going for him with the BBWAA, and it sounds like your model agrees with that.

Larkin has a problem with All-Star games, and that problem is named Ozzie Smith.

Now, I'm not going to badmouth Ozzie. Ozzie is a fully-qualified Hall of Famer, and he had a stretch of 5 years or so in which he was the best SS in the NL, often by a large margin. I was one of the Hall of Merit voters who put Ozzie ahead of Trammell on my ballot when the two of them were both eligible. (Although I admit that last is a very close call.)

But starting in 1988, Larkin was a better player than Ozzie. And the sheer momentum of Ozzie's enormous popularity with the fans won him many more ASG starting spots over Larkin, even after the title of best SS in the league had moved on.

Shoot, I only gave him 7 all-star games in one calculation and I gave him a 1 for the dummy variable of reaching 10,000 PAs (he only had 9057), so his predicted first year % is 34.6, just a little different.

Per BB-REF, Larkin was voted to start in 5 games, played in 4 others, and was selected for but did not play in 3 others. I imagine that this is generally recorded as having been an All-Star 12 times, and that is the number that the writers will use.

Okay, according to Larkin's BB-Ref Page, he was selected to the All-Star team 12 times, broken down as follows:

Starting SS - 5 times: 1993, 1995-6, 1999-2000
Played but didn't start - 4 times: 1988, 1990-91, 2004
Selected but didn't play - 3 times: 1989, 1994, 1997

And I was also wrong about Freehan.

According to BB-Ref, Freehan, with 11 selections, is the only player eligible for the Hall of Fame with 10 or more All-Star Game selections who is not in the Hall of Fame. If either Larkin or Alomar (12 apiece) is not elected this year, they would become the HOF-eligible player(s) with the most All-Star Game selections not in the Hall of Fame.

Edit: And, of course, a coke to Srul. Damn my long-windedness.

I understand the non-linearity of the model, but quite frankly, if the sole difference between candidate A and candidate B is two All-Star games, at the level already established (11 and 9), and the vote percentage predicted goes from 17% to 76%, that model is a broken predictor. That's basically saying that ASGs are the sole causal factor for votes into the HOF. I think you're not properly dealing with something there, and one of those things is: did you control for 2 ASGs in the 60s?

Maybe I will redo the analysis and use all-star teams named to instead of games. Maybe the results will be better. I actually lost some data and statistical output recently and what data I do have on this Hall of Fame voting is at the office, so it may take few days or a week to do it. Watch my blog to see if I get it done.

According to BR, Larkin's salary went from $9,000,000 in 2003 to $700,000 in 2004. Is that a record drop for a player staying with the same team?

I only looked at guys whose first year of eligibilty came beginning in 1990, so I don't think that the two all-star games in one season is an issue. One reason why I only began in 1990 is that if you go farther back, then you start getting huge changes in players and voters, so I was trying to contain that. I tried all kinds of models and the ones I reported were the most accurate. It does seem strange about the all-start games, I agree. In the 2nd model, they don't matter as much. Going from 5 to 10 games increase the % of the vote by 30 points.

I understand the non-linearity of the model, but quite frankly, if the sole difference between candidate A and candidate B is two All-Star games, at the level already established (11 and 9), and the vote percentage predicted goes from 17% to 76%, that model is a broken predictor. That's basically saying that ASGs are the sole causal factor for votes into the HOF. I think you're not properly dealing with something there, and one of those things is: did you control for 2 ASGs in the 60s?

Without looking at the math, I agree that 17% - 76% seems like a big swing, but looking at the list of top All-Star game selections, it really is an extremely strong correlation with Hall-of-Fame status. Also, the BB-Ref link that I had in #73 shows # of All-Star selections both by season and by # of games (e.g., Hank Aaron was selected in 21 seasons but appeared in 25 games, which leads both categories).

Right, but by happenstance, and maybe this is just perception, but an abornmally higher-than-expected number of inner-circle HOFers were in their peaks during the 2 ASG phenomenon.

Oh, and one more edit to my #73. Mark McGwire had 12 All-Star selections, is HOF-eligible, and is not in the Hall of Fame. But I'm not sure how to count him for what are, I assume, obvious reasons (if I was doing something like Cyril's analysis, I'd be tempted to simply add a "Mark McGwire dummy variable" for him).

Freehan, with 11 selections, is the only player eligible for the Hall of Fame with 10 or more All-Star Game selections who is not in the Hall of Fame.

Freehan is in the Hall of Merit. First eligible 1982, elected 1985.

Of course that line for Larkin will read first eligible 2010, elected 2010.

