Baseball for the Thinking Fan

Login | Register | Feedback

btf_logo
You are here > Home > Hall of Merit > Discussion
Hall of Merit
— A Look at Baseball's All-Time Best

Sunday, May 11, 2008

Ranking the Hall of Merit First Basemen - Discussion

These are the Hall of Merit first basemen to be voted on (in alphabetical order):

Dick Allen
Cap Anson
Ernie Banks (just kidding, but if Deacon White allegedly is a third baseman… :-)
Jake Beckley
Dan Brouthers
Will Clark
Roger Connor
Jimmie Foxx
Lou Gehrig
Hank Greenberg
Keith Hernandez
Harmon Killebrew
Buck Leonard
Willie McCovey
Mark McGwire
Johnny Mize
Eddie Murray
George Sisler
Joe Start
Mule Suttles
Bill Terry

The electon starts May 18 and ends Sunday on June 1 at 8 PM EDT.

John (You Can Call Me Grandma) Murphy Posted: May 11, 2008 at 10:20 PM | 160 comment(s) Login to Bookmark
  Related News:

Reader Comments and Retorts

Go to end of page

Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.

Page 2 of 2 pages  < 1 2
   101. TomH Posted: May 20, 2008 at 06:56 PM (#2787897)
for those who haven't seen the history of the Dick Allen debate, or who look in shock at my ranking, I offer this:

A player's job is to help his team win. 99% of the time, his individual batting, fielding, and running stats correlate well with how much he helped his team win.

Dick Allen IMHO is the other 1%.

While we should be wary of taking press reports and drawing broad conclusions, especially when the press may have had biases (particularly racial ones!), the data correlate well with the reports that Allen contributed to bad clubhouse karma or what have you.

We should also be wary of correlating team wins with player movement helter-skelter, since baseball is a team game, and teams have much else going on than one player. But when teams consistently don't improve despite getting what seems to be a great, great player for often very little in return... I must conclude Dick Allen's value was less than his numbers suggest.

Here's a quick summary; please check it out yourself to see if I've mangled anything.

1964: Allen arrievs in Philly, hits great, replacing a very poor hitting Don Hoak. Phils improve a little (4 games), mostly due to better record in close games, but they flop in September, and Allen hit very poorly against the league winning Cards.

1965-69: Allen has great numbers. Phils get slowly worse.

1970: Traded to StL in big package. Allen has fine year. Phillies improve by 10.5 games in his absence. Cards decline by 11 games with Allen.

1971: Traded to LA for Ted Sizemore. Cards improve by 14 games [it must have been Ted Sizemore! :) ] Dodgers improve by 1.5 games with Allen's fine year.

1972(-74): traded to Chi Sox for ... not much. Dodgers do same without him. Sox do improve from 79 wins to 87, then win 77 and 80 games in subsequent years with Allen, who still is one of the best hitters in the game.

After this, he becomes a part-time player.

I feel bad for Allen; he got a raw deal in many ways. But I think his claim to Hall of Fame (Merit) status is tenuous.
   102. Chris Cobb Posted: May 20, 2008 at 07:22 PM (#2787965)
Tom H,

There are lots of examples of great players on good teams. We know why teams didn't want Dick Allen around by the late 1960s, but the history you chart is meaningless as an assessment of Allen's impact on his teams. In 1969, St. Louis had a 121 ERA+. In 1970, they had a 102 ERA+. Do you blame Dick Allen for the St. Louis pitchers having a bad year?
   103. Charter Member of the Jesus Melendez Fanclub Posted: May 20, 2008 at 07:25 PM (#2787975)
There probably is a good argument that Allen hurt his teams, but I don't think simply reciting the improvement and decline in team records is evidence. Is Alex Rodriguez also hurting his teams off the field? (Well, maybe some would argue so, but not on Allen's level.) I think it would be more useful to review the actual controversies he was involved in. I suppose that information is in his thread.
   104. John (You Can Call Me Grandma) Murphy Posted: May 20, 2008 at 07:39 PM (#2788010)
There probably is a good argument that Allen hurt his teams,


The times he was out surely hurt his team, but that's recorded in his stats. The same goes for poor performances due to pouting.

With that said, I seriously doubt he aggravated his teammates to the point that it affected their performances. If it did (based on what we know Allen actually did during that time), it would reflect more poorly on those players, IMO. The guy wasn't a bully or pulling scams. He was moody and sulky. I think we could all handle that type of silliness as a teammate.
   105. TomH Posted: May 20, 2008 at 08:03 PM (#2788068)
Do you blame Dick Allen for the St. Louis pitchers having a bad year?
--
No, but when you trade Curt Flood (and replace him in CF with Jose Cardenal) to get Dick Allen, you gain a big bat and lose a lot of D!

I don't mean to say that Dick Allen cost his teams 5 (2, 6, pick-a-#) games a year with his attitude. But if a guy was said to be clubhouse trouble, what evidence IS there to substantiate or refute it? If his teams won, wouldn't lots of people be saying that the evidence controverted the claim?? Team performance most certanly IS evidence. Enough to convict? Hardly. Could be circumstantial. But the prosecution rests, awaiting the defense to come up with more than "naaaaahhh".

I also realize I take a different tack than most, and evaluate players as a GM would; if I sign this guy, will he stay around long enough to help my team win for many years? YMMV, but if I guess I have to trade a guy for a box of doughnuts in a few years, that's not gonna make me too warm and fuzzy about obtaining him.

Again, I'm not trying to persuade. Just explaining my own perspective.
   106. John (You Can Call Me Grandma) Murphy Posted: May 20, 2008 at 08:29 PM (#2788110)
Team performance most certanly IS evidence.


Only if, by elimating all other possible factors for the poorer team performance, all indicators point to Allen as the culprit. I just don't see it here.
   107. Dizzypaco Posted: May 20, 2008 at 08:48 PM (#2788145)
With that said, I seriously doubt he aggravated his teammates to the point that it affected their performances. If it did (based on what we know Allen actually did during that time), it would reflect more poorly on those players, IMO. The guy wasn't a bully or pulling scams.

I agree with you if you're talking about Dick Allen in the 1960's, or up to 1972. I disagree with you about Allen near the end of his career. He quit on his team in the middle of a pennant race. From what I have read, he really did cause problems in the clubhouse in his last few years. I would not want him on my team at any price after 1972, and therefore I credit him with zero value.

Now, would I vote for him based on what he did from 1964 to 1972? Quite possibly. But if there was ever someone who was less than his stats suggest, I think it was Dick Allen.

Bill James wrote back in the 1980's about team leaders. They are going to be your best players, whether you like it or not. If your best players hustle, are always trying to improve themselves, and are willing to be unselfish on behalf of the team, than the other guys are sure to follow. But if your best players are moody and selfish, refuse to hustle, refuse to practice, and refuse to make the most out of their talents, its going to cause problems that you just don't want.

There has always been this assumption around here that all of these GMs who couldn't wait to get rid of the guy were idiots, that they just didn't know how good or valuable he was. I disagree - I think they knew he was really good, and they just didn't want him around anyway for some pretty good reasons. And the fact that their teams did improve after he left suggests they weren't so dumb after all, even if win-loss record isn't definitive proof.
   108. Blackadder Posted: May 20, 2008 at 11:05 PM (#2788293)
Dan, that is interesting. Two questions:

1) Have you done any studies on whether FRAA/FWS are as poor for other positions?

2) Do the fielding numbers on your WARP spreadsheet incorporate the retrosheet-style PBP metrics for 1956-1986, or is it FRAA/FWS all the way up until Zone Rating?

Thanks.
   109. John (You Can Call Me Grandma) Murphy Posted: May 20, 2008 at 11:06 PM (#2788296)
I would not want him on my team at any price after 1972, and therefore I credit him with zero value.


What did he do to make you think he was worthless for those years?

There has always been this assumption around here that all of these GMs who couldn't wait to get rid of the guy were idiots, that they just didn't know how good or valuable he was.


Getting rid of a cancer like Denny McClain made total sense. That's a guy who could really disrupt a team. However, Allen wasn't McClain.

And the fact that their teams did improve after he left suggests they weren't so dumb after all, even if win-loss record isn't definitive proof.


The only thing that it suggests is that they improved. It does not necessarily suggest that it was because of Allen, just as it doesn't say anything bad about A-Rod before and after with the Mariners, either (the latter drives me up the wall even more so, BTW). One has to do some detailed analysis first before you can say a player is to be blamed or not.

Bill James is my favorite baseball writer and analyst, but I frankly wish he had put a little bit more thought in his sketches of Allen and Hornsby.
   110. Chris Cobb Posted: May 20, 2008 at 11:50 PM (#2788380)
No, but when you trade Curt Flood (and replace him in CF with Jose Cardenal) to get Dick Allen, you gain a big bat and lose a lot of D!

Yes, but that's hardly Allen's fault. That's genuine GM responsibility, there, if the loss of Flood was indeed a significant factor in sinking team ERA+.

Let’s look more closely at what happened to the 1970 Cardinals, using BP’s WARP1 as a guide to changing team quality.

