Why Not Lou?

December 28, 2017

A good friend of Stat of the Week, ESPN Chicago radio host Mike Murphy, presented us with a question as to why Lou Whitaker has not been elected to the Hall of Fame when he had better numbers than his recently elected Tigers’ teammate, Alan Trammell. Lou Whitaker hit .276/.363/.426 with 2,369 hits and 244 home runs in 9,967 plate appearances. Trammell hit .285/.352/.415 with 2,365 hits and 185 home runs in 9,376 plate appearances. Whitaker won rookie of the year, was a five-time All-Star, and won three Gold Gloves and four Silver Sluggers. Trammell was a six-time All-Star and won four Gold Gloves and three Silver Sluggers. He was also named World Series MVP in 1984.

Bill James presented his reconfigured Hall of Fame Monitor in The Bill James Handbook 2018 that was released this past November. The system tracks accomplishments that are more common among Hall of Fame players than among non-Hall of Fame players. It doesn’t determine who should be in the Hall of Fame, but rather how each player’s career stacks up when looking at what Hall of Fame voters may consider to be a Hall of Fame career. By using this system, Whitaker has 65 points and Trammell has 70. Generally, if a player has more than 100 points in the Hall of Fame Monitor by the end of his career, he likely will be selected for the Hall of Fame. If he has less than 100, he likely will not. James defines the range of 70 to 130 as a gray area. By this standard, we see that Trammell was at the lowest point of the gray area and Whitaker was just below. Teammate Jack Morris's recent selection was predicted by the Hall of Fame Monitor with a score of 112.

So, why might it be that Trammell squeaked by in the eyes of the voters whereas Whitaker did not? In looking at the components of the Hall of Fame Monitor calculations, we found that the biggest discrepancy between the two players was in the points awarded for batting average. In seasons with at least 400 plate appearances, Alan Trammell had seven seasons with a batting average greater than or equal to .300, whereas Whitaker achieved this feat only once. Even though their career batting averages were close—only .009 apart—maybe voters were looking more closely at individual season totals.

Here are some other players with similar Hall of Fame Monitor scores to Lou Whitaker who have been inducted into the Hall of Fame:

Player Position HOF Induction Year HOF Monitor Score
Larry Doby* CF 1998 69
Lou Boudreau SS 1970 69
Phil Rizzuto* SS 1994 68
Edd Roush* CF 1962 67
Pee Wee Reese* SS,3B 1984 66
Lloyd Waner* CF 1967 65
Ross Youngs* RF 1972 64
Gabby Hartnett C 1955 63
Bill Mazeroski* 2B 2001 60


*Note: a player with an asterisk next to his name was voted in by the Veteran's Committee.

We can see fellow second baseman, Bill Mazeroski, was inducted into the Hall of Fame in 2001 with five fewer points than Whitaker. His batting line was worse than that of Whitaker, .260/.299/.367. However, he won eight Gold Gloves and was a ten-time All-Star.


COMMENTS (31 Comments, most recent shown first)

I'm adding my voice to this late, but I hope I can assist. In regards to positional adjustments for bWAR and fWAR. They are not made based on anything to do with the offensive or defensive performance of other second basemen in the league. WAR is determined based on a theoretical "replacement player" that all MLB teams could theoretically place in the lineup. bWAR answers the question of how much more valuable "Lou Whitaker" (or any other player) would be over the theoretical replacement player. That replacement player, by the way isn't an "average" player. He's below average, since even an average MLB player is difficult to find. However, MLB organizations have replacement level players stocked in their minor league organizations.

Hope this helps.

I grew up in Michigan and watched probably 1,900 of those games that Sweet Lou and Tram played in the same lineup. I don't think one was much more valuable than the other. Having Trammell in the Hall with Lou out doesn't compute to Tiger fans. They really were twins in a sense, as Bill James once pointed out in one of his early 1980s Abstracts.
3:08 AM Feb 26th
Just BTW: Tom hasn't been inconsistent about what name he uses.
He's always Tom Tango in terms of authorship and public persona, and he's always tangotiger in his posts and comments on here -- just like all of us are whatever-is-our-actual-or-public-name in public, and whatever is our user name on here.
8:11 PM Jan 2nd
There is a selection problem. Whether it changes the results, we don't know.

Tangotiger (or Tom Tango -- he's been inconsistent over the years regarding what pseudonym he uses) is aware of the problem from the first link I provided. My recollection is that he tried to address it, but I searched high and low for it and couldn't find it.

