Don't call him Doody.

Better than Jim Rice, I don't understand why Rice is on the cusp of the HoF and Evans couldn't get 5%.

I don't understand why Rice is on the cusp of the HoF and Evans couldn't get 5%.

1978

I don't understand why Rice is on the cusp of the HoF and Evans couldn't get 5%.

R

B

I

70 RBI is the difference between HoF and 5%?! Hits and HR are pretty much even also but Dewey has Rice beat in OBP by 20 points.

First impressions are probably a big factor; Evans took a while to reach a really high level, while Rice was a major star as a rookie.

Someone make the case for Dwight Evans as a HOMer showing why is he a better case than
a) the top candidates at his own position
b) the top candidates at LF and 1B
c) the top candidates at C, 3B, 2B (the less-well represented positions).

I should start by saying that I think Dwight Evans was better than Rice.

But its not hard at all to figure out why Rice is getting support and Evans is not. Very, very easy. And its not just RBIs.

On the HOF monitor, where a score of 100 is borderline, Rice has a score of 147 (clear Hall of famer) and Evans is at 70 (most likely not a hall of famer).

Rice was an all star eight times. Evans was an all star three times. Evans hit, in various years, 46, 39, 39 and 39 homeruns. He regularly hit over .300, and had a career average of .298. Evans never had single season homerun totals like this, and was generally a .270 hitter, not a .300 hitter.

Rice won an MVP. Evans did not. Mainstream media, at the time they both played, were far more impressed by Rice than by Evans.

All in all, I'd be shocked if Evans received anywhere near as much support as Rice.

I agree with Eric, I don't understand the early love for Evans. If I have time I will try and post a comparison arguing AGAINST Evans for induction (item #1, a low peak) compared to other corner outfielders. Right now I am not sure why I wouldn't place Parker above Dewey.

Dan R,

How does your system seem Evans? In looking at WARP and WS, I am just not very impressed. Are they missing something that your system is possibly picking up?

I'm mystified by the lack of support for Evans, and, if true, the low WS/WARP ratings. He was a gold glove outfielder who was consistently in the top ten in adjusted OPS during the 1980s, and also had a relatively long and productive career. Is it possible that WS and WARP are not giving him enough credit for his defense? He also played during a time when it was harder to dominate than almost any other era - is this taken into account? One of his best years was during the strike - how is this handled? He was consistent and he was durable. I'd be suprised if you could name more than a couple of backloggers who are clearly more qualified.

Right now I am not sure why I wouldn't place Parker above Dewey.

If either Win shares or WARP has Parker as better than Dewey, than something went very, very wrong with the statistics.

IIRC, one of Evans biggest years was 1981. The strike didn't help his cause. And I just noiticed that Dizzypaco mentioned this.

The Hriniak/Lau approach to hitting seems to have gone out of vogue, but Evans was probably Hriniak's best disciple.

He was a gold glove outfielder who was consistently in the top ten in adjusted OPS during the 1980s, and also had a relatively long and productive career.

Making a little tweak here...

He was a gold glove outfielder who was consistently in the top ten in adjusted OPS during the 1910s-1920s, and also had a relatively long and productive career.

Harry Hooper.

Evans was top 10 in OPS+ 6 times and Hooper 4. You can say the same thing about...Dave Parker

He was a gold glove outfielder who was consistently in the top ten in adjusted OPS during the 1975-1985, and also had a relatively long and productive career.

Parker was a five-time OPS+ leader in this period. And he won three GG, and he was an MVP.

You can say it about Mickey Vernon, too, though he didn't win any GG since they came along in his twilight. This is just among current unelected players at the corners over 9500 PAs. Beckley might fit this mold, though his defense isn't reputed to top-notch. On the other hand plenty of corner-fielding HOMers likely also fit the same description, and, yes, Evans's performance is better than the others I've cited. But I don't think it's a strong argument by itself for induction.

My all-time favorite Sock. Work the count into your favor, look for your pitch, then swing really hard at it. doo-WEE!

I don't think he belongs in either Hall.

Remember the Bob Newhart show where one guy always introduced his brother Darrell and his other brother Darell? :)

Dwight Evans is scarily like the other brother Darrell. Good D, low BA, hi walks, hi power, long career, last name, first initial, same era... and right about overll as good, above my in/out line, good enough to be a #1 in a backlog year.