BP sees the St. Louis pitching staff declining slightly, from 61 PRAA to 46 PRAA from 1969 to 1970, while it sees the team's defense as absolutely tanking, going from 47 FRAA to -39 FRAA, a change of 86 FRAA, worth about 9.5 wins. As BP's WARP sees it, St. Louis downgraded at almost every position:

C McCarver -> Simmons/Torre -2 FRAA
FB Torre -> Allen/Hague -10 FRAA
2B Javier -> Javier -> +4 FRAA
3B Shannon -> Shannon/Torre -16 FRAA
SS Maxvill -> Maxvill -16 FRAA
LF Brock -> Brock -12 FRAA
CF Flood -> Cardenal -29 FRAA
RF Pinson-> Lee/Hague -3 FRAA
P Pitchers -> Pitchers -3 FRAA

So by getting rid of McCarver and Flood in favor of Allen and Cardenal, the Cardinals lost two strong defenders up the middle, shifted Torre from the one position where he was a defensive asset to two where he was a defensive liability, and replaced those strong defenders with weak defenders. These changes alone cost the team 55 FRAA, which is about 6 wins! With Brock going from ok to poor defensively and Maxvill going from great to good defensively also, the defense overall fell off a cliff.

6 wins would be a lot for Allen and Cardenal to add with the bat over their predecessors. Let’s see what happened offensively

C+1B+3B+CF (the positions affected by the change of personnel) created

82+98+72+89 = 341 EQR in 1969, which equals about 77 BRAR or 8.5 wins
90+108+91+82 = 371 EQR in 1970, which equals about 86 BRAR or 9.5 wins

Note that offensive production did improve at the positions in which Allen was involved: these added 37 EQR, or a bit over 1 win. Cardenal loses a couple of BRAR vs. Flood, so overall he was about a 3.5 win downgrade from Flood.

Adding Allen himself, then, cost the team about 2 wins: 3 wins on defense with 1 win gained back on offense. That win was washed out, of course, by Javier and Maxvill and Shannon being complete out machines in 1970, and by the decline in pitching.

In retrospect, the Cardinals had no business trading for Allen. The sensible thing to do would have been to hand McCarver’s job to Simmons, leave Torre where he was at first base, and find someone who could play third. Shannon was in decline in 1969, but St. Louis may not have counted on him being totally washed up. Either that, or they thought Allen could handle third base, even though he had been a poor third baseman in Philly for years, and the Phillies had moved Allen completely off of third in 1969. Somewhere here there were several significant errors in judgment by the management, mostly involving the defensive value of players.

One can conclude, then, that adding Allen did not make the Cardinals better, but that was mostly because he didn’t actually fill any of the team’s gaps. He could have significantly improved a team in need of a good-hitting first baseman, but that team was not the Cardinals! The change from Flood to Cardenal in CF was much more damaging to the team, however, and the team also had bad luck, underperforming Pythagorean expectations by several wins (Note that this was not a pattern for Allen’s teams! Their overperformance/underperformance was pretty well mixed and washes out to close to zero over Allen’s career).

It’s worth adding that Allen’s 1970 was one of the worst years of his career. He missed a lot of time (injuries or sulking?), his offensive performance was the worst of his prime, and his defensive performance was, too. If Allen had performed as well in 1970 as he did in 1969, when he played about as many games but hit and fielded better, the Cardinals would have about broken even with the addition of Allen. If they had gotten the Allen of 1968 or 1971, he would have added about 2.5 wins to the team. That gain still would not have offset the downgrade from Flood to Cardenal, so if these two changes were indeed interrelated, the Cardinals would not have strengthened the team unless they had gotten the Allen of 1964-67. Given the clear attitude issues with Allen, that level of performance could hardly have been expected from him again (though the White Sox got it in 1972).

Overall, Allen played poorly, for him, and cost the Cardinals, but even if he had played up to expectations, the team still would have declined as a result of its changes in personnel. The team was probably right to make moves to replace Flood and McCarver, both of whom the team had good reason to believe to be in decline, but, aside from bringing up Simmons, their moves to replace these two were not effective, and they didn’t address the declines of Shannon, Maxvill, or Javier, which were severe.

A significant cause of the Cardinals’ rebound in 1971 was not Allen’s departure, but the departures from the lineup of Javier, Shannon, and Cardenal (Maxvill would hang on until 1972).

In sum, the case of Allen and the Cardinals should at least remind us that baseball is a team sport. When teams change half their starting lineup each year for two seasons, it’s not accurate to attribute the team’s changes in fortune to adding and then subtracting one player.

As I read it, the Cardinals in 1970 were an old team. They had been a great team, but now they were in decline. They thought with Allen that they could hold off on rebuilding, except where they knew they had a good piece to add (Simmons). They were wrong, or, rather, they didn’t make the moves they would have needed to make to compete with their aging core for another year.
   111. David Concepcion de la Desviacion Estandar (Dan R) Posted: May 21, 2008 at 06:37 AM (#2789230)
Blackadder:

1. No, definitely not. I am going position by position and will post my results as I get them--it's a bit time consuming and I've been quite busy of late. There's also the $64,000 question of how to incorporate non-range/hands effects: double plays for second basemen, scooping for first basemen, and throwing for outfielders.

2. It's definitely FRAA/FWS all the way up until Zone Rating--Simple Fielding Runs and TotalZone didn't even exist when I first did my WARP! I may calculate a WARP 3.0 incorporating these new metrics--not just empirically accurate regression equations for the range/hands component of defense, but also DP runs for 2B, way improved catcher D stats, and significantly better baserunning stats (distinguishing between stealing 2B and 3B, etc.).
   112. bjhanke Posted: May 21, 2008 at 09:26 AM (#2789244)
This is the big Sisler post I promised. I tried to put it into the Sisler/Konetchy thread, but I can't find it, for some reason. Since the post is long, if the moderators want to move it to the Sisler thread, I have no objection. I just promised to post it up today, so here it is. Sorry, Brock Hanke. Wow. I just had to break the post into two sections, because it's over the post length limit. Please move it to the Sisler thread, someone. Thanks.

INTRO

This is a large post, because 1) I have thought a lot about George Sisler over the years, and 2) I had access to two people who were heavily into baseball when Sisler was at his peak: my father and the sportswriter Bob Broeg. I didn't pay any attention to how good they thought Sisler was, because they both were irrational on the subject. My dad was an 11-year-old boy in 1922, when the Browns nearly won the pennant. Bob Broeg was about to start writing the Cardinal beat for the St. Louis daily paper at that time. But they were able to give me info about the details and shape of Sisler's play that I would not otherwise know. The main reason for my writing such a huge post is that info. All I can hope is that it helps and that you don't think you wasted your time reading this.

ON SISLER'S DEFENSE

I'm going to start with a Bill James quote, from his latest Historical Baseball Abstract. "Also, while articles INVARIABLY (emphasis mine) claim that he (meaning Sisler) was a fine fielder, the Win Shares system is unable to see that this is true." Now, "invariably" is a pretty strong word, and Bill doesn't throw such around much. The strength of this quote, therefore, allows me to quote my father and Bob Broeg with more freedom than I would otherwise take, because of their bias. I was worried that Sisler's defensive rep was a local St. Louis thing until I read that sentence. Now, I feel much more free to repeat what I've heard in Sisler's city.

Anyway, it was very odd. I asked both men what Sisler's greatest strengths were, including both offense and defense. There was no point is asking them how good he was, because I already knew they both thought he was a first base god, so I asked them to break it down instead, to get a handle on what contemporary observers thought they were seeing. Both men started the breakdown with almost the same quote, which was "Sisler was a Gold Glove shortstop who happened to be left handed. His specialty was pouncing on bunts hit to the first base side, to the point that no one would bunt in his direction." There are two important things here. First, neither man started with offense. Not the two .400 seasons, not the 19 homer campaign. No, they both started with defense at first base. The second thing is that they both singled out bunt defense as Sisler's particular greatest strength.

Trying to put this stuff together, I looked at Total Baseball's Fielding Runs for Sisler. They are high; Pete Palmer's methods have Sisler as a fine defensive first baseman. TB also does NOT rate Hal Chase highly on defense. Chase is the other first baseman that Bill's DWS doesn't rate highly, but who had a great defensive rep. TB struck me as more likely than the DWS result. Chase's defensive stats SHOULD be much lower than his rep. The man was in the business of selling ballgames, and there are limits. If you start selling the games by not trying to get hits, you end up out of the league for lack of offense. But if you are very flashy on defense, and Chase doubtless was, you can sell the games by making defensive "mistakes," and the stats of the 1910s aren't going to catch you doing it. Bill mentions this in his Hal Chase comment, "A favorite trick was to break late in covering first, thus failing to make a catch while seeming to make a good effort...."

Normally, I trust Defensive Win Shares much more than Fielding Runs, but in this case, I realized that there might be a variant of the Johnny Bench problem that I mentioned in the HoM catcher discussion. Bill's system works largely by looking at the context that the fielder operates within. That is, he uses the synergy between the various infielders' stats as part of his examination. So what would happen to all those stats if a first baseman was so good that no one would bunt his way in a league that bunted a lot? Well, assists for third basemen would go up, because in those situations where a team is going to bunt no matter what, they are always going to bunt towards third. Putouts for second basemen would also increase a little as they covered first to take the third baseman's throws. First base assists and putouts would both be down, of course (no tag plays on bunter/runners who bunted too hard to first). What's that going to do to Win Shares? Make the first baseman look weaker than he really was. His stats are going to be too low, and the context is going to be too high. So, it's very possible that DWS ranks Hal Chase low because Chase really did have weak defensive stats as a result of selling games, and Sisler low because the system cannot adjust for which direction hitters bunt toward.

I tried to check this out. I wondered if Brownie third basemen had a drop in assists in 1923, when Sisler missed the season, and then had them go back up in 1924, when he returned, and before the league figured out that he couldn't do what he used to be able to do. Unfortunately, the Browns' third base position is a mess at this time. ("Joe Dugan, the Damn Yankees, and 1922" is a curse word among old Brownie fans.) They are trying everything. Platooning. Trying multiple third basemen every year in an effort to find a good one. They did, eventually, but after the three-year window that I needed to look at.

So, overall, I have to go with the rep and with Total Baseball. I have to suggest here that people ignore DWS, treat Sisler as a tremendous defensive player before the injury, and adjust rankings according to how much effect you think suppressing bunts in one direction might have. Then, of course, you have to drop the ranking down after 1923, so that's a lot of factors involved in evaluating one first baseman's defense. Overall, in case anyone wants to know, I give Sisler an A+ through 1922, a B- from 1924 on, and an A- overall. If you generally operate with Win Shares, I suggest adding one whole Win Share (we're talking about a medium-level effect at best) per year through 1922, and nothing after. It's the best I can do. I just think that there is no reasonable doubt that George Sisler was a real defensive phenom before his eyes went bad.

And I think it is worth remembering that at least two serious fans of Sisler considered defense, and bunt defense in particular, to be his PRIMARY strength; a greater one than anything involving offense.
   113. bjhanke Posted: May 21, 2008 at 09:26 AM (#2789245)
ON THE BALLPARK

As you no doubt know, Bill James considers Sportsman's Park in St. Louis, where Sisler played all his St. Louis home games, to be a ridiculous bandbox, and there is no doubt, no matter what method you use, that it was a great general hitters' park, perhaps the best in the league at the time. However, as with Fenway in Boston, the park is not symmetrical, and therefore, its park effects are not the same on righty and lefty hitters.

Hobbyhorse alert: It is my opinion that the single most useful thing that sabermetrics could do and has not done would be to compute park effects separately for righties and lefties, both hitters and pitchers. I'm serious. I think this is project #1. I can't do it, because I lack the database needed. But the database should be out there, and I really do wish that one of the organizations that has it would compute those historical ballpark effects splits, both for handedness and also for individual stats. All I can find now is splits for homers. I want them for everything, and I want that lefty/righty split.

OK, enough of me riding my hobbyhorse. Regarding Sportsman's Park and Bill James and George Sisler: Bill actually got himself in some trouble over this issue, when, in his first Historical Abstract, he claimed that Rogers Hornsby's home run totals should be discounted for the ballpark effects. As it happened, one of his readers had gone through the box scores and recorded Hornsby's homer splits, and found out that they were neutral. Bill, whose publishing ethics are one of his strong points IMO, apparently spent money to purchase the reader's data and published it in his next Historical, along with a retraction of that particular statement.

I believe, based on that and on observation (I probably saw a couple of hundred ballgames in that park before it closed), that the park was in fact neutral for righty hitters. That means, of course, that the entire weight of the ballpark's effects falls on the bats of the lefty hitters there. Like George Sisler. And Stan Musial. Not to mention Ken Williams or other lesser players.

It kills me to say this, but that means that you should increase the ballpark adjustment against Sisler (and Musial, et al), when evaluating them. And you should ignore it when evaluating Hornsby and other righties (Ken Boyer, et al).

Except. When I mentioned this to Bob Broeg, he pointed out that the ballpark's effects were caused almost entirely by a small right field territory, due to the dimensions of the city block that contains the ballpark. That field is so small that the team eventually had to put a big fence, similar in effect to the Green Monster, on top of the right field wall, in front of the right field stands. Broeg admitted that Sisler played before the screen went up in the thirties, but he also pointed out that the effect of the park - the effect countered by the screen - was to increase home runs. Which George Sisler didn't hit anyway.

Broeg wasn't stupid; he had a point. If the main effect of the ballpark was to increase home runs over the right field fence, well, that kills Ken Williams, but it doesn't have much effect on George Sisler. In fact, it may have actually hurt Sisler, in that right fielders could play shallow in there, because there wasn't much room to hit balls over their heads that would stay in the park. That is, they could have played shallow and cut off some of Sisler's doubles power and maybe even some of his line drive and pop fly singles. I don't know. As I said, it's my hobbyhorse.

Well. Talk about your cheap instant karma. I wrote the above about 20 minutes ago, and have been watching the Cards play the Astros down in Houston. Troy Glaus has just hit a line drive down the left field line, a double is almost any park. But because Houston's left field is small (which results, statistically, in a large hitter-friendly adjustment, particularly for righty power hitters), Houston's left fielder was able to hold the Cardinal righty power hitter to a single. The announcer's comment? "Well, that happens a lot here in Houston." I don't know about "a lot," but this is exactly what I'm talking about with Sisler back in Sportsman's.

In any case, I think you have to be careful in discounting George Sisler's stats for the ballpark, except when considering home runs. You have one factor (the lefty/righty thing) pulling one way and another (the feature of the park doesn't affect Sisler's type of hitting as much as power hitting) pulling the other way. Until someone gets my hobbyhorse built, or at least some splits for Sisler, I don't know what to tell anyone, except to be careful about this with Sisler. The ballpark effects that are printed now may not turn out to be accurate for George. Adjusting his stats down may not be justified.

DOMINANCE of a PERIOD OF TIME

This is an odd adjustment that I've made for years, but I don't know if anyone else makes. As you all know, Sisler's reputation was very very high when the Hall of Fame opened. He was elected early and easily. Why? Well, you can get the clue by looking at the Historical Abstract. Bill has Sisler ranked about as low as anyone, at 24th among first basemen. But check who is above him. There are, literally, only three first basemen ranked ahead of Sisler who played earlier than he did: Connor, Broughers, and Anson, all from the 19th century. That is, when the Hall of Fame made the statement, essentially, that George Sisler was the best first baseman of the first 25 years of the 20th century, they were just acknowledging fact. I tend to give a player credit for that - for being the best at his position for a long period of time. Home Run Baker: there is some chance that when we get to third basemen, I will actually have him in first place. He's the dominant third baseman in 20th century baseball - probably in baseball going back into the 19th c. - until at least Eddie Mathews. I tend to give credit for that. I mean, every team as to play a third baseman. If you have the best one over a 75-year period, you've got something. Sisler isn't that strong, but he is the dominant first baseman for 25 years. I don't know if you give credit for that, but I do. And, according to Bob Broeg, so did the Hall of Fame. That's part of why he was elected so early. Makes sense to me.

As a side note, Sisler is probably the real answer to the question of who is the best dead ball era first baseman. True, he only played about 4 years in the dbe, but his skills, throughout his career, are the skills of dead ball play. And he was the best at those skills. Again, I don't know if that means anything to anyone else, but it does help me get a handle on what kind of player I'm dealing with.

NOTES

Here are a couple of small things.

First, if you look up Sisler, you will find that stolen base stats are erratic at best during his career. But when they are available, Sisler is outstanding, both in volume of attempts and in success rate. He's actually in a class with Eddie Collins in the same years. Stolen bases don't count for a whole lot, but it is a small plus.

Second, if you're looking at Total Baseball, the career ranking for Sisler is wrong, for the reason that everyone complains about with Pete Palmer's methods: assuming that the average player has zero value. The back end of Sisler's career has a bunch of small negative numbers, since TB thinks that Sisler, then, was just a little below league average. For his career, they give him 24.7 Total Player Ranking. But if you count his TPR up through 1922, you get 30.5 TPR. That is, TB says that Sisler's teams would have been better off if he had just quit the game when he got hurt. But what does that assume? It assumes that Sisler could have been easily replaced by an average player. That's not true, of course. I don't know exactly what the replacement player is, but it isn't league average. I don't know what everyone will do with that info, and I don't make much adjustment for it myself, but it is there.

IN HIS CONTEXT

So, where does George Sisler rank among his contemporaries, meaning all players, not just first basemen? Surprisingly to me, the standard sources, Win Shares and Total Baseball, don't agree, especially on where to rank his peak. On the back end of his career, pretty much everyone agrees that he was a decent, but mediocre starter. But the peak ranking varies.

Total Baseball uses the title Total Player Ranking for their leader board of position players (not pitchers), which they mercifully print up for you. Win Shares does not print annual leader boards, but I think I managed to compile the numbers without more than one or two mistakes. Anyway, here are Sisler's rankings, among position players only in the American League only, starting in 1916, which is his first full season, and running through 1922. A zero means he's not on the leader board, which lists the top 5 players. The columns are Year, TB (Total Baseball), and WS (Win Shares):

Year - TB - WS
1916 - 0- 0
1917 - 5 - 0
1918 - 2 - 0
1919 - 4 - 0
1920 - 2 - 4
1921 - 4 - 3 tied
1922 - 1 - 3 tied

The TB entries are genuinely impressive. They put Sisler, for 6 years, in the company of Ty Cobb and Babe Ruth and Tris Speaker and Eddie Collins: the megastars. Sisler ranks right among them, not at the bottom of the list or anything like that. Win Shares doesn't see it. They have him a rung below the megastars. Perhaps among the superstars, but clearly a rung below what TB thinks.