What can be done to better resolve the issue though? We don't have data on how everyday players (other than the occasional Ben Zobrist and Tony Phillips) perform when they are asked to play other positions. Also, would it matter a lot? Tangotiger has noted in a couple places (including in comments on a Bill James Online article) that his positional adjustments are similar to what others come up with.

So I'm agreeing that your point is valid, but don't see a better way to compute the positional adjustments that are needed to estimate a player's value.

My original point still stands though, that (to the best of my knowledge) WAR's evaluation of Lou Whitaker's value doesn't vary depending on whether other AL second baseman were especially strong or weak hitters during his career.
12:28 PM Jan 2nd
Re that last thing:
I don't know that there's such a "selection bias." I suspect there is, but don't know, and I don't pretend to know how big it would be.

But the thing is -- and the reason the objection would seem to be valid -- is that if the system is based on what you said, the system assumes there isn't any such 'bias.'

BTW, to be clear about it: The "selection bias" I'm talking about is that the difference between a given utility-type player's fielding at different positions might not at all be representative of the the actual difference in difficulty or importance of the positions per the fielding of regulars at the positions.
9:53 AM Jan 2nd
I looked and couldn't find it on Tangotiger's old blog.

Somewhat related is www.insidethebook.com/ee/index.php/site/article/replacement_level_fielding/ WAR assumes that the difference between an average and a replacement level player is all in the batting, that a replacement level player fields on average as well as an average MLB player does. The above link documents that that assumption is close to reality.

I believe that's related because using data from players who switch positions to calculate positional adjustments might not lead to as much selection bias as we think.
8:18 AM Jan 2nd
Maris: Your objection is a valid one, although you are (of course) not the first to raise it. Tangotiger anticipates it in the first link that I provided for example. I'm sure it's been addressed (whether it's been done so adequately is of course something else).
8:02 AM Jan 2nd
My understanding of Tango's research was that it indicated simply that shortstop is harder to play than second base. I don't think that was much of a surprise, given the longer throw (which is not entirely evened out by the easier pivot).
7:29 AM Jan 2nd
Geez -- maybe 'yes' after all.

MWeddell, I hope you'll clarify. Upon further look (and some dissection), it looks like you really did mean what I thought in the first place.....
11:38 PM Jan 1st
No no.....I think I might have to say, as the announcers do, "Check that."

MWeddell: I'm wondering if I just mistook altogether what you meant by talking about players who divide their time between different positions.

I thought you meant it's what Tom has said the adjustments are based on -- but, I'm wondering if that's not it at all, and you were just saying it as an example.

If so: Forget my objection. It wouldn't apply.
10:35 PM Jan 1st
Funny -- My impression of the general take on "WAR" that we find here is the opposite -- i.e. very highly favorable -- although I think it has been somewhat moderating over perhaps the last year or so.

I would like to take some credit for the latter. :-)

And, I'd better add, even I have a quite favorable take on "WAR" and the related metrics, like WAA.
My main problem with it is what many people seem to do with it, i.e. making more of it than is merited.


Thanks for what you added about your understanding of what that thing is about.
It's essentially what mine was, although "operationalizing" it -- i.e. being more specific about exactly how it is derived -- would be harder than what either one of us said.

Although actually :-) .....maybe this thing I said is pretty close:

looking at the players' runs-saved (or some such) in comparison to all players at the position, and then comparing those players' own runs-saved-per-inning (or some such) at each position with their own runs-saved-per-inning (or some such) at the other positions they played


And, assuming it's something like what either of us said, here's my problem with it:

It seems hardly valid, and frankly a little nuts, to base a bedrock aspect of such an important metric on the play of non-regulars at the positions -- especially since those players not only aren't regulars but also are switching between different positions.

I wouldn't at all assume that the relative quality of play at the different positions by (let's call them, for the purpose of this) shuttling journeymen is indicative of a basic thing about the respective positions or about the players who play the positions regularly.
10:17 PM Jan 1st
My understanding is that Tango looks at runs saved (lost) versus average and each position. To illustrate (I'm making up numbers, you understand), if those players who play both 2B & SS average three runs per 150 games below average in the field as SS but two runs per 150 games above average at 2B, then (if that was the sole data point), we might conclude that an average shortstop is 5 runs per 150 games more valuable than an average second baseman. This is my high-level understanding. For example, it might not be per game. Whether I have stated it accurately enough for you to see whether your objection has merit, I don't know.