Just using WARP3 for a minute, comparing Evans to other OF/1Bmen:
career WARP3 yrs (tossing out seasons with WARP 2 or less)
Evans... 119 18
Johnson 100 13
Smith..... 91 15
Bonds.... 91 12
(Wynn.... 90 12)
Beckley. 107 18
Perez..... 99 14

So, Dwight amasses about 23 more wins than most, in about 4 seasons. That is well above most anyone's replacement level. To me Dwight is the highest 1b/of candidate.

Harry Hooper.

Harry Hooper's career OPS+ was 114. He surpassed 125 three times in his entire career. He never finished in the top five in OPS+, despite playing in a league that was segregated, with less teams and easier to dominate. Evans' career OPS was 127. He finished in the top five in OPS four times, and was valuable in his other years as well. There's no contest whatsoever.

Parker was a five-time OPS+ leader in this period.

Parker was as good as Dewey in his prime. But otherwise, Parker didn't have a lot of value - a fat, bad outfielder with an OPS around 100 and a lot of time on the DL, not to mention a clubhouse problem. Evans outside of his prime was a very useful player, both offensively and defensively, as well as being durable and consistent.

You can say it about Mickey Vernon, too

Have you looked at Vernon's career recently? He had a few very good years, a couple of good ones, and was otherwise pretty worthless.

Doc, you are usually pretty rational in your observations of players, but it seems like you are going out of your way to make Evans' career appear far worse than it was. Are you sure you are looking at the right player?

Dewey was one of my favorite players, which is odd because I hated the Red Sox. I remember him having a very odd batting stance, where his front foot actually pointed slightly towards his back foot. I also him having a very good arm for an older player. Very underrated player. I think his lack of All-Star appearances hurts him because most people probably think of him as a good, not great player.

I had forgotten about the strike, that will change things some. On the other hand, using WARP3 to compare Evans with Beckley and Bob JOhnson is highly dubious because of its HUGE timeline. In some ways, the same could be said of players like Bonds and Smith.

I just looked at 1981 again. Evans is given 26 WS, he played 108 games but I am not sure what the average team played in 1981. If I give him 9 more WS, he has 35. That makes his top five seasons 35,31,29,25,24 with 356 in his career. Parker is at 37,33,31,29,25 with 327 in his career. In top five seasons, Parker has Evans licked and they are only 29 off for their careers. As a peak guy, I see little reason (by this measure) to put Evans above Parker.

Now WARP isn't terribly interested in Dave Parker or Dwight Evans and Evans would come out slightly ahead there. But If I were to use WARP as my only measure I am not sure that either would be in my top 40. As it stands, right now I will probably have both in my top 30-35 but not in my top 20. I dont' see either are really good candidates for induction.

My argument isn't that we should be voting for Parker instead but that we should temper our excitement on Dewey.

Is it possible that WS and WARP are not giving him enough credit for his defense? He also played during a time when it was harder to dominate than almost any other era - is this taken into account? One of his best years was during the strike - how is this handled?

In terms of WARP, Evans is credited with 60 FRAA, concentrated in the early part of his career. This is consistent with the perception that he was a very good defensive player. His 1981 season is easily his best season according to WARP3, which accounts for shortened seasons by multiplying WARP2 x (1 + (2/3 x fraction of games missed)). Thus, his 8.9 WARP2 becomes 11.9 WARP3. If you believe that using the straight-line method would be appropriate, you can add 1.5 WARP to the total. However, he is afforded substantial strike credit.

My biggest concern with WARP relates to your second question. I think that its treatment of league difficulty, or its adjustment from WARP1 to WARP2, is very flawed. BPro doesn't say how this adjustment is made, but the numbers are questionable. Evans played the equivalent of about 16 full seasons, with nearly the entirety of his value between 1974 and 1989. His WARP adjustment from 105.1 to 115.9 is 10.3%, or .68 per full season. Similarly, Albert Belle was a corner outfielder in a DH-league. Belle played the equivalent of 9.5 full seasons, with nearly the entirety of his value between 1993 and 1999. His WARP1 to WARP2 adjustment from 69.1 to 86.1 is 24.6%, or 1.79 per full season.