So who do you believe, and why? That's the question for where you put him on your ballot. If you think that he was, at his peak, the equal of Cobb, Speaker and that group, you're going to rank him higher than if you think he's in the next tier down. My personal decision is to agree mostly with Total Baseball. My main reason is that I don't think that Win Shares has his peak defense right. My secondary reason is that his reputation, in his own time, squares more with TB than with WS. In the absence of any evidence that is precise enough to distinguish between superstar and megastar, I'm going with the rep. If there were conclusive evidence, yes. But I don't think that's the case here.

Please note that I do NOT intend to rank his CAREER up there with the megastars. But his career is several years of ordinary play, plus the peak. The peak is the only thing he really has to offer the level of the HoM ballot, so it makes a lot of difference where you rank that peak. All I can hope is that I've given enough info here to help some of the rest of you make up your own minds.

Thanks for reading!

- Brock Hanke
   114. OCF Posted: May 21, 2008 at 12:30 PM (#2789267)
I have a couple of responses to this, especially the second post.

We've talked some about differential effects of ballparks on different kinds of hitters. We're not all of one mind about this, but here's the position I have staked out for myself: it doesn't matter. If you want to argue that Gavy Cravath took unusual advantage of the Baker Bowl, or Mel Ott took unusual advantage of the Polo Grounds, or Joe DiMaggio was unusually hurt by Yankee Stadium, then you're arguing a "what if." What would this player have produced hitting in a different park? And I don't want to go there. I don't want to talk about the hypothetical value that a player would have had in a different circumstance; I'd rather talk about the value a player did have in the circumstances he played in. A run in the Baker Bowl, in the Polo Grounds, in Sportsman's Park, in Yankee Stadium, has a certain win value. That's what simple, same-for-everyone park factors tell you - coupled with offensive levels, they tell you the win value of a certain amount of run production. Jim Rice's ability to pull balls over or off the Green Monster resulted in real wins for his team - that counts. (It's not enough to get him a sniff of HoM consideration, but that's a different story.) That Jimmy Wynn seems somehow not to have been hurt by the Astrodome like everyone else was - that counts, too, and resulted in real wins for the Astros.

As for Sisler's "unique" dominance over the first quarter of the 20th century - I'm not so sure. I'll claim that, offensively, the best four or five years of Sisler's career do not top the best four or five years of Frank Chance's career. Chance is a little different as a hitter - lower BA but more walks, playing in (on average) lower-run circumstances than Sisler. I do put Sisler ahead of Chance overall, but that's mainly a question of career value. Sisler did have that long coda to his career, and although it wasn't worth much, I agree with you that it's better to count it as a net positive rather than a net negative. Chance had (counted in games) a very short career, and he had a significant problem with in-season durability. But when he was in the lineup, Chance was as good an offensive player as Sisler, and while I don't have that clear a sense of Chance's defense, it's clear that he wasn't a defensive liability.

Defense mattered, of course. But in the sportswriting of the time, the glorification of defense approached the level of being an established religion. Chase, who had the reputation, was praised far beyond the level warranted, while a masher like Jack Fournier had trouble even sticking in the majors despite abundant evidence that he could hit. I grant that Sisler was the best first baseman between ABC and Gehrig, but the "what ifs" include both "what if Chance could stay healthy" and "what if they'd let Fournier play."
   115. DL from MN Posted: May 21, 2008 at 01:42 PM (#2789301)
> George Sisler was the best first baseman of the first 25 years of the 20th century

If you only consider white first basemen. I prefer Ben Taylor to George Sisler. This is the same bias that leads people to overrate Pie Traynor when the best contemporary 3B were in the other league.
   116. DL from MN Posted: May 21, 2008 at 05:54 PM (#2789553)
> I may calculate a WARP 3.0

Hopefully the pitchers are in this one, also 2006-2008 stats.
   117. David Concepcion de la Desviacion Estandar (Dan R) Posted: May 21, 2008 at 10:15 PM (#2789808)
OK, using my new regression equations, I've calculated revised 1B fielding stats for the post-1956 mid-ballot guys. The results are somewhat surprising. Unsurprisingly, the standard deviation of 1B fielding is meaningfully smaller here than it was in the previous version of my WARP. This is because the last version forced the 1B defensive stats to have the same standard deviation as Chris Dial's Zone Rating-based data, despite the fact that they were less reliable. The standard deviation of these new numbers reflects *precisely* the level of confidence we can have in them--68% (so with an effective 32% regression to the mean) for 1987-2002, 54% (an effective 46% regression to the mean) for 1956-1986, and 14% (a whopping 86% effective regression to the mean) for 1893-1955.

But the relative ranking is also notable. In particular, Keith Hernandez appears merely human. DRA is nuts about him (+191) and TotalZone is appropriately bullish (+112). But SFR is *extremely* unimpressed, crediting him with just 18 runs above average excluding his 1983 and 84 (which are missing from its data set), and SFR is the stat that has *by far* the strongest correlation with the 2003-07 PBP average (over 70%; the next-highest is 61%). Given the magnitude of Hernandez's reputation, and the fact that every stat besides SFR sees him as superlative, I think we can comfortably conclude that these results understate the magnitude of Hernandez's defensive contribution. But this is what the equation spit out. Note that this does not include scooping, which I understand was a strength of Hernandez's--Sean Smith has him saving a further 28 runs on scooping alone just from 1985-1990, at the tail end of his career.

Furthermore, the guys who were clobbered in my old rankings--Killebrew and McCovey--fare much better here. Killebrew shows up as actually *above* average at first base; he did his damage in the field when he was played elsewhere. (Just how much damage we'll have to see). And McCovey emerges with a very respectable score for such a long career, although that doesn't include his misadventures in the outfield when Cepeda manned first. I imagine that, given his nickname, he was probably a competent scooper as well, which might have been enough to make him average overall at the position. That would be enough to move him up substantially in my rankings.

I will post revised charts and rankings for all the post-1893 MLB first basemen as soon as I can--hopefully before the end of this ballot.


Eddie Murray +46 (high of +9 in 1987 and +8 in 1979, only below average in 1988 and 1992-93)
Keith Hernandez +38 (high of +8 in 1985 and +7 in 1983; only below average in 1989 and 1990)
Mark McGwire +9 (+23 from 1989-1991 and -16 from 1996-99)
Harmon Killebrew +5 (only counting years when 1B was his primary position)
Will Clark -3 (+6 in 1989)
Willie McCovey -24 (league average from 1959-1972 excluding his OF seasons)
Dick Allen -26 (only counting years when 1B was his primary position; +5 in 1972 but -14 in 1969 and -9 in 1975)
   118. bjhanke Posted: May 22, 2008 at 01:49 AM (#2790398)
DL from MN says, "If you only consider white first basemen. I prefer Ben Taylor to George Sisler."

Do you mean "If you do NOT only consider white...."? I've never heard of Ben Taylor, and when I looked, the only one I could find played 52 games at the start of the 1950s. I doubt that's who you mean, if he's white. I looked for other first named Taylors, and that didn't help. If Taylor is a Negro League guy, I don't know anything about him, although I'm certainly willing to learn.
   119. RobertMachemer Posted: May 22, 2008 at 02:38 AM (#2790530)
I posted a similar question in the catchers thread earlier today, but this appears to be where posters' attention is, so I'll re-work my question here.

Compare Gene Tenace and Bill Terry.

Terry played 1721 games (most of which were at first base. According to bb-ref, he had a 136 OPS+ and 333.4 adjusted batting runs.

Tenace played 1555 games (about 900 at catcher, 600 at first base). According to bb-ref, he had a 136 OPS+ and 266.8 adjusted batting runs.

Even granting that Terry was likely a better fielder at first than Tenace was at catcher, the fact that Tenace was roughly as good a hitter (in fewer games, admittedly) while getting about 60% of his games in behind the plate ought to do a lot (!) to narrow the gap in value, yes?

And yet Terry is in the Hall of Merit and Tenace (in the "2008" vote) was tied for 92nd. Why? Why is there such a discrepancy between what you perceive to be Tenace's value and what you perceive to be Terry's value (or, as I asked in the catchers thread, Freehan's value)? I'm not saying you have it wrong -- far from it, I trust your guys' evaluations a lot -- I'm just surprised by your collective evaluation of Tenace as compared to Terry and Freehan (who are easier to compare to Tenace for their both having similarly, though slightly longer, short careers).
   120. RobertMachemer Posted: May 22, 2008 at 02:39 AM (#2790532)
Aargh, I can't edit my post? Shoot, sorry.

"Even granting that Terry provided more offensive value and was likely a better fielder at first than Tenace was at catcher..."
   121. OCF Posted: May 22, 2008 at 04:58 AM (#2790644)
Well, by the method of my post #26, which involves RC from a Stats Handbook, turned into RCAA (but not position-adjusted) and adjusted for the run value of wins:

Terry .  69 68 59 53 47 38 37 35 28 18  8  6  1  0
Tenace 
53 46 39 38 36 30 28 18 17 15 13 12  1 --4
Freehan  53 46 28 28 25 21 13  4  2  1  1  1 
--4-10-12 


I do think we might well be shorting Tenace in the voting. If he were evaluated as if he were all catcher, then his offense would be about as good as Cochrane/Dickey/Hartnett, and that's saying a lot. But we wasn't all catcher. What this looks like to me: Tenace simply wasn't as good a hitter as Terry. (But I haven't done Dan R. style standard deviation adjustments, which might close part of that gap.) On the other side, Freehan wasn't, overall, the hitter Tenace was - but he did have a couple of peak years in there.
   122. bjhanke Posted: May 22, 2008 at 08:47 AM (#2790687)
About Dick Allen. As you all now know, I'm from St. Louis, and certainly old enough to have been aware of at least the propaganda that the Cardinals put out regarding Allen. Here are some highlights:

The Cardinals claimed that someone in their organization, who they trusted, thought that Curt Flood was about done. The person thought - I'm serious, this is what came out of the organization - that Flood's body was aging quickly compared to a normal person's. This sounds like idiocy and ass-covering, and there are any number of reasons why Flood would have fallen apart after he left the Cards, but that's what they said and they said it before he collapsed in Washington. I don't really give it any credence, but it is there.