I will be mildly surprised if you generate a useful discussion of WAR on these message boards. The consensus of the message board community is decidedly anti-WAR in my perception such that members frequently slide into ad hominen arguments.
7:01 PM Jan 1st
I don't do well with his explanations (nor does he do well with my inquiries.) :-)

If you or anyone can explain it simply or directly (which IMO he tends not to), it would be welcome.

Perhaps to clarify my question: (as I said on Reader Posts)....

How does the method look at "fielding abilities of players playing multiple positions"?
Does it mean just that it looks at the players' runs-saved (or some such) in comparison to all players at the position, and then compares those players' own runs-saved-per-inning (or some such) at each position with their own runs-saved-per-inning (or some such) at the other positions they played?

That's my guess, but I don't know, and if we want to judge it, we need to know clearly.
3:57 PM Jan 1st
Thanks for that last comment.
I didn't know this:

The position adjustments used in both fWAR and b-refWAR are the same. They are from research a decade ago by Tangotiger on the fielding abilities of players playing multiple positions. For example, players who play both 2B and SS tend to be more a few runs less (per 150 games) at SS than they are at 2B.

.....and in fact I don't understand what you're saying, maybe in part because of the little typo (the word "more") but I think I wouldn't understand it anyway. You're saying great stuff but I think not clearly. It looks like one of those things that if you already know it, it's clear, but if you don't, it isn't. I think I might be close to getting it, but it's important to know it clearly in order to be able to to know what we think of it.

The reason I'm particularly interested is that it smells like a thing that if I understood it, I'd be able to point out a clear flaw in it.

Of course I'd welcome clarifications from anyone else too. I'm going to post this in Reader Posts as well, because I'm not sure how many people are still looking at this article.
10:31 AM Jan 1st
The position adjustments used in both fWAR and b-refWAR are the same. They are from research a decade ago by Tangotiger on the fielding abilities of players playing multiple positions. For example, players who play both 2B and SS tend to be more a few runs less (per 150 games) at SS than they are at 2B. So, yes, the fielding adjustments are based on data, but they aren't based on batting data. They probably should be modified somewhat as one goes back in time, but they aren't.

My point is, to the best of my knowledge about WAR, that you've asserted a flaw in WAR that isn't there. Even if Lou Whitaker played during a time when his fellow second basemen were unusually poor batters, that wouldn't affect his WAR calculations (except to the extent that it affects the average of all batters, not just second basemen).
9:06 AM Dec 30th
They have to be something, Weds. Either they are random, or they are tied to data. I think they are tied to data, but since it's not a one-year datapoint maybe that makes you think it's not?

Since most of these guys played for over a decade, they have to be represented fairly heavily compared to some guy who only played 20-40 games and wasn't any good, which is kind of my point.

My point is that there were an unusual number of second basemen playing hundreds of games, rather than dozens of games, at a level that really needed to be replaced. Even the false normalization of defensive statistics can't make up for all those sub-.650 ops seasons.
1:54 AM Dec 30th
I didn't say the positional adjustments were random or arbitrary, nor did I say that they weren't tied to data.
5:34 PM Dec 29th
I didn't design either b-refWAR or fWAR, so don't blame or credit me with any of it. The positional adjustment is constant because it is based on estimated fielding value differences. It isn't computed based on the quality of the batters at that position.
5:33 PM Dec 29th
Weddell, that doesn't make any sense. Are you saying position adjustments are just random, or arbitrary? Not tied to data? Because if they are tied to data, they are tied to Wilfong, Kuiper and all those guys. If they are not tied to data, what the Lee Lacy are they tied to?
4:51 PM Dec 29th
The position adjustment doesn't change based on whether other second basemen or uncommonly strong or weak batters though.
2:23 PM Dec 29th
It's in the position adjustment.
1:00 PM Dec 29th
To my knowledge, having weak batters overall would increase Whitaker's WAR, but having weak second basemen specifically wouldn't increase Whitaker's WAR. The WAR calculations aren't first assigning wins to second baseman and then apportioning that fixed pool of wins among the second baseman.
11:55 AM Dec 29th
I expect Whitaker to get his plaque within a decade, rendering the argument moot, but it's an interesting question. I'm honestly surprised they don't both do better on a Hall of Fame prediction metric.

Because of something else I did - the Test - I found some subtle differences in their respective comparison groups that won't show up in most metrics. For some reason, half the league's managers were still playing guys like Duane Kuiper and Rob Wilfong at second base, so WAR makes good players like Grich and Whitaker look like superstars. I'm not saying they weren't really good, mind you, but their values were exaggerated a bit.

Trammell's competition wasn't quite as bad, mostly because the weaker hitting shortstops could field. Lots of those supposed good secondbasemen were just as bad in the field as they were atbat. I think it was the old "second base, second hitter" syndrome. The older managers wanted a little-ball player at second base.
11:49 AM Dec 29th

I agree, but Whitaker was a slightly better hitter most of his career, ending up with a higher on base and slugging percentage. I think they are essentially interchangeable as players in regards to value.

I grew up in Baltimore and watched Bobby Grich his whole career, Whitaker was that type of player. If you didn't watch him closely you didn't realize how valuable he really was.
10:50 AM Dec 29th
I admit to being a Trammell fan and always have thought he should have been the MVP in 1987. I would just add the argument that shortstop is a more difficult position than second base. Therefore, Trammel was slightly more valuable to his team than Whitaker in the field, everyting else being equal.
9:29 AM Dec 29th
.....on second thought (meaning a second later)....

No, I see that it's not about stronger or weaker, just that it seemed like I meant a different kind of thing.
And indeed I did a sloppy usage.

What I meant was guideline for us in how we see the prospects, not guideline for the voters -- which is exactly what you're saying too.
9:26 AM Dec 29th
Steve: But that's exactly what I meant by "guideline"!
It's exactly why I chose that word (yes, it was a thoughtful choice). :-)

I guess it seems like a stronger word to you than to me....
9:22 AM Dec 29th
Mazeroski is not a good comparison. He was almost certainly elected because he is perceived to be (and very possibly is) the greatest defensive player in modern baseball history, and hitting one of the most famous homers ever didn't hurt. I would have voted for Maz, and so would have Bill James in 1985 (but he had changed his mind by 2017).

I expect Whitaker to get in eventually. I would have voted for both of them, but not Morris (although I don't think his induction hurts the Hall).
9:16 AM Dec 29th
For those of you who don't mind looking at WAR numbers from Fangraphs, this graph illustrates how Trammell's best seasons were better than Whitaker's best seasons even though Whitaker ended up with slightly higher career value: www.fangraphs.com/graphsw.aspx?players=1013157,1013846

Whitaker's graph is very flat, with his 15th best season essentially the same as his 8th best season.

Whitaker added power (and Sparky Anderson sat him against tough lefties) as he aged, roughly offsetting his loss of range in the field. In contrast, Trammell was essentially done at age 34.
8:47 AM Dec 29th
Maris, I'm sure you're aware of this, though your use of the word 'guideline' raises doubts: Bill has told us any number of times that the Monitor is intended to be predictive, not prescriptive.

That said, instead of asking "Why not Lou?" one might just as easily ask "Why Tram?"

For the record, Trammell's election doesn't seem offensive (unlike, say, Bowie Kuhn's), but it doesn't seem necessary, either. I feel the same way about Morris.
7:55 AM Dec 29th
First, there's a thing right in your article that goes a little way (indirectly) toward explaining it, but you're putting it seemingly as though it argues the other way:

.......[Bill's] reconfigured Hall of Fame Monitor.....Whitaker has 65 points and Trammell has 70....Trammell [is] at the lowest point of the gray area and Whitaker [is] just below....
So, why might it be that Trammell squeaked by in the eyes of the voters whereas Whitaker did not?

Yes, the rest of what you say (i.e. what is represented by the initial "......") does argue for Whitaker -- but this part argues against him. Granting that the system is just a guideline (but a good one), if a player is in the gray area, he's in the game; if he's not even in the gray area, he isn't.

More concretely I'd say a key difference between them is that Trammell's best seasons were better than Whitaker's best seasons, whether on "WAR"-type stuff or Win Shares, although not hugely so, and although to a great extent it's sort of just one year of Trammell's that makes the difference. (BTW Whitaker shows better on career value by both methods.)
Similarly, Trammell did much better in MVP voting.
So, there are some objective rational explanations although we could debate whether they're large enough to account for the huge difference in how they've been considered for the Hall of Fame.
As probably most of us know, race is sometimes raised as a factor, and while I do see the objective differences as being enough to possibly explain the disparity, I do suspect some outside factor going on, whether it's that or something else, because despite the noted things in favor of Trammell, to the main extent I see them as a unit, almost like Tinker-Evers-Chance, and I would have expected them to get in together or not at all.

BTW, on baseball-reference.com, Trammell is Whitaker's 2nd most similar player (albeit by a fair margin, behind Sandberg), and Whitaker is Trammell's 4th most similar.​
7:55 PM Dec 28th
©2019 Be Jolly, Inc. All Rights Reserved.|Web site design and development by Americaneagle.com|Terms & Conditions|Privacy Policy