If you look at the coefficient values on all-star games in the logit model, it is alot bigger than it is for MVP awards or 100 RBI seasons (I don't think it is directly comparable to the coefficient on average for example). And remember that it is the exponential function, so that is why things rise dramatically. I also have that linear model, too where we don't have this problem. I don't know which one is better.

retired players, Highest OPS+, 1500+ G at SS,

Unfortunately that's not how the BBWAA tends to define "SS". Your list excludes Yount and Banks who were both viewed as SS despite spending half their careers at other positions. Also Vaughan who just missed your cutoff but wasn't voted in by the writers. Also Vern Stephens. Your list would also exclude AROD and Nomar who didn't make the 1500 g cutoff. If you look at career OPS+ for players with 1000+ games at SS (including current), Larkin ranks 10th.

And who knew that Jim Fregosi would compare so favorably to Larkin. :-)

Cyril, there's nothing wrong per se with your logit model, it's just an odd sample (and model).

Note the odds ratio for career average -- it's 2.8 x 10^29. This is because you're measuring it as a decimal and a "1 point" change in BA is a "1000 point" increase in BA. Lord only knows what that's doing to your model. Measure it in 000s of points.

You also have 3000 hits. You say you removed Rose but, as far as I know, every player with 3000 hits is in the HoF -- that coefficient should "blow up." You can't include a perfectly predictive variable in a logit.

Your C indicator also has a suspiciously large effect on the odds -- but maybe it's OK (it probably just highlights how poorly the standard criteria work for Cs).

Looking at the data, the AS game thing is pretty clearly a threshhold effect -- if you made 10, you were in unless you were Garvey (this is the reason he has such a high probability. It's also highly correlated with your 3000 hits indicator.

The main "issue" though is the data. Primarily your model distinguishes between guys who were almost certain to make it and guys who never made the 5%. Look at your predicted probabilities -- almost everybody's either 95%+ or 12% or lower. You've only got 5 guys in the middle. And sure, in a sense, those are ideal results but all it's mainly done is distinguish the no-brainer inductions from the no-brainer non-inductions -- something any human being could do looking at the list.

And your model is seriously whack in some fashion if it only projects McGwire to have a 12% chance since you don't have a steroids indicator ... and that's 47% below Rice and 85% under Garvey.

I'm going to guess you ran a lot of models, taking variables in and out, etc. That leads to a data-dependent model (especially in small-sample scenarios) which is unlikely to replicate well. This is also likely the reason you get a largely different set of predictors in the %vote in the first year model.

That second model is more reasonable but has similar issues. It's dominated by the 3000 hits and 500 HR indicators which is stuff we already know. Technically speaking, you should transform the dependent variable since it is bounded 0/1 and you can get predicted vote percentages outside that bound with a linear model. The logit would be appropriate except you probably have a lot of zero votes. You could try the odds which only have a lower bound (and won't cause you trouble until somebody gets 100%). Also you shouldn't include squared terms or interactions in a regression unless you also include the lower-order terms -- or at least such is the standard advice.

It's not easy for a question like this. The population of interest is hard to define. There's no need to answer the question of why Gwynn made the HoF -- it was obvious. Similarly nobody's the least bit curious about what kept Al Bumbry out of the HoF. Over any 20-year period, there probably aren't more than 20 players who are in any real sense borderline.

An approach would be to run a 'tobit' model which combines probit (similar to logistic) and linear regression. You let the probit bit predict whether someone got at least 5% of the vote in their first year then the linear model would predict the vote percentage of the group who did pass the 5% threshhold.

But may main piece of advice is to stop over-fitting models. There are cases where such exploratory modeling is justified but this really isn't one of them. First, there's been lots of previous research so you should already have a good idea of what variables are important. Second, you just don't have enough data.

The main "issue" though is the data. Primarily your model distinguishes between guys who were almost certain to make it and guys who never made the 5%.

This right here is exactly what I was trying to say.

Walt

Thanks. You raise alot great points. JC Bradbury did a probit model and I have a link to his study and others at one of my posts.

How should I transform the dependent variable in the linear model? (anyway, as I recall, there were some guys who were predicted to have more than 100%, but not many and I don't think it was too much above 100% but I will have to check).

Not sure if my current package can do tobit, but I will look into it.

I think the biggest point is the issue with the things like 3000 hits, that they all get in. I just had not thought about that (and in the linear model, it has a value of .43, more than half way to getting in!). I think that pretty much wipes out even looking at the logit analysis. I could try just looking at all the guys who had less than 3000 hits, but then that is even less data to look at.

So on batting average, you're saying use 295 instead of .295?

Wish I could tell you what the correlation was between All-star games and 3000 hits, but I don't have that data here at home since I just had a hard drive crash and bought a new computer. It may be at the office.

You say that the two models had largely different predictors. They had some different and some the same and some similar.

On McGwire, I could put a dummy in for the steroid issue, but what if he is the only one (I may have actually done this but I can't recall since it has been a few months. I will check). Plus, there are other scandals. Someone suggested that Raines was hurt by his drug use. It could get hard to tell what to put in.

You are probably right that there is only some small number of borderline guys in any given time period. We could expand the time frame to get more data but then you get guys in lots of different run environments and you get the voters changing. One generation might have looked at 3000 hits or all-star games differently. Maybe some kind of time variable can be put in. The studies I have links to might have done that. I just don't recall.

On my squared terms in the linear model, they might be weirder than you think. I capped the all-star games at 10, for example. So up to 10, the power of AS games is increasing exponentially while and AS games beyond 10 have no further power (that was another one that resulted from trial and error). I will try it with the normal and squared terms, like you say, when I get a chance. But I think it might make sense to cap it. I can see the voters not caring too much past a certain point.

As for the earlier studies, I will look at those again to see what variables they used.

Thanks again. But again, the point that really seems most important is the one about so many guys being no brainers, either being in or out. Makes me wonder how much it is worth it to keep looking at it.

Cy

But again, the point that really seems most important is the one about so many guys being no brainers, either being in or out. Makes me wonder how much it is worth it to keep looking at it.

Cyril, I really respect the effort that you've made here, but I do think you may have nailed it here. Really, as Walt said, the interesting questions are guys who didn't reach one of the "automatic" milestones (3,000 hits, 500 HRs) but who have some kind of plausible HOF case (i.e., as Walt said, not Al Bumbry - although I loved the Bee as a kid, and if I had a HOF vote, I'd be tempted to toss him a token 1st-ballot shout-out). The problem here is how you define a guy with "a plausible HOF case". Ken Singleton, for example, gets occasional love in the Hall of Merit and has a reasonable case for having been a better hitter than Hall-of-Famer Jim Rice. But, on the other hand, Ken Singleton got no Hall-of-Fame votes his only year on the ballot and, honestly, I don't think anybody was the least bit surprised by that. Do you include him or not?

Really, once you throw out the "automatics" one way or the other, I suspect you're going to end up with maybe 30-50 players and at that point, you're going to ask your model to distinguish between, for example, these two guys:

Jim Rice 8 ASG (4 starts), 0 GG, 2 SS, 1 MVP, 9,058 PA, .298 BA, 382 HR
Dale Murphy 7 ASG (5 starts), 5 GG, 4 SS, 2 MVP, 9,040 PA, .265, 398 HR

I doubt you can do it.

Going back to the subject of this thread, is Barry Larkin - 12 ASG (5 starts), 3 GG, 9 SS, 1 MVP, 9,057 PA, .295 BA, 198 HR, 379 SB

more like Ryne Sandberg - 10 ASG (9 starts), 9 GG, 7 SS, 1 MVP, 9,282 PA, .285 BA, 282 HR, 344 SB

or Alan Trammell - 6 ASG (0 starts), 4 GG, 3 SS, 0 MVP, 9,375 PA, .285 BA, 185 HR, 236 SB

I doubt that you'll be able to find a mathematical model that's going to answer that question any better than the first 70-some comments of this thread did - which basically boils down to, "who the hell knows".

In modern baseball, how much weight should be given to Larkin being the "best shortstop in the league" in years where he was the best in the NL but behind as many as three (or maybe more) guys in the AL?

Even when Larkin won the MVP in 1995, you could make a not terrible argument he was the 2nd best shortstop in baseball that year. And that was even before the ARod, Garciaparra and Jeter eras started (and after Ripken and Trammell had started to slide considerably).

I'm not arguing against Larkin as a candidate necessarily, but I think this particular argument harkens back to the "Ozzie Smith was the best shortstop in the league in the 80s" argument: well sure, but then Ripken, Trammell and Yount were in the other league.

I think Larkin should be judged against his peers (a judgment I think is very favorable to him) and to me his peers include guys from the AL.

Even when Larkin won the MVP in 1995, you could make a not terrible argument he was the 2nd best shortstop in baseball that year. And that was even before the ARod, Garciaparra and Jeter eras started (and after Ripken and Trammell had started to slide considerably).

I'm not arguing against Larkin as a candidate necessarily, but I think this particular argument harkens back to the "Ozzie Smith was the best shortstop in the league in the 80s" argument: well sure, but then Ripken, Trammell and Yount were in the other league.

I think Larkin should be judged against his peers (a judgment I think is very favorable to him) and to me his peers include guys from the AL.

One thing to keep in mind is that most of this thread has been discussing whether we think the BBWAA will elect him. I don't think anybody here thinks that Barry Larkin DOESN'T belong in the Hall of Fame. The "best shortstop in the NL since Ozzie Smith" is being framed by most people here as a "BBWAA-style" argument - hence, for example, the discussion of All-Star game selections and Gold Gloves and MVP voting.

And, as you say, even if you add Ripken (better), A-Rod (better), Trammell (worse), et al. into the mix, Larkin still looks very, very good. Debating whether Larkin is better or worse than Ryne Sandberg or Derek Jeter are probably interesting debates that reasonable people can take either side on (without studying it too much, I think I'd say he's better than both). But those debates would take place well inside the Hall-of-Fame in/out line.

Even when Larkin won the MVP in 1995, you could make a not terrible argument he was the 2nd best shortstop in baseball that year.

I assume you're talking about John Valentin. If you somehow decided you're not going to give the award to Randy Johnson, then a not terrible argument can be made that Valentin deserved the AL MVP. Getting to that level - plausible MVP - takes some of the sting out of that comment. Valentin didn't stay at that level long enough to be a plausible HoF candidate, but he was all of that for that particular year.

Debating whether Larkin is better or worse than Ryne Sandberg or Derek Jeter are probably interesting debates that reasonable people can take either side on (without studying it too much, I think I'd say he's better than both). But those debates would take place well inside the Hall-of-Fame in/out line.

Yep. For sure. Or at the very least, well inside the Hall of Merit in/out line.

(Oh, and the true 1995 NL MVP was obviously Greg Maddux.)

Cyril,

as 86 points out, I'm not sure there's a good way to do this. I wasn't really seriously suggesting a tobit -- they're kinda odd models, kinda hard to interpret and really something you should only undertake after some experience. Modelling vote totals, as you have, might have some utility but #86 points out the issues there as well. If you go that route, you could transform your dependent variable as the odds (p/(1-p)) or, if you have no 0s or 1s, the logit (log (p/(1-p))).

And yes, I was saying put in BA as 295 not .295 -- I'm not sure it will "fix" things but it will look better. :-)

My point on McGwire wasn't to suggest adding a steroid indicator. It was that a model that predicts McGwire as having only a 12% chance of making it in is doing something wrong -- especially if it's got Garvey at 96%. We all know the reason McGwire wasn't in on the first try or, at absolute worst, got 50% + and would be in by now was steroids. Since you (rightly) didn't include a steroids indicator, there's something odd going on there.

I'm not sure I'd rank Larkin ahead of Trammell, though I suppose that depends on what you thought of Trammell's defense. Trammel is one of those players that has a really clean peak from 25 to 30 where he was a dynamite baseball player. His misfortune was to have that overlap with Ripken's best overall years as well. He also was the main victim of one of the most absurd MVP decisions I've ever seen in 1987 (though Boggs was equally as screwed there).

So was his defense equal to his rep? If so Trammell is a clear Hall of Famer just like Larkin is. It is neat that each comes out as each other's most similar player on BBRef. I think the lack of support for him in the HOF voting is troubling.

Getting to that level - plausible MVP - takes some of the sting out of that comment.

The point really isn't to dock Larkin, the point is just to point out that separating out "NL Shortstops" from "MLB Shortstops" doesn't make a lot of sense in terms of evaluating him. It hurt Alan Trammell a lot more in comparison to say Ozzie Smith, but it's worth mentioning in Larkin's case since for most of his career the AL Shortstops were generally a stronger group (Trammell and Ripken early, the trio late with Valentin having some decent years in between).

Kiko

I think Singleton had about 310 career Win Shares and I think that is the highest for anyone who got zero votes in my study. I sure don't know if you include him or not in the list of possibles. I think Will Clark is in a similat boat. I think he got less than 5% his first time yet had over 300 WS.

Walt

When you say "you could transform your dependent variable as the odds (p/(1-p)) or, if you have no 0s or 1s, the logit (log (p/(1-p)))" is that p the percentage of the vote they got or a probability and if a probability, how is that calculated?

Cy

p would be the proportion of votes (percentage divided by 100). But it's not likely to make much of a difference what your dependent variable looks like and then it becomes difficult to interpret the results. Those were just minor comments.

Thanks.