Dick Allen had a bad injury during his one season with the Cards. That's why he's missing playing time. Until the injury, he looked like he might actually break the team's all-time record for home runs, despite it being 1970 and the park being death valley. He hit the most mammoth shots I ever saw in the ballpark, including McGwire's. As far as the fans were concerned, including me, he was a fan favorite. I still have a soft spot in my heart for Allen.

When the team traded him for Sizemore, the reason that they gave was that Sizemore was a middle infielder who could hit .300. He had indeed hit .306 for the Dodgers in 1970. I think this may have been a damage control excuse, and the team had just decided that they had to get rid of Allen, but I only have a couple of insider sources for that speculation. The fans were up in arms, but the .300-hitting middle infielder claim did calm things down somewhat.

Mike Shannon was a time bomb by the late 1960s. Shannon has some sort of congenital body problem whose name I have forgotten. He was apparently a star-in-the-making in college football until the doctors told him that continuing that sport would kill him (sources: Bob Broeg and Bing Devine). He switched to baseball, but by 1968, he was having to ask doctors every spring whether he was allowed to continue his career. In 1970 they told him no more, and he quit. He was also pretty bad that year, but it was the medical stuff that did him in.

I don't know how much that helps anyone get hold of Allen, but it's all come up in the thread, so I posted up.
   123. bjhanke Posted: May 22, 2008 at 09:25 AM (#2790690)
OCF said, We've talked some about differential effects of ballparks on different kinds of hitters. We're not all of one mind about this, but here's the position I have staked out for myself: it doesn't matter. If you want to argue that Gavy Cravath took unusual advantage of the Baker Bowl, or Mel Ott took unusual advantage of the Polo Grounds, or Joe DiMaggio was unusually hurt by Yankee Stadium, then you're arguing a "what if."

I don't think that "what if" is what I'm advocating. I think I'm advocating better ballpark adjustments. Simply using a runs adjustment is far too simple and inaccurate for me. I think you really should adjust each stat individually and also adjust separately for lefties and righties. If you have access to any of my old books, you will find that I have thought this since the 1980s, as I printed those adjustments every year as part of each team's stat package. I found that looking at those adjustments was really informative and more accurate than the ones involving runs only. Since I've thought this since the 1980s, you probably can't talk me out of it. On the other hand, if everyone thought exactly the same things, sabermetrics would be solved like tic-tac-toe. Boring.

OCF also said, "I'll claim that, offensively, the best four or five years of Sisler's career do not top the best four or five years of Frank Chance's career."

Win Shares thinks it's close, but Total Baseball clearly separates Sisler from Chance. Win Shares has Chance in the middle of its leader board for NL position players in 1903, 04 and 06, but in no other years. That's about what they have for Sisler. TB has Chance fourth in 04, but that's it. I tend to lean on those leader boards a lot when looking at baseball way way back then, because it's so foreign to the game we now have that I have troubles getting hold of what is really what. TB thinks that the big advantage Sisler has over Chance is the two huge years when he hit over .400. Those years dwarf anything Chance put up. Win Shares does not think so much of those years, and has them essentially even with the best Chance years. Since I think that Win Shares has Sisler's defense underrated, I go with TB (note that I am making no claims about ballpark adjustments). Still, I will concede that Frank Chance's peak does not look silly compared to Sisler's. I can see where you're coming from, but I do still think Sisler was better.

All in all, those were good comments. Thanks.
   124. DL from MN Posted: May 22, 2008 at 01:38 PM (#2790790)
bjhanke - Ben Taylor was a Negro League player. You can learn more about him at the Hall of Fame website because he was elected to the Hall of Fame in 2006. Or you can go here:

http://en.wikipedia.org/wiki/Ben_Taylor_(Negro_Leagues)

I meant what I said - Sisler is the best 1B of the first 25 years of the 20th century if you only consider white players. Ben Taylor compares favorably to Jake Beckley and Keith Hernandez - good defensive reputation, long career, just enough bat with high OBP.

Machemer - You're trying to argue that Tenace should be in by saying he's almost comparable to two of our most marginal HoM players. You should realize that there are a LOT of players at the margins that Tenace competes with for 15 ballot slots. That he gets any votes at all is really a significant accomplishment.
   125. bjhanke Posted: May 22, 2008 at 03:20 PM (#2790937)
DL from MN gives me the info, "bjhanke - Ben Taylor was a Negro League player."

Thank you! I will look him up tonight, at least in Wikipedia. Always glad to find out about another good player. Um. About the ballotting. If I should decide that Ben Taylor was better than some of our 20 first base candidates, can I substitute him for the weakest one, or is it the policy to stay with the 20 we've got? I'd stay with the 20, but I don't know HoM policy. You never know - I may think that Ben Taylor is better than Keithe Hernandez, to whom he sounds most comparable.
   126. Chris Cobb Posted: May 22, 2008 at 03:33 PM (#2790962)
If I should decide that Ben Taylor was better than some of our 20 first base candidates, can I substitute him for the weakest one, or is it the policy to stay with the 20 we've got?

These ballots are strictly for ranking the players elected to the Hall of Merit. Some voters may believe that some unelected players are in fact superior to players who have been elected. If so, they are welcome to indicate that in their comments, but unelected players are not included in these rankings. But our next opportunity to vote for players not yet elected to the Hall of Merit will be in the balloting for the 2009 election to the Hall of Merit, which will take place this fall.
   127. bjhanke Posted: May 22, 2008 at 09:25 PM (#2791088)
Chris - Thanks for the info. Would you believe that I actually did not know what exactly this vote was about? Now it makes sense. The list of 20 is not a list of candidates; it's a list of already-elected members. That makes sense. I have a much better idea of what to do here now. Thanks again! - Brock
   128. RobertMachemer Posted: May 22, 2008 at 10:33 PM (#2791153)
Machemer - You're trying to argue that Tenace should be in by saying he's almost comparable to two of our most marginal HoM players.
Well, sorta. Are Freehan and Terry both better candidates than Tenace? The fact that they're in and Tenace is finishing 90th in the votes suggests that collectively the HOM voters think that Freehan, Terry, and roughly 90 (or more) players are better than Tenace. I'm not sure that even Freehan and Terry are.



Here are rankings of the first basemen and catchers voted into the HoM by adjusted batting runs according to bb-ref (remember, Tenace had 266.8 for his career):
Gehrig    982.9     Torre      304.0
Foxx      782.9     Bench      267.2
Brouthers 639.0     Cochrane   256.2
McGwire   589.2     Dickey     247.7
Anson     569.1     Harnett    242.9
Connor    552.3     Berra      238.6
McCovey   548.4     Simmons    215.2
Mize      513.1     White      206.4
Killebrew 508.4     Fisk       192.2
Allen     476.3     Bresnahan  188.2
Murray    462.9     Ewing      180.7
Greenberg 455.1     Carter     162.4
Clark     404.6     McVey      142.5
Terry     333.4     Campanella 141.5
Hernandez 328.4     Bennett    100.7
Beckley   290.4     Freehan     99.2
Banks     259.8     Gibson     
*****
Sisler    253.5     Mackey     *****
Start     106.4     Trouppe    *****
Suttles   *****     Santop     *****
Leonard   ***** 


I recognize that it's difficult to know what to do with Tenace because he split time between two positions. But Ernie Banks did too, provided roughly the same overall offensive value as Tenace (while spending a larger percentage of his time at 1B), and sailed into the HoM in his first year of eligibility.

Here are some players who have been voted in since Tenace became eligible. (I'll skip the Negro Leaguers who have no adjusted batting runs to be compared). Here are the bottom ten players (by adjusted batting runs) who have been voted in since Tenace became eligible.

McGwire     589.2
Schmidt     576.1
Yastrzemski 535.7
Jackson     519.7
Brett       504.7
Morgan      496.9
Boggs       475.0
Murray      462.9
Winfield    440.7
Gwynn       437.6
Carew       418.0
Clark       404.6
Rose        394.6
Browning    379.6
DwEvans     362.4
Molitor     349.2
Raines      332.8
Hernandez   328.4
Wynn        298.2
Keller      295.2
Beckley     290.4
DaEvans     278.6
Grich       271.5
Bench       267.2
Roush       251.7
Whitaker    227.2
Yount       218.1
Dawson      216.4
Simmons     215.2
Jones       208.2
Ripken      197.8
Fisk        192.2
Bresnahan   188.2
Boyer       165.0
Carter      162.4
Sandberg    150.5
Trammell    123.9
Nettles     118.2
Randolph    110.4
Fox         
-81.8
Smith      
-132.3 


Obviously, Tenace shouldn't get consideration over players like McGwire, Schmidt, and Yaz. But there are 41 players on that list, and 17 of them had fewer adjusted batting runs than Tenace, for whatever that's worth. Obviously many of those received consideration for playing difficult defensive positions for the bulk of their career, but Tenace played a difficult position for half his career too.

It would not surprise me if Tenace has one of the highest adjusted batting runs totals outside of the Hall of Merit right now. (Anyone want to check that?) If that's true, I'm not sure why he continues to languish as far down on the ballot as he does, given that he spent half his career at catcher.

I'm not saying he should be voted in. I'm suggesting that his merits be reconsidered.
   129. Paul Wendt Posted: May 22, 2008 at 10:51 PM (#2791172)
DanR #119
the last version forced the 1B defensive stats to have the same standard deviation as Chris Dial's Zone Rating-based data, despite the fact that they were less reliable. The standard deviation of these new numbers reflects *precisely* the level of confidence we can have in them--68% (so with an effective 32% regression to the mean) for 1987-2002, 54% (an effective 46% regression to the mean) for 1956-1986, and 14% (a whopping 86% effective regression to the mean) for 1893-1955.

Dan,
You have a flat 86% discount for 1893-1955, then a flat 46% discount for 1956-1986. The "curve" goes from 14 to 54 over a 93-year period, and it doest that all in one giant step up from 14 to 54 over the winter of 1956. There is no way anyone's confidence in the ratings should follow that curve even imprecisely. I don't know precisely ;-) what you have here but I know you are overselling it!
   130. Blackadder Posted: May 22, 2008 at 10:52 PM (#2791176)
There were quite a number of eligible players with more adjusted batting runs than Tenace. Quickly eye-balling the list, the highest eligible player not in seems to be Bob Johnson at 391, although Fred McGriff at 447 will probably be an interesting argument.
   131. David Concepcion de la Desviacion Estandar (Dan R) Posted: May 22, 2008 at 10:58 PM (#2791179)
Paul Wendt,

Yes, there is absolutely a way that everyone's confidence in the ratings should follow that curve precisely. 1956 is currently the beginning of Retrosheet data. As Retrosheet pushes further back in time, we'll be able to extend that 54% confidence level with it.
   132. OCF Posted: May 22, 2008 at 11:03 PM (#2791187)
I'm looking at posts #112 (Chris Cobb) and 124 (bjhanke) - hey, you guys are talking about the team I followed! I should add that I didn't read the St. Louis press, since I lived in Oklahoma - what I knew at the time came mostly from Harry and Jack, delivered by radio. What Brock knows is a lot closer to being "inside" information than what I remember.

I do believe that it was kidney disease that ended Shannon's career. In retrospect, look at him: he never really hit enough to be a star corner outfielder, and while that nominally had more value at 3B, he fielded that position like an out-of-position outfielder. I know he was very popular, but it's hard to see him as a cornerstone of the franchise.

As for the decline of the pitching staff: One part of it is Gibson. He was a deity in 1968 at the age of 32, then a demi-god in 1969 at 33, then a good workhorse #1 starter in 1970 at 34. I mean, I'll still take the 1970 version of Gibson for my team - but the slope is downward. Given how high the peak in 1968 was, how could the slope be anything but downwards? My equivalent records are 26-7 for 1968 (the best post-1920 year for anyone in my system), 24-10 for 1969, and 21-11 for 1970. He would slip to an equivalent 15-12 for 1971, then rebound to 21-10 for 1972, before then getting old.

But what happened to the young starters? Nelson Briles was, at age 23, one of the major heroes of 1967, as a swingman pulled into the rotation by Gibson's broken leg, followed by two years as an average pitcher with a lucky-stiff W-L record, followed by utter collapse in 1970 at age 26. And then there was Steve Carlton, a year younger than Briles. Everyone knew he was a major talent, and he was good, and sometimes (as in 1969) he was great - but was he really making progress? There was Mike Torrez, there was Jerry Reuss - boy they had some young arms. Briles rebounded some - after he got traded away. Carlton burst out with his greatest season immediately after being traded away. Some of the others had their best success elsewhere. Yes, some of it might have been the defense; the decision to play Cardenal in CF was unfortunate. (Couldn't they just have held onto Bobby Tolan?) But who was the pitching coach at the time, and why didn't they get more out those young arms?

Side bit: has there ever been a better recovery from an injury than Gibson's recovery from his 1967 broken leg? He wasn't having all that great a season when it happened. But once it did happen, he was out less then 2 months, and he was instantly bulletproof as soon as he came back. His Sept. 1967 starts and his 1967 WS work would fit comfortably into his unreal 1968 season, and his 1969 season was a great season even if not quite 1968. It took 2+ years for whatever effect it was to wear off.

And then there's the racial climate issue. We all know about the long-term history of St. Louis; it's not a pretty picture. But it's amazing what sheer fandom will do, and Gibson and Brock were (and still are) beloved figures for Cardinal fans. And Gibson, in particular, had an image and reputation as a man - a tough, no-nonsense, stand-up-for-himself man. For whatever reason, St. Louis did not erupt in riots in 1968 when so many other cities did. In all that, I don't sense that the fans had all that much racially-tinged distance or coldness towards Allen. Of course, he wasn't Gibson or Brock - no one else could be - but I don't think it was a hostile climate for him.

(That was all kind of rambling and pointless. But we should all get a chance to ramble pointlessly some times.)
   133. Paul Wendt Posted: May 22, 2008 at 11:12 PM (#2791200)
Brock #115
Hobbyhorse alert: It is my opinion that the single most useful thing that sabermetrics could do and has not done would be to compute park effects separately for righties and lefties, both hitters and pitchers. I'm serious. I think this is project #1. I can't do it, because I lack the database needed. But the database should be out there, and I really do wish that one of the organizations that has it would compute those historical ballpark effects splits, both for handedness and also for individual stats. All I can find now is splits for homers. I want them for everything, and I want that lefty/righty split.

I copied that because I planned to reply usefully. But I have replied too expansively below (writing from the bottom up) and run out of time. I suspect that you overestimate what is in the databases. Maybe I misunderstand. Maybe someone else can comment usefully on what I have quoted.

. . .
Regarding TPR (and TPI for pitchers) you are certainly right about the long runs of small negatives, short runs of large negatives, ups and downs like -3 +1 -3 +1 -3 +1 that come to a clear negative in aggregate. There is a long long list of mlb players whose teams should have released them or benched them even a decade before their actual retirements, if Pete Palmer's benchmark (zero) is routinely available even to most teams at most fielding positions most of the time --not to mention if that is "freely available talent" in a perfect labor market.

Ironically, I first noticed the magnitude of the effect in case of Dave Concepcion, the favorite player --or favorite sabrmetric focus-- of Dan Rosenheck. (See #119 and see also a thread on Dan Rosenheck's uber-stat unfortunately called WARP, same as Davenport's. For several good reasons DanR's is not included in "battle of the uber-stats" but it is another uber-stat.) In the 1970s the Philadelphia Bulletin and the radio team and, well, everyone knew that Larry Bowa and Dave Concepcion were the two great shortstops. Twenty years later when I bought TB3 --the only one distributed in a paperback edition-- I looked up Dave C and found that he was a Hall of Fame quality player for a decade and then a big negative for a decade. Or so he was according to TPR.

(If that is "ironic" that is only true here, where DanR has raised the local profile of Dave Concepcion very high.)


#125:
<i>I don't think that "what if" is what I'm advocating. I think I'm advocating better ballpark adjustments. Simply using a runs adjustment is far too simple and inaccurate for me. I think you really should adjust each stat individually and also adjust separately for lefties and righties. If you have access to any of my old books, you will find that I have thought this since the 1980s, as I printed those adjustments every year as part of each team's stat package. I found that looking at those adjustments was really informative and more accurate than the ones involving runs only.


You may be misinterpreting wholly or partly.

1. As OCF said, many of us take the general position "it doesn't matter". And many do not.

2. I believe that most who say "it doesn't matter" will agree about your work as you describe it, measuring particular ballpark effects rather than a single effect on runs or "scoring environment". Namely, that particular work may be more informative and more accurate for the purpose of evaluating skill (who is the best player?) or predicting achievement (who will do most for us next year?). For that you should take the player up out of his past ballpark context and assess him either transcendently or down in your future ballpark context. But it doesn't matter for the purpose of evaluating past achievements (who was the most valuable player?).

- I believe that most of us here agree with that. The disagreement concerns how central is value to the Hall of Merit? (or to "merit" lowercase? or to the Hall of Fame?, etc) Some would say to OCF, yes, you are right about value, but here I should try to vote for "the best player" not the one with most valuable past achievements.
   134. Paul Wendt Posted: May 22, 2008 at 11:13 PM (#2791202)
Oops.
All three of my quotations are one paragraph long. "Ironically" begins my reply to the second one.
   135. Mark Donelson Posted: May 22, 2008 at 11:15 PM (#2791204)
Robert:
We're all kind of engaged in a different analysis right now—ranking the guys who are in—so you're not getting a lot of attention to your Gene Tenace inquiries. I suspect others could respond to this more fully, but to get a fuller sense that we at least did look into Tenace's merits when he became eligible, as well as why Freehan was elected, you might check out these threads:

Gene Tenace thread (the beginning and end--the middle is a long digression about how to pronounce his name, and lots of other things)

Bill Freehan thread (See in particular Tiboreau's post #58, which addresses Freehan's defense--the main reason, I think, that he made it and Tenace did not. It comes down not only to the years Tenace spent at first, but also to Freehan's significantly stronger defense at the catcher position.)