According to WARP, the league difficulty gain for Belle is roughly 2.5 times as large as that for Evans. Based on everything I have seen, the era from the mid-70s through the 80s in which Evans played was the most difficult in MLB history. Belle's excellent production in the mid- to late 90s coincided with two league expansions and no major increase in the talent pool, which should indicate an easier to dominate league. Whether you agree with this assessment or not, I wonder how anyone can justify the WARP adjustments which appear to show that Belle's league was significantly tougher than that of Evans. I believe that it will become clear that WARP systematically underrates those who peaked in the 70s and 80s as this project continues.

I just looked at 1981 again. Evans is given 26 WS, he played 108 games but I am not sure what the average team played in 1981. If I give him 9 more WS, he has 35.

The average team played 108 games. Evans played in every game for the Sox. Why would you only give him 9 more win shares?

Two points on Evans' case:

1) In comparing Parker to Evans using win shares or WARP1, adjusting for the DH is crucial. Without an adjustment, both systems will underrate Dewey relative to the Cobra.

2) I am a strong supporter of Evans, but I recognize that his best case is his career case, because his batting and his fielding peaks did not coincide. He was a great defensive outfielder from the time he entered the league through 1981. He became an excellent hitter in the early 1980s, so he was a good, not great player for the first 6-7 years of his career when he was a great fielder and an ok hitter, he was a great player for1 season (1981), which was diminished by the strike, and then he was a very good player for another 7+ years, because he was an excellent hitter, but his defensive value tailed off, especially when he was spitting time between outfield and first base (he was not a good defensive first baseman) or between outfield and DH.

Now WARP isn't terribly interested in Dave Parker or Dwight Evans and Evans would come out slightly ahead there. But If I were to use WARP as my only measure I am not sure that either would be in my top 40.

If you adjust WARP for the DH, it likes Evans plenty.

I can see why the peak voters are not excited about Evans, but I think Evans' case is one that shows the weaknesses of a focus only on peak, especially if one is making the claim that Dave Parker is to be prefered to Evans. Parker was truly better than an average player in only about six seasons in his career. Evans was better than average for twice that many. He only had one year (1981) that was a match for Parker's top 3, but he is better than Parker _everywhere else_ in his career.

For the Evans supporters out there, let me note that, despite the rather critical tone of this thread, I think it's highly likely that Evans is going to be elected, and soon. Most of the criticism is coming from peak-oriented voters. This is going to be a case where the career voters will be a (relatively) silent majority.

The average team played 108 games. Evans played in every game for the Sox. Why would you only give him 9 more win shares?

I don't know if 9 is the right answer, but we can't just prorate his numbers without taking into account that players regress toward the mean the longer a season is. IOW, Evans' 26 WS in 108 games doesnt necessarily translate to 39 WS.

But, as has been pointed out in this context before, 26 WS in a 108 game season contributes as much toward a pennant as does 39 in 162 game season.

But, as has been pointed out in this context before, 26 WS in a 108 game season contributes as much toward a pennant as does 39 in 162 game season.

Correct. The question isn't what Evans would have done with 54 more games, its how valuable was he in the 108 that were played. He had precisely the same amount of value as someone with 39 win shares in an 162 game season.

ANNOUNCEMENT: My system likes Dewey.

Yeah, somewhat surprising. And while the OPS+ is unimpressive, the career is really good, good enough to sustain his peak. Add the defense (WARP isn't necessarily unkind to him), and a splintered backlog, and I definitely see him elected.

IMO, he's better than the recent "rushed newloggers": Sutton, DaEvans, Hernandez.

but I recognize that his best case is his career case, because his batting and his fielding peaks did not coincide.

I'm not sure if that is a negative though... while it certainly is a negative in these "systems" in actuality on the field, it may not be... or it might not matter.

This example is exagerated, but stay with me. Say a player is an average player for his 10 years. In the first 5 years he is +10 with the glove, and -10 with the bat. That is a good player to have around, as you can leverage his glove, pinch hit for him... spot rest him, manage his usefulness based on park and pitcher ect... for the second half of his career he is the opposite +10 with the bat, -10 with the glove... you can work with that, move him down the spectrum, defensive replace him... manage his playing time optimally ect... Those types of players MIGHT have more value to a team, than the guy who is average both ways for 10 years...

The reason I think this point has merit (pun intended), is because guys with really good skill will almost always find a job somewhere.

OK, here's one of our never-ending series of OPS+ comparison charts. This time it's prime RF candidates, as well as LF and 1B. Includes all unelected guys at each position with 100+ points in last election with a couple add-ins that I wanted to know about. These are 250+ AB seasons so that the Tony Perezes will be in their best light. I don't want anyone to accuse of me of Norm Cash bias! ; )

I ordered them by career OPS+ which I will list above for space considerations.

Cravath 148*
Tiernan 138
Smith 137
Singleton 132
Bonds 130
Evans 127
Staub 124
Parker 121
*ROUGHly estimated to include MiL credit

Howard 142
Johnson 138
Burns 114

Cash 139
Cepeda 133
Beckley 125
Perez 122



A few things.

1) Why doesn't Mike Tiernan get more votes?

2) Why is the 1B support nearly inverted from this chart?

3) Let me break out Evans here. At each stop, here's his rank among the group of 8 RFs:

6 4 5 5 6 4 7 5 5 2 4 2 3 3 1 1 1 1 1

No one is surprised that his peak is good but not as good as other hitters, nor is anyone surprised he'll improve in relative standing once the better hitters retire, leaving him, Beckley, and Perez to duke it out among the trailing, long-career group. I don't see that as an immediately HOM-worth offensive record, except in that Evans did have an obp heavy OPS, and so he may gain some advantage over at least Beckley and Perez. The guys above him all had good to great OBPs, so he doesn't pick up as much on them.

4) Parker is clearly the worst RF candidate as a hitter.

5) Norm Cash stays ahead of Beckley the whole way until Cash's last 250 PA season, after which point Beckley's remaining seasons are likely at or below the positional average for 1Bs.

6) Tony Perez is not impressive. He loses it very quickly, and 'it' was good for four years but as good as, say, Frank Howard.

7) This is Reggie Smith's best light, if we up the PA threshold to, say, 400, then he'd lose a 133, a 153, and a 125. If we pushed it to 500 to reward consistent, full-time play, he'd also lose a 150 and a 116.

OK, let's make junk-stat game out of it. For each of these guys, let's:
-subtract a point for every sub-120 season
-add a point for every 120+ season
-add a point for each ten points above 120 they are in a season (so 1 point for 130, 2 points for 140, etc....)
-award small bonsues for playing time
add a .2 points for each season over 300 PA in a 120 OPS+ or higher season
add another for each 100 thereafter.

It's arbitrary, I know, and, I'm only eyeballing sked adjustments, and also Cash, Cepeda, Burns, and Cravath are at a potential disadvantage due to the PA-suppressive effects of low-run environs. But I actually don't know how it will turn out, and I'm just trying to get some kind of relative ranking of the offensive contribution of these fifteen or so players. So when we do this we get:



Right. So Evans is bottom-middle, Parker subzero, brrrrrr. Reinforcing some things I've said here and elsewhere, Perez is down the bottom too, and Cash near the top. I didn't ding Johnson for the war, by the way.

Clearly this is not definitive, just more information, but if OPS+ is a strength for Evans, it's relative to the league not to HOM-level players. Even when we cherry pick his good seasons (as in this last exercise), he doesn't leap out (he's behind Beckley after all!!!---kidding, Karl...). As noted OPS+ is not perfect, but mostly the guys above Evans also benefit from the OBP underrating effect of OPS+. And defense and baserunning aren't taking into accout either. Those may or may not be strengths for Evans or his competition, but in general corner guys aren't getting by on baserunning and defense, so if there are coarse differences, legs and gloves may not close those gaps.

If anything, what I'm getting is that none of these guys is a tremendous candidate, but that Gavy Cravath, with MiL credit, does stand above the fray on offense and may be HOMable.

But, as has been pointed out in this context before, 26 WS in a 108 game season contributes as much toward a pennant as does 39 in 162 game season.

Have we had this context before?

We've been vigilantly using this logic for 162 vs 154 and the slowly increasing schedule of the 1900s, but this specific question about 1981 has never come up has it? 1972 is a barely noticable year. 1918 is a precedent, but it was also a war year with many players out of action.

I certainly think we need to adjust up, but 50% is a big extrapolation. Mike Schmidt goes up to 45 WS.

And if we do incur some sort of regression, its only fair to also do it for players who had off years in 1981. Reggie Jackson comes to mind. I'm not sure how many non-moot examples of this are.

I'm not strongly arguing either way, but I'm just wondering if "what to do with 1981" has been fully debated here before.

And if we do incur some sort of regression, its only fair to also do it for players who had off years in 1981. Reggie Jackson comes to mind. I'm not sure how many non-moot examples of this are.

I agree, David. Regression should be at both ends.

And if we do incur some sort of regression, its only fair to also do it for players who had off years in 1981. Reggie Jackson comes to mind. I'm not sure how many non-moot examples of this are.

Oddly enough, a guy like Dave Parker.

Doc,

While I appreciate your analysis here, I don't think that George Burns belongs in this group. These guys are hitters - you wouldn't have any of them lead off like Burns, and they won't steal bases like Burns, and they couldn't play CF like Burns.

Burns (as my ballot comment has stated for the past couple of elections) should be compared to Roush or Brock or Duffy or GVH.

Dizzy,

I do not do linear projections for shortened seasons. I did teh same thing with pet candidate Charlie Keller, docking ahim a few WS from the straight line projection for 1945.

Also, why would one adjust for the DH league in WARP? Doesn't WARP3 already have a league quality adjustment? I can see doing it in WS since there are a ceratin number of hitting WS to be divied up, but since WARP is not connected to real world wins and therefore any player can earn a high amount of WARP, why adjust?

OK, let's make junk-stat game out of it.

This I would agree with.

The stat doesn't really seem arbitrary - instead, it seems designed to make Evans look as bad as possible. Ignoring the fact that it doesn't include defense (which Evans has a sizeable advantage over almost everyone on the list), you give points at 120, 130, 140, 150, and so on, while Evans had seasons at 119, 129, 149, and so on. It appears to be heavily weighted toward peak, whereas Evans beats several of these people in terms of career value. It doesn't give much credit at all for seasons between 110 and 120 (which Evans had a few, and when combined with gold glove defense is clearly a valuable season). It ignores playing time and injuries - Evans missed very few games in his best years, while, for example, Reggie Smith missed nearly 50 games in one of the seasons you have listed. And so on.

Its probably possible that you could create a junk stat that makes every candidate that's even close to borderline look especially good or especially bad.

Oddly enough, a guy like Dave Parker.

His 1980-82 OPS+ line goes 116-106-113. Plus he missed a considerable fraction of his teams games in both 1981 & 1982. If we regress (I'm still just supposing on that point), he'd get a tiny bump I suppose, but only a few points of OPS+ maybe... nothing major.

Guys with low-1981 years in 1980-82 that I can find right now:

Dale Murphy-135-100-142
Ted Simmons-142-87-112 (with a trade league-switch between 1980 & 1981)
Jack Clark-152-128-138
Reggie Jackson-172-119-147
Robin Yount-130-114-166
Lance Parrish-121-99-135
Jim Rice-123-117-131
Hal McCrae-123-110-147
Al Oliver-131-124-150
Gary Carter-126-112-146

etc...

I'm at work so I don't have time to be thorough at the moment.

Off years happen and its unfair to regress them away, and in some cases the big years in either 1980 or 1982 might have been career years, but there's likely some mis-timed slumps going on above that would regress up a bit.

Anyhow, I still haven't gone either way on this, I'm just dumping out information. :-)

Ignoring the fact that it doesn't include defense (which Evans has a sizeable advantage over almost everyone on the list), you give points at 120, 130, 140, 150, and so on, while Evans had seasons at 119, 129, 149, and so on.

That's easily fixed. Just change the formula to (OPS+ - 110)/10. That removes the effect of the 9's.

I do not do linear projections for shortened seasons. I did teh same thing with pet candidate Charlie Keller, docking ahim a few WS from the straight line projection for 1945.

1945 was not a shortened season. They played 154 games, just like they did in 1946. When you are adding war credit, its not about the value of what a player actually did, its a best guess of what he would have done had he been playing.

The strike was different - it really was a 108 game season, and the pennant was decided based on those 108 games. There were no "additional games" for which to project. Evans had 26 win shares in a 108 game season, which has precisely the same value as 39 win shares in a 162 game season. This isn't my opinion - its a fact.

Burns (as my ballot comment has stated for the past couple of elections) should be compared to Roush or Brock or Duffy or GVH.

You're right, Max, I just wanted to see him in this group because my own system supports his candidacy.

The stat doesn't really seem arbitrary - instead, it seems designed to make Evans look as bad as possible.

Honest Abe, I wasn't trying to make Evans look bad, Diz. Just thought it up on the spot over lunch as a quick-n-dirty. Sorry if you thought I was purposely trying to sabatoge him.

It ignores playing time and injuries - Evans missed very few games in his best years, while, for example, Reggie Smith missed nearly 50 games in one of the seasons you have listed. And so on.

That's not true, though it's possible I wasn't clear about it. I was awarding points based on how many PAs they had in those years in which they had a 120+ OPS. Though it is true that I decided that I wanted OPS+ to count more than playing time, so you'd have me there.

Its probably possible that you could create a junk stat that makes every candidate that's even close to borderline look especially good or especially bad.

Absolutely! I wanted a quick/dirty way of assessing the backlog of corner guys for offensive performance and secondarily for durability. Those kinds of little cutoffs and stuff always produce problems, nature of the beast. It's just showing that Evans is in the middle of the pack. If you account for those 9s, well, then he's probably around Cepeda instead. He's still not leaping upward toward the top of the group.

Actually wasn't the pennant decided after a series of playoffs where the first place teams from the first and second halves of the season played each other? If you want to say that 26 WS in 108 games is exactly worth 39 WS in 162 games, maybe you shoudl split the season up as it was actually plit up and go from there.

For me, I see them (war and strike credit) as one and the same. I choose not to do a linear translation. You have your way and I have mine.

Even a linear translation, however, does not change the fact that Parker licks Evans in top 5 years while neither is lacking in career value. That was my larger point and it still holds.

The strike was different - it really was a 108 game season, and the pennant was decided based on those 108 games. There were no "additional games" for which to project. Evans had 26 win shares in a 108 game season, which has precisely the same value as 39 win shares in a 162 game season. This isn't my opinion - its a fact.

Well, we can construct ridiculous WS amounts for Levi Meyerle's 1871 season when his team only played 28 games. Does anyone think he would have hit .492 if he had played a much longer season?

When comparing players from different eras, you have to use regression or your analysis will be skewed. I'm not presumptuous enough to say that is a "fact," but I don't see any other choice.

Actually wasn't the pennant decided after a series of playoffs where the first place teams from the first and second halves of the season played each other?

Yes. In the NL, the Cardinals & Reds would have been in the NLCS if the season was not split. Neither team won either half so they both watched the playoffs on TV.

I'm not making a point with this... its just a good anecdote. ;-)

And if we do incur some sort of regression, its only fair to also do it for players who had off years in 1981. Reggie Jackson comes to mind. I'm not sure how many non-moot examples of this are.

I'm not strongly arguing either way, but I'm just wondering if "what to do with 1981" has been fully debated here before.

Mike Schmidt, Steve Carlton, and Tom Seaver were in another league and Eddie Murray will be.
Grich is in. Valenzuela is moot, too.

That leaves Dwight Evans, Andre Dawson, and Rollie Fingers with career years in 1981 and open questions about their Hall of Merit cases. (I guess that Evans will go in this year without needing serious 1981 strike talk but I've been wrong before.)

RIGHT FIELDERS

cravath 213 172 171 160 153 150* 147 147* 142* 136* 136 119 106
tiernan 163 160 158 156 152 147 134 132 127 124 121 87
r smith 167 161 157 153 150 143 143 137 133 130 128 127 125 116 101
singleton 165 156 153 152 147 142 135 132 131 119 118 109 101 62
bonds 153 146 143 136 135 133 132 123 123 119 118 117 72
evans 163 156 149 147 137 136 131 129 125 124 120 119 115 111 110 110 106 103 93
staub 166 152 147 147 138 137 136 132 131 129 118 118 117 116 113 108 100 84 79 78
parker 166 149 148 144 141 132 118 117 116 113 111 106 103 103 97 92 80

As noted OPS+ is not perfect, but mostly the guys above Evans also benefit from the OBP underrating effect of OPS+.

One aspect of the imperfection of OPS+ that is rarely considered is its dependence on league quality. Staub's 166 OPS+ in 1969 (10th highest in MLB) is not nearly as impressive as Parker's 166 OPS+ in 1978 (#1 overall - 161 was 2nd highest). When looking at players in a given year or span of years with similar playing time, OPS+ works pretty well as a measure of hitting. When looking at players over different time intervals, it is much less effective.

Below is a chart of 3-Year Averages in the MLB Top 10 OPS+ figures from the past 45 years, using the Mean of the Top 10 OPS+ from each year and averaging it with the Means of the previous 2 years. For example, 1975 shows the average of all the Top 10 OPS+ figures from 1973-75.

Year OPS+ Rank
1962 161.1 2
1963 161.1 2
1964 158.5 2
1965 158.1 1
1966 161.4 2
1967 164.2 2
1968 166.6 2
1969 170.3 3
1970 167.7 3
1971 168.9 3
1972 165.8 2
1973 163.0 2
1974 160.5 2
1975 156.8 1
1976 156.0 1
1977 156.1 1
1978 155.7 1
1979 157.0 1
1980 157.1 1
1981 158.4 1
1982 157.0 1
1983 154.5 1
1984 151.3 1
1985 152.7 1
1986 151.2 1
1987 155.0 1
1988 155.5 1
1989 160.9 2
1990 160.3 2
1991 160.4 2
1992 161.7 2
1993 163.8 2
1994 171.2 3
1995 170.9 3
1996 174.3 3
1997 170.4 3
1998 171.4 3
1999 167.3 3
2000 169.2 3
2001 174.1 3
2002 179.2 3
2003 177.1 3
2004 173.5 3
2005 167.7 3
2006 166.4 2

I divided the 45 years into 3 ranking categories: 1 for the lowest 15 years, 2 for the middle, and 3 for the highest 15 years. A definite trend emerges in which OPS+ figures in the 70s and 80s are considerably lower across the board than in the 60s and 90s/00s. One could guess that this lower OPS+ period at the elite level is a byproduct of a lack of hitting talent at the top of the league relative to the other eras and that perhaps a greater proportion of elite talent was on the pitching side in this era. To test this, I did the same thing with 3-Year Averages of ERA+ figures.

Year ERA+ Rank
1962 137.9 1
1963 142.0 1
1964 149.7 2
1965 151.8 2
1966 153.8 3
1967 148.4 2
1968 153.6 3
1969 155.8 3
1970 152.7 2
1971 148.6 2
1972 147.4 2
1973 151.6 2
1974 149.2 2
1975 145.4 1
1976 141.5 1
1977 143.5 1
1978 146.6 1
1979 146.3 1
1980 144.4 1
1981 143.6 1
1982 144.2 1
1983 142.5 1
1984 137.4 1
1985 146.7 1
1986 148.1 2
1987 150.4 2
1988 146.6 1
1989 146.6 1
1990 150.0 2
1991 148.2 2
1992 152.9 2
1993 152.2 2
1994 159.1 3
1995 160.7 3
1996 163.0 3
1997 165.2 3
1998 164.7 3
1999 163.7 3
2000 159.9 3
2001 155.2 3
2002 158.2 3
2003 159.7 3
2004 162.8 3
2005 160.5 3
2006 152.1 2

These numbers display the same trend. In fact, for the periods 1972-85 and 1990-2006, the rankings are identical. The best players begin to put up lower OPS+ and ERA+ numbers in the early 70s and do not show any deviation at all from this low productivity until the mid- to late 80s. The top players then begin to increase their OPS+ and ERA+ output around 1990, culminating in a big leap forward in the mid-90s that is sustained until quite recently. I suppose it is possible that the league's best hitters and pitchers suddenly became worse in the 70s and then greatly improved in the 90s just as quickly. I find it more likely, however, that differences in the level of the average player primarily attributable to league expansion are responsible for these vast and sudden disparities. If this is the case, OPS+ and ERA+ figures need to be adjusted to account for it.

I believe the subject of the 1981 short season has come up in discussions of Rollie Fingers.

1945 was not a shortened season. They played 154 games, just like they did in 1946. When you are adding war credit, its not about the value of what a player actually did, its a best guess of what he would have done had he been playing.

1945 was a shortened season for Charlie Keller, who returned from military service in time to play in the last third of the season. The issue of regression is thus pertinent in his case.

It's no surprise that there are serious playing time issues for everyone at the top of Eric Chalek's list, else they would not be outside the Hall (maybe Tiernan would, a la Ken Singleton). Reggie Smith, Norm Cash, Frank Howard, and Mike Tiernan --newcomers, take a look-- didn't play enough. Gavy Cravath and Bob Johnson didn't play enough in the majors either. Their HOM cases depend on some serious consideration of their late arrivals. Most of us grant that consideration and many give some "credit" after considering, enough to keep Cravath and Johnson viable. (Cravath and Johnson would both be on my ballot. Of the others only Frank Howard would have a chance, one depending on consideration of what caused his limited early playing time. I would need to give that consideration if I were voting.)

That leaves Dwight Evans, Andre Dawson, and Rollie Fingers with career years in 1981 and open questions about their Hall of Merit cases. (I guess that Evans will go in this year without needing serious 1981 strike talk but I've been wrong before.)

Nolan Ryan's got a major outlier season in 1981. Ryan was overrated and if the HOM were a bit smaller, it might matter, but there's just too much career value there and he'll sail in easily.

When comparing players from different eras, you have to use regression or your analysis will be skewed. I'm not presumptuous enough to say that is a "fact," but I don't see any other choice.

John, there are two different issues here. The first is how valuable a player was in a given season. The second is how that season should be viewed when voting for the HOM. We can, and I believe do disagree on the second of the two issues. But that does not change the first - its a mathematical fact that 26 win shares in a 108 game season has the same value as 39 win shares in a 162 game season. I understand the view that Evans should not automatically be credited with a 39 win share season as a result in HOM considerations.

I find it more likely, however, that differences in the level of the average player primarily attributable to league expansion are responsible for these vast and sudden disparities. If this is the case, OPS+ and ERA+ figures need to be adjusted to account for it.

I appreciate the work you did, and find it helpful. My only quibble is that I wouldn't assume that expansion is the only or even primary factor. Personally, I believe that a general shift toward offense over defense in the early 90's, unrelated to expansion, is more of a factor, although I have no evidence one way or another.

Well, we can construct ridiculous WS amounts for Levi Meyerle's 1871 season when his team only played 28 games. Does anyone think he would have hit .492 if he had played a much longer season?

Just because it looks funny doesn't mean it isn't true. We shouldn't be in the business of taking value away from players because we don't find the numbers aesthetically pleasing.

Meyerle's 1871 Win Shares or batting average can reasonably be adjusted downward for quality of competition or standard deviation or whatever, but schedule adjustments are equivalencies, not approximations.

John, there are two different issues here. The first is how valuable a player was in a given season. The second is how that season should be viewed when voting for the HOM. We can, and I believe do disagree on the second of the two issues. But that does not change the first - its a mathematical fact that 26 win shares in a 108 game season has the same value as 39 win shares in a 162 game season. I understand the view that Evans should not automatically be credited with a 39 win share season as a result in HOM considerations.

You have to forgive me, Diz. My thinking is almost always entwined with the HoM. With that said, I agree with your first issue as you have framed it.

Just because it looks funny doesn't mean it isn't true. We shouldn't be in the business of taking value away from players because we don't find the numbers aesthetically pleasing.

I'm not taking anything away from him, Sean. Whatever he did within those 28 games is a fact. Prorating his numbers, however, is not factual for HoM, since he he didn't actually play 154, 162 or whatever amount of games that you want to use. I'm 100% confident that, while he would have been the best hitter with terrific numbers, his numbers would have regressed with a larger schedule.

OK, it's probably high time I did this since I'm the doubting Thomas. The following is what my system shows. I use WS to answer the answerable Keltner questions, and Keltner is helpful because it examines players from many points of view and offers them many opportunities to excel.

A Win-Shares-oriented, Keltner-Based Analysis of Dwight Evans’ HOM case

1) Could he ever have been described as the best player in his league?

I don’t believe so no. Using a three-year average of his percentage of the AL MVP’s WS total, there’s little evidence he had a strong argument for being considered the best player in his league at any time. His career really picks up around 1980, when I’ve got Brett as the top guy. Soon thereafter, