As I said in another thread, I think Concepcion will be the test case (8-10 years from now) as to the extent to which Joe Morgan becomes Frankie Frisch

If he (Joe) is still around, wielding more & more suasion with the VC--then his buddies like Concepcion will get in

and that's the only way he'll get in

Great analysis -- you really turned me around on Trammell.

One question, though: yes, Trammell was a lot better than "replacement level" -- but isn't that partly because it was such a weak era for shortstops? How much credit should Trammell get for his peers being terrible? Isn't it fair to consider that relatively few SS from the 70s & 80s are on par with today's breed?

A clarification: I did not use CAD to look at Concepcion, but a variant of Win Shares. I did not look formally at Trammell at all.

IAE, I looked at the period from 1974-1980 and 1982, and Concepcion was an great fielder, easily the best shortstop in the NL, though Ozzie caught up to him at the end. Oddly, the Reds were not, as one would think, a great fielding team -- the Red outfielders were just okeh, the third basemen were not any good at all (Anderson played Pete Rose there because he had four good hitting outfielders and no good hitting third baseman), and Bench and Morgan, while good fielders early in the period, dragged the team as they aged.

Scott -- I think Joe is wrong here. The early 1980s is as good a time as any for shortstops, with Trammell, Ozzie, Concepcion, Ripken, Yount and Dickie Thon. Their hitting stats don't look as good as the current crop's do, but that's true for all positions.

We all yammer on about the current golden age of SS, but the 80's seems to do very well. Yount, Ripken, Ozzie, Trammel, and Larkin all did big things in that decade and would in my opinion qualify as top 10 at their position all time (though someone might get nudged out in another 10 years). Or another way of putting it. What happens if Jeter continues to decline, Nomar keeps getting hurt, and Tejeda has a good peak but not much else due to his sudden aging? Suddenly the golden age would pale in comparison.

I always liked Concepcion as a player, but I think he falls short.

Trammell made it to both my MLB and STAT Internet ballot, so I'm sold on him. I think he was a little bit better than the Wizard, but an argument can be made for Ozzie.

Thanks guys . . .

Mark:

Nice catch on the turf issue with Concepcion. It's definitely an interesting tidbit.

From Charles:

"Scott -- I think Joe is wrong here. The early 1980s is as good a time as any for shortstops, with Trammell, Ozzie, Concepcion, Ripken, Yount and Dickie Thon. Their hitting stats don't look as good as the current crop's do, but that's true for all positions."

I should have said 1970's for SS's. You are right Charlie, around 1980 (when Yount really turned it on) the tide changed. But Ripken wasn't playing SS full-time until 1983, Thon wasn't full-time until 1982, Ozzie wasn't a decent hitter until 1982, etc. . . . the comment was directed at Concepcion's contemporaries from the 1970s. Sorry for the confusion.

Scott:

"One question, though: yes, Trammell was a lot better than "replacement level" -- but isn't that partly because it was such a weak era for shortstops? How much credit should Trammell get for his peers being terrible? Isn't it fair to consider that relatively few SS from the 70s & 80s are on par with today's breed?"

The replacement level studies I've seen show that even when the star level moves up (like AL SS 1996+, NL CF 1950s, etc.) the replacement level doesn't change much. When you remember that Major League baseball players are in the far right of the bell curve, once you are out to replacement level at a position the field is big enough that moving up a spot or two (to accomodate the new star glut) doesn't really make a difference.

The replacement numbers I used when comparing Trammell to Ozzie were generic and would apply in any era, I wasn't specifically comparing them to other 1980's SS, I was just using a 1980's level of offense, and I guesstimated what a repl. SS would hit at those offensive levels, the only differnce over a 2000s comparison would be to raise the offense for the era as a whole, not the SS's specifically.

Joe Ross:

I'm as big a Minoso champion as there's ever been. I just forgot about him, most of the guys I mentioned were a little more recent. He's my number one candidate on the new Vet's ballot, I hope he's alive when he finally gets in.

Jason:

I always remembered it that way too. I remember the game was on a Sunday afternoon on ABC, and Al Michaels and whoever was announcing with him picked a Giant-Tiger World Series. I always remembered it as grass too.

But when I went to either BR or Retrosheet to refresh my memory on the events, DET was on the left and TOR on the right (they are BOLD for winner, so that wasn't it) and I assumed that meant home team on the right, as in DET @ TOR. Sorry for the mixup.