Tenace does continue to have support on a few ballots, though not many. But you do have to remember that due to the way voting works, if every voter but the few who vote for Tenace had him at #16, the results would look the same--if you're not in someone's top 15, you don't get any points. And, of course, Freehan himself only squeaked in, so many voters likely do rank Tenace higher--it's just that most of them who do would appear to have neither on their ballots, or (possibly) terribly close. In other words, the difference between them in our consensus is probably not nearly as great as the difference between "elected" and "92nd" make it sound.
   136. Mark Donelson Posted: May 22, 2008 at 11:28 PM (#2791229)
One other thing I forgot to mention, of course, is that most of us don't vote solely (or even largely, in many cases) on adjusted batting runs. By WS, Freehan is ahead career-wise, 267-231. By WARP, this week Tenace leads 73.1 to 71.7, a near tie.

And of course there are various peak vs. career things going on with various voters.

So to me, while Tenace perhaps deserves more attention than we've given him, it doesn't seem as if he's a slam-dunk "we missed him."
   137. Mark Donelson Posted: May 22, 2008 at 11:31 PM (#2791234)
And, of course, Dan R. has provided some hard numbers that go into this much more fully on the catcher results thread. I knew one of the heavy hitters would get to it soon! :)
   138. Chris Cobb Posted: May 23, 2008 at 12:17 AM (#2791297)
Mark's well-phrased explanations cover the Tenace case clearly.

I agree that Tenace is a player deserving of more attention than we have given him, and Robert Machemer has done a good job of reminding us of that.
   139. bjhanke Posted: May 23, 2008 at 04:56 AM (#2791560)
Paul Wendt says, "I believe that most of us here agree with that. The disagreement concerns how central is value to the Hall of Merit? (or to "merit" lowercase? or to the Hall of Fame?, etc) Some would say to OCF, yes, you are right about value, but here I should try to vote for "the best player" not the one with most valuable past achievements."

Ah! I think I understand you now. It makes sense. I use the same distinction to describe the Historical Abstract Win Shares Method (what DO you call the application of WS to the HA?). Before Win Shares, Bill made adjustments of all kinds as though he were trying to get the player into a normalized ballpark. Now, he adjusts for results like actual team games won. It occurred to me that, if you were considering a Hall of Fame candidate's career - that is, if you wanted to know "how good" the player had been, you would use Type 1 criteria, from before WS. If you were considering a one-year MVP, though, and wanted to know which player had had the most effect on his team's record for that season, regardless of how much effect he might have had on any other team, if he had played elsewhere, the WS system adjusts for that goal. It gives you, to quote Paul, "the most valuable past achievements." I've even taken to calling the two approaches The Hall of Fame Approach and The MVP Approach to sabermetrics. I should have seen that OCF and I were having that discussion more than anything else. Thanks for the mind-clearing.

And yes, OCF is right in that his (or her) approach, like WS, is better suited to MVP-style analysis than mine is here. I guess what I should say is, "I think that, if you are going to make ballpark adjustments, they would be better if you included the splits I suggest." And also yes, you may be right about the database. To do what I want would indeed involve a play-by-play database. I was able to print those stats in my books because STATS, Inc. has just such a db. I've seen enough splits that I just thought someone had compiled all the pbp there was. I could be wrong. The splits we have may have been the result of specialized research kept narrow because there was a goal and a deadline.
   140. DanG Posted: May 23, 2008 at 01:01 PM (#2791666)
I agree that Tenace is a player deserving of more attention than we have given him, and Robert Machemer has done a good job of reminding us of that.

Last month, I made an attempt (mainly in the "Once We Catch Up..." thread) to bring this project to a systematic focus on Tenace and the other top non-HoMers. My suggestions were largely dismissed, as most everyone here is only interested in studying "our" HoMers.

I tossed out a list of catchers deserving of comparison to the HoM catchers:

21. Howard
22. Schang
23. Parrish
24. Munson
25. Lombardi(HOF)
26. Tenace
27. Porter
28. Petway
29. Boone
30. McGuire
-------------------------
~40. Ferrell(HOF)
~45. Schalk(HOF)
   141. Paul Wendt Posted: May 23, 2008 at 07:26 PM (#2792154)
Paul Wendt,

Yes, there is absolutely a way that everyone's confidence in the ratings should follow that curve precisely. 1956 is currently the beginning of Retrosheet data. As Retrosheet pushes further back in time, we'll be able to extend that 54% confidence level with it.


I stand corrected.
14% is horribly low confidence if I understand the metric and I am hopeful that one can do better without nearly complete Retrosheet data (play-by-play). Perhaps one can argue by direct comparison that the earlier estimates are more than one-quarter so good as the later ones. What if we look at all regular players who seem to be at relatively stable periods in their careers (not Jackie or Brooks Robinson). Compare 1954-55 ratings with 1956-57 ratings. Does it appear that the earlier methods, lacking pbp data, do incorporate only one-quarter of the effective information, the rest being noise?

By the way, getting to Brock Hanke's questions and presumptions about the data, have you tried using the 1921-1922 data that Retrosheet has published? --the box score project, filling play-by-play gaps with Retro-standard box scores. A season is complete when there is a box score for every game.

--
Brock continued
And yes, OCF is right in that his (or her) approach, like WS, is better suited to MVP-style analysis than mine is here. I guess what I should say is, "I think that, if you are going to make ballpark adjustments, they would be better if you included the splits I suggest." And also yes, you may be right about the database. To do what I want would indeed involve a play-by-play database. I was able to print those stats in my books because STATS, Inc. has just such a db. I've seen enough splits that I just thought someone had compiled all the pbp there was. I could be wrong. The splits we have may have been the result of specialized research kept narrow because there was a goal and a deadline.

Retrosheet has released "all the play by play" there is, and that is every major league season nearly complete, back to the mid-1950s (1956 at last incorporation by DanR, see just above). And also for NL 1911 and 1922, but those seasons have been completed by including box scores for many games --contrast the reliance on few or zero box scores for leagues in the Retrosheet era. For 1872, 1874, and 1921 box scores only.
Retrosheet directory of years

For most of major league history, complete games were common and many of the incomplete games were "one-hand complete": the starting and relief pitchers used by one team were all right- (usually) or all left-hand throwers. So the switch-hitters batted from only one side of the plate. Maybe some private organizations may have covered some seasons approximately, using
- box scores or official daily records from the league offices and the
- throwing hands of the starting pitchers as proxy for entire games

Home-away splits are the more important for measuring those particular ballpark effects such as you advocate. Home-away splits may be compiled from box scores or official dailies and a good game log, which Retrosheet has published for every season from 1871 to 2007 (select a year, then a team, then Game Log). Maybe some private organizations have covered some seasons by home-away or even by ballpark, but without any "hands" data (pitcher throwing hands and switch-hit batting hands).
   142. Paul Wendt Posted: May 23, 2008 at 07:38 PM (#2792172)
I don't believe private organizations are paying many people or paying much for historical data or its compilation. One thing that does happen, I know, is that simulation players compile data and distribute it freely. A strat-o-matic simulation player compiles starting lineups for all of the 1890 Players League games and not only uses it for reference in replaying the season but gives a copy to strat-o-matic and posts another to a users group. Strat-o-matic and Diamond Mind may both put that 1890 PL starting lineup data into their proprietary formats within a few years.
   143. Paul Wendt Posted: May 23, 2008 at 07:50 PM (#2792190)
23. David Concepcion de la Desviacion Estandar (Dan R) Posted: May 23, 2008 at 01:18 AM (#2791562)
TomH, can we get some explanation on Murray over Mize?

24. TomH Posted: May 23, 2008 at 09:41 AM (#2791703)
Sure. Murray has a WS advantage of 100. With war credit, Mize would get closer, and if you ue WSAB (above bench), it's pretty even.

Mize was a better hitter by measures above avg (bbref's batting wins, or BRAA, OWP times PT), but
1. the career length diff is huge,
2. I rate Murray slightly better on D,


27. Chris Cobb Posted: May 23, 2008 at 10:23 AM (#2791755)
First Base Ballot.
. . .
5. Johnny Mize. Total = 373. Could mash and was sound defensively. Players of his generation (who had the bulk of their career in the 1940s) who rank ahead of him in my system are Ted Williams, Stan Musial, Josh Gibson, Satchel Paige, Joe Dimaggio, and Luke Appling. All but Appling are inner-circle greats, and he’s about even with Appling.

6. Dan Brouthers. est. Total = 360. Could mash! Best hitter of the 1880s. WARP2 rates him as equal to Mize in an all-time context. But he was a liability in the field and a bit injury-prone, so that drops him a step below Connor and Mize, to whom he is otherwise almost identical.


Johnny Mize sound defensively? Brouthers a liability? That is a big difference to my ear but I'm not sure how good is "sound"?
I think Mize is commonly supposed to be a weaker fielder than Gehrig or Foxx, or Murray, or Anson and Connor among those who (uncommonly) know they played first base.

Allen(?), Beckley, Brouthers, Greenberg, Killebrew, McCovey, Mize, Suttles.
I think these are the ones some of us call weak or liability fielders. Is that right?
   144. Paul Wendt Posted: May 23, 2008 at 07:52 PM (#2792193)
Not Anson at 45, Murray at 40, McGwire at 35 but everyone in his prime seasons
   145. John (You Can Call Me Grandma) Murphy Posted: May 23, 2008 at 08:06 PM (#2792206)
Allen(?), Beckley, Brouthers, Greenberg, Killebrew, McCovey, Mize, Suttles.
I think these are the ones some of us call weak or liability fielders. Is that right?