Another late bit of info from a discussion with some other Primer folks (Tango, Charlie, Mike Emeigh and Chris Dial) makes me hedge even less on Trammell over Ozzie: Fielding repl level is probably a little bit ABOVE average. To quote:

Tom: "The bigger question is to ask how can BOTH the replacement level on fielding and hitting be set to 20 runs below average. After all, a replacement level 2b who can't field, is probably a little below average hitter. And a replacement level 2b who can't hit, is probably a little above average fielder. But to be both -20 runs from average as a fielder *and* a hitter? We're talking about a bad triple-A player here."

Chris: "Tom's dead-on. Replacement level for a fielder is above average. That is the starters in a given league field worse than their colletive replacements.

League average fielding (for ZR) is higher than the collective fielding of the starters."

Because of this, I have bigger issues with the Prospectus (Davenport) fielding numbers than I did when I wrote the article, and I'd say Ozzie's is probably considerably overrated by FRAR.

Another late bit of info from a discussion with some other Primer folks (Tango, Charlie, Mike Emeigh and Chris Dial) makes me hedge even less on Trammell over Ozzie: Fielding repl level is probably a little bit ABOVE average. To quote:

Tom: "The bigger question is to ask how can BOTH the replacement level on fielding and hitting be set to 20 runs below average. After all, a replacement level 2b who can't field, is probably a little below average hitter. And a replacement level 2b who can't hit, is probably a little above average fielder. But to be both -20 runs from average as a fielder *and* a hitter? We're talking about a bad triple-A player here."

Chris: "Tom's dead-on. Replacement level for a fielder is above average. That is the starters in a given league field worse than their colletive replacements.

League average fielding (for ZR) is higher than the collective fielding of the starters."

Joe, could you clarify what this means? I am having a hard time figuring out what Tom and Chris are saying.

Templeton, Concepcion, Bowa, Ozzie, Trammell, Thon, Yount, Ripken... they ushered in an era where shortstops were more than just leather....

How did Bowa get mixed up in this group? I remember him just being leather when he played.

I challenge you not to judge the shortstops and others from 25 years ago, based on today's standards.

You're speaking to the wrong group here, piper. The guys who live on conventional stats, the media, are the ones who are leaving Trammell off the ballot. On the STAT Internet Hall of Fame ballot, Trammell had 43% of the vote, compared to only 14% for the BBWAA.

Let's set the standard high. Only great players belong in the Hall. Let's define greatness. For the purpose of brevity let's say greatness is defined as "the cream rises to the top." Playoffs and WS are where stars shine and great players step it up.

In 3 postseason series Trammell hit .333/.404./.588 (including a .450 average in his only World Series.) He did what great players are supposed to do - lead teams to championships and for this reason I vote him in.

For me, postseason performance is much more relevant than All-Star games. I would replace #12 with this question: In pennant races, playoffs and World Series, did the player perform beyond expectations? All-Star games are really a popularity contest so it is not germane to evaluating a candidate for the HOF.

Perhaps this is a tangent but I think the field of sabrmetrics and Win Shares in particular misses a key component - quality of competition. That is, why not weight a players stats by the level of competition he faces? For example, if Davey Lopes faced Carlton 60 times, but Morgan had only 30 AB against him would not we give Lopes credit for the higher quality pitching he faced? Does Win Shares equalize competition this way?

I think sabrmetrics will have to deal with the quality of competition issue soon because expansion has diluted talent. We should figure out which hitters are getting their hits off mediocre pitchers - pitchers who would not have been #4 or #5 starters prior to expansion - and discount their career statistics. We should figure out which players would not be playing today were it not for expansion. Rafael Palmeiro comes to mind. He is a great hitter but is he still in baseball because there are more teams and therefore more jobs to go around? His counting stats are piling up but if we decide he benefitted from expansion how do we adjust his career statistics?

Sabrmetricians should try to figure out which 10-year period had the highest level of talent and give credit to those who played in that era. From best player to worst player. The opposite of dilution. Concentration of talent. Which era had the most Hall of Famers playing? Which era had players shoved out because there was someone better - not because they were too old to hit.

I would enjoy your thoughts on this issue.

Ain't been on this site too much, eh? Baseball's drawing from a larger talent pool than ever, by a much larger factor than the increase in Major League teams

I'd like to expand this idea a little further from Jonathan's excellent post. Any time there has been an expansion year in baseball, the standard deviation between the best and worst hitters has increased. There is no doubt that there were more inferior hitters in 1961, 1962, 1969, 1977, 1993, and 1998 than the years before. However, within a few years, the standard deviation reverts back to its pre-expansion levels. I have no doubt that there are more fine hitters than there ever have been in the majors now.