I don't think Beckley was considered a bad fielder, Paul.
   146. David Concepcion de la Desviacion Estandar (Dan R) Posted: May 23, 2008 at 08:07 PM (#2792207)
Paul Wendt, 14% is the r2 I got from a regression of DRA, FWS, and FRAA against my PBP average from 2003-07. I don't see how looking at 1954-57 data would make a difference in the accuracy of our estimates.

I don't calculate my own fielding metrics. I'm just trying to find the best mix of the available ones, plus regression to the mean to account for uncertainty, to help us rank players.
   147. DL from MN Posted: May 23, 2008 at 08:19 PM (#2792220)
It is possible that FRAA was better at estimating runs in 1960 than it is at estimating runs in 2007 due partially to strikeouts.
   148. Paul Wendt Posted: May 24, 2008 at 12:09 PM (#2793130)
You may be able to (in)validate the model that uses 2003-2007 pbp data.
For example if DL is right and the effect is large enough, you should be able to invalidate it.

I was thinking that you roll your own (although you have explained otherwise more than enough times). If not then you need to rely on others to try that 1922 pbp data which may be a long wait.

Here is the other part of my idea which may be applicable. Take the best ratings for 1954-55 (without pbp data) and the best ratings for 1956-57 (with pbp data). Compare them --maybe quantitative, maybe rank-order-- for those players who were mid-career regular players both before and after. Do the same for four consecutive seasons when you have the best pbp-based ratings for all four seasons. Hm, probably that should be 1956-1959.

I was thinking that some of the basic ratings (not your composite use of them), like Runs Created by Bill James, are calculated differently depending on what data are available, so they would differ in formulation for 1954-55 and 1956-57. Now I think not. Then my idea probably boils down to comparing the with-pbp and without-pbp ratings for 1956-1960 rather than 2003-2007.
   149. Paul Wendt Posted: May 24, 2008 at 12:23 PM (#2793137)
Checking Retrosheet directory again, NL 1911 and NL 1922 seem to be the two league-seasons released by Retrosheet with pbp data where available.
If I count correctly, pbp is missing for 22 of 156 New York games in 1922 (Dave Bancroft 1922 at retrosheet).
For 1911, the majority of Cincinnati games are missing (Dick Hoblitzell 1911), 35 games missing for Chicago (Jimmy Sheckard 1911).
   150. Paul Wendt Posted: May 24, 2008 at 12:34 PM (#2793139)
Over in the ballot thread someone called George Sisler a "slap hitter".
More like George Brett, a fine extra base hitter.

George Sisler 1916-1922 rank in league
(0 = ten)
4 2 3 4 1 3 1 Hits
4 3 5 7 1 5 1 : singles (*never* higher than his rank in hits)
- 4 0 9 2 7 3 : doubles
- - 8 2 2 1 1 : triples (developing power)
9 - - 2 2 9 - : homeruns
0 8 0 3 2 6 3 Extra Base Hits
7 4 5 3 1 5 2 Total Bases
0 4 4 5 2 3 1 Runs Created

6 5 1 2 2 1 1 Steals! (literally his strongest suit)
   151. Paul Wendt Posted: May 24, 2008 at 12:36 PM (#2793142)
'0' means rank ten and '-' means not in top ten
   152. Howie Menckel Posted: May 24, 2008 at 06:34 PM (#2793306)
Beckley supposedly had an erratic arm; I don't think he was considered a bad glove man.
   153. Howie Menckel Posted: May 24, 2008 at 07:18 PM (#2793341)
Career OPS+s, excludes Leonard and Suttles.
just a starting point; there's defense and era and league quality and many other adjustments.
(1 3 4 10) means the player's top 10 finishes in adj OPS+ are 1st, 3rd, 4th, and 10th.
separated by broad eras.

Dan Brouthers..........170 OPS+ in 7658 PA. (1 1 1 1 1 1 1 1 2 2 3 7)
Roger Connor...........153 OPS+ in 8837 PA. (1 2 2 2 2 2 2 3 3 3 3)
Cap Anson..............141 OPS+ in 11319 PA (1 2 2 3 4 4 4 4 6 6 7 9 9 9 10)
Jake Beckley...........125 OPS+ in 10470 PA (3 5 6 8)
Joe Start..............121 OPS+ in 4911 PA. (5 7 9 10) give big 1860s bonus

Lou Gehrig.............179 OPS+ in 9660 PA. (1 1 1 2 2 2 2 2 2 3 3 5)
Jimmie Foxx............163 OPS+ in 9670 PA. (1 1 1 1 1 2 2 4 4 4 5 8 9)
Johnny Mize............158 OPS+ in 7371 PA. (1 1 2 2 2 2 2 3 3) give big WW II bonus
Hank Greenberg.........158 OPS+ in 6096 PA. (2 2 2 2 3 4 4 9) give big WW II bonus
Bill Terry.............136 OPS+ in 7111 PA. (4 4 5 7 7 9)
George Sisler..........124 OPS+ in 9013 PA. (2 3 3 4 5 6 8)

Mark McGwire...........162 OPS+ in 7660 PA. (1 1 1 1 2 6)
Dick Allen.............156 OPS+ in 7314 PA. (1 1 1 2 2 3 5 6 7 8)
Willie McCovey.........147 OPS+ in 9686 PA. (1 1 1 3 4 4 5)
Harmon Killebrew.......143 OPS+ in 9831 PA. (3 3 3 4 4 4 4 4 5 5)
Will Clark.............137 OPS+ in 8283 PA. (2 2 2 7 7)
Eddie Murray...........129 OPS+ in 12817 PA (2 2 2 2 3 5 7 8 10)
Keith Hernandez........128 OPS+ in 8553 PA. (3 3 4 5 6 8)

..........

pct of games by position (estimates for Negro Leaguers)
Start 100, Gehrig 100, Mize 100, KHernandez 100, Beckley 100, Terry 99, Brouthers 98, WClark 98, Sisler 97, McGwire 96, Leonard 95, Connor 88, McCovey 88, Foxx 87, Anson 83, Greenberg 83, Murray 81, Suttles 70, Allen 47, Killebrew 40

hybrids, with percentages:
Killebrew 40 1B, 33 3B, 20 OF
DAllen....47 1B, 38 3B, 15 OF
Suttles...70 1B, 30 OF
Murray....81 1B, 19 DH

other HOMers with at least 10 pct at 1B:
Banks 51, Carew 50, Wilson 45, Stargell 40, Stovey 37, Torre 36, Charleston 35, Musial 35, DaEvans 32, McVey 31, Rose 27, Jennings 26, Lloyd 25, Yastrzemski 23, Heilmann 22, Ewing 19, Kelley 16, Delahanty 15, Hines 12, Lajoie 12, Mantle 11, FRobinson 11, Spalding 10, O'Rourke 10, Dihigo 10, JRobinson 10, Irvin 10
   154. Howie Menckel Posted: May 24, 2008 at 07:20 PM (#2793343)
hybrids, with percentages (adds Greenberg):
Killebrew 40 1B, 33 3B, 20 OF
DAllen....47 1B, 38 3B, 15 OF
Suttles...70 1B, 30 OF
Murray....81 1B, 19 DH
Greenberg.83 1B, 17 OF
   155. Bleed the Freak Posted: May 29, 2008 at 05:25 AM (#2798119)
Considerable discussion regarding Adjusted Batting Wins prompted me to compile a list of the top 10 and 2 part time catchers that are eligible for the HOM:

1. Bob Johnson 392 OF
2. Jack Clark 389 OF
3. Albert Belle 372 OF
4. Norm Cash 371 1B
5. Ken Singleton 366 OF
6. Reggie Smith 356 OF
7. Rusty Staub 356 OF
8. Frank Howard 354 OF
9. Orlando Cepeda 344 1B
10. Chuck Klein 338 OF
~20 Brian Downing 284 ~700 G at C, 1500 OF, 800 DH
~30 Gene Tenace 267 ~900 G at C, 600 1B

This list reinforces the candidacy of Reggie Smith, the only man above with a quality glove.

Also, during the 7th inning stretch of the Cubs/Dodgers game last night, Tommy LaSorda declared Smith the best player he ever coached.
   156. DL from MN Posted: May 29, 2008 at 01:33 PM (#2798197)
> Reggie Smith, the only man above with a quality glove

Bob Johnson is regarded as a good corner outfielder. Norm Cash was a very good fielding first baseman.
   157. Esteban Rivera Posted: May 29, 2008 at 01:59 PM (#2798219)
This list reinforces the candidacy of Reggie Smith, the only man above with a quality glove.


Reggie's problem was that the glove and the bat didn't coincide for him very much. He either had the glove early on or the bat later on, but not both together.
   158. DL from MN Posted: May 30, 2008 at 01:37 PM (#2799484)
Aren't we supposed to be figuring out how to deal with outfielders and pitchers while we vote on the infield?

I'd keep LF and RF separate primarily to make the lists smaller. For pitchers there are clear groups based on era that make sense and I think we should vote by era.
Page 2 of 2 pages  < 1 2

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
The Ghost fouled out, but stays in the game
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Hot Topics

Syndicate

Demarini, Easton and TPX Baseball Bats

 

 

 

 

Page rendered in 0.8591 seconds
49 querie(s) executed