I'm less confident with expansion pitching though. Because of the higher attrition levels of pitchers, there is less available talent to acquire.

Does anybody know of a study highlighting pre- and post-expansion pitching using standard deviation?

Did Ozzie have measurable ability and value on offense beyond his hits, walks, and steals?

Reading Joe Marshall's post about the Cardinals offense (up nearer the top, and reprinted below) reminded me of the first time I got to talk with someone from the Soviet Union. (Most surprising thing: the Estonian hadn't seen Star Wars but had seen Battlestar Gallactica!) Joe wrote about his perspective, as a--gasp!--'80's Mets fan, on the Whiteyball offense. I was a Redbirds fan in St. Louis with the same impression of Whiteyball from the other side (the good guys) of the rivalry.

Vince, enabled and continued by Ozzie and Willie, drove pitchers crazy and scored runs. Tommy Herr was part of that mix, and Jack Clark was the big stick. Joe raises a fascinating issue about stuff that happened beyond the H, BB, SB, & CS that show up in the players' stats. Were other events happening at unusual rates when these guys were on base: balks, wild pitches, passed balls, errors on the catcher and pitcher and maybe infielders, extra bases successfully taken on hits...

No doubt these things stuck in our minds when they _did_ happen, but Joe and I--from opposing perspectives--sure thought they were happening to a mad degree.

I'd love it if one of our skillful Retrosheet analyzers could look into these issues of the Cards offense. If you're a bit ambivalent about how much fun walks, whiffs, station-to-station, and three-run-homers are, that's extra motivation for you. =)

I'd focus on rates of events when Vince was on base, as he was _the_ extreme defense-distractor if such a thing measurably existed. But the same stuff for Ozzie and Wille would be good. And sorta conversely, what was the effect for and of the others (Ozzie, Willie, Tommy, maybe Jack) when a distractor (or simply anyone) was on base? Did the distractors make it easier or harder for the next guy to succeed at the plate? Absolutely, and relative to an average of base-stealers? Joe notes Ozzie's remarkable control of the bat and the strike zone, seemingly adding to and exploiting pitchers' frustrations with a speed demon on base. Willie also was a low K hitter with terrific speed; he was high-average and low walks.

How did these qualities intertwine to make stuff happen? Did they?
I'd look at '87, or '85-'87, or possibly '85-'89.

Joe's post:
"One comment about Ozzie, and the Cards of the mid to late 80's in general: they were such a unique (and effective) team offensively, I think it's difficult to fully capture the impact of Ozzie in the lineup. Ozzie's ability to take pitches without striking out (I think he led the league a couple of years in fewest strideouts), and then put the ball in Play ON THE GROUND contributed mightily to Vince Coleman's huge runs scored total. I remember Vince was balked over something like 22 times one year, and God knows how many wild pitches and passed balls were caused by the speed of Ozzie, Vince, and Willie McGee. I was a Mets fan during those years, and that team drove me crazy, though I admired them, and wish we could see more baseball like that again."

Another late bit of info from a discussion with some other Primer folks (Tango, Charlie, Mike Emeigh and Chris Dial) makes me hedge even less on Trammell over Ozzie: Fielding repl level is probably a little bit ABOVE average. To quote:

Tom: "The bigger question is to ask how can BOTH the replacement level on fielding and hitting be set to 20 runs below average. After all, a replacement level 2b who can't field, is probably a little below average hitter. And a replacement level 2b who can't hit, is probably a little above average fielder. But to be both -20 runs from average as a fielder *and* a hitter? We're talking about a bad triple-A player here."

Chris: "Tom's dead-on. Replacement level for a fielder is above average. That is the starters in a given league field worse than their colletive replacements.

League average fielding (for ZR) is higher than the collective fielding of the starters."

Joe, could you clarify what this means? I am having a hard time figuring out what Tom and Chris are saying.

Not to put words in Joe's, Tango's or Chris' mouth, but I think what they're getting at is how to accurately determine replacement level.

There are two types of replacement players (non-starters): defensive replacements (good glove, no-hit types), or pinch-hitters (good offensively, passable at best with the glove). Generally anyone who's really good with the bat will eventually find a starter's role, so I would assume most of the bench players would fit into the first category.

So non-starters, as a group, tend to be below average offensively, but average, or better (I think tango claimed 1-2 runs below for the 2001 season over on fanhome) on defense. So the replacement player baseline should be that, rather than the -20 offensively, -20 defensively that BP seems to use. This leads to overrating good defensive players by the BP Fielding Runs method.

Maybe we should have a separate thread on this topic alone. I don't buy the argument that the defensive replacement level should be higher due to the existence of many backup defensive wizards. These guys would almost never be the regular for the team since they are such lousy hitters.

Let's take the argument to the extreme and see where it goes. Suppose that there is a class of ballplayers who are defensive wizards, the equivalent of Ozzie Smith, Bill Mazeroski or anyone else you care to name. However, they are weak hitters, in the extreme suppose they can't even put the ball in play (the classic automatic out). Suppose further that each major league team keeps one or two of these type guys on the bench for late inning defensive replacements. The argument given above would suggest that these players would lead the defensive replacement level to be quite high (higher than the league average).

But that's just not right. A replacement level is supposed to reflect what quality of player a major league team would be forced to field if it didn't have its regulars. In the scenario above, if the team's regular shortstop, say, was injured, the team would not play the automatic-out defensive wizard in his place (certainly not for more than a few games), but would be forced to go out and acquire a new shortstop from somewhere.

I guess the problem I am having is in the splitting of the offensive vs defensive replacement level quality. It just doesn't work that way. And the bizarre situation described above would not arise if a composite replacement level quality were adhered to.

Comments encouraged!

I guess the problem I am having is in the splitting of the offensive vs defensive replacement level quality. It just doesn't work that way. And the bizarre situation described above would not arise if a composite replacement level quality were adhered to.

I think an important piece of the replacement level question is how teams behave when a starter goes down. Do they bring up a AAA player who's more well rounded than the defensive specialist (who fills a need when the starters in the game) or do they let the specialist play and have that new AAA players as the backup?

Let's assume that the defensive specialist is -30 runs offensively and +10 defensively, and the AAA player is -10 offensively and -10 defensively. Obviously they are both -20 players, but the choice of which to play affects the caluculation of the replacement player.

This seems to be an example of chaining which Tango and David Smyth, among others, have discussed quite a bit on Fanhome.

I don't think it's necessarily as simple as I portrayed it in my previous post, but I do think that setting the replacement player as far below average both offensively and defensively is the wrong way to go - since if you're that bad at both, you're almost certainly not a major leaguer.

In the 1988 Abstract, James comments on the uniqueness of the 1987 Cardinals offense. One of the more interesting findings was the Cardinals good record in low-scoring games, and the lack of games with 0 runs, both phenomenons that James attributed to the extreme amount of one-run strategies.

I agree, it is silly to partition replacement levels by facet of play, and then combine it. The offense/defense thing is intertwined. The league replacement level should be set at about 75% or 80% of league average. And then you apply a historical positional adjustment. See my note on fanhome (link posted above) for more details.

Let's assume that the defensive specialist is -30 runs offensively and +10 defensively, and the AAA player is -10 offensively and -10 defensively. Obviously they are both -20 players, but the choice of which to play affects the caluculation of the replacement player.

I think that (setting aside arbitration clocks or ego issues) the obvious thing to do in that situation is to play the AAA player and continue using the defensive specialist as his defensive replacement. Jonathan's description of Lou Pinella's use of Charles Gipson indicates that Lou "gets it."

NL "ABB" Leaders 1985-1987 Vince Coleman* 39 Willie McGee* 18 Tony Gwynn 14 Ozzie Smith* 14 Chili Davis 11 Eric Davis 10 Von Hayes 10 Tony Pena 9 Kal Daniels 8 Tom Herr* 8 Steve Sax 8

I don't have the time to add everyone's steal totals, but I bet there's a strong correlation between steals and "Advanced By Balk"s. The first non-Card on the list is Tony Gwynn and he stole 107 bases over those three years (mostly 1987). I suspect we've been ignoring some added benefits of high SB totals here. Where are ABB numbers available? I wonder if the added benefit would be as strong today or if teams have compensated for opposition running games enough that they're less likely to overcompensate with balks. Hmm....

I agree with Tim DeWalt's conclusion that defensive replacement level is relatively higher at "defense-first" positions like SS than, for instance, 1B. Intuitively, it just strikes me as true.

If I were re-writing this article, I'd re-analyze the FRAR aspect. FRAR are NOT a valid metric to use, IMO. I would use FRAA, as replacement fielders are basically average, see Tango's work for more on this.

FRAA (2003 version) has Ozzie at 252, Trammell at 62. That would lower my guesstimate on Ozzie's defensive edge, and definitely move Trammell from basically even to 'slightly ahead' of Ozzie overall.

Bob, you're going to have to give me more to change my mind on Concepcion. Or point out exactly where in my analysis I may have missed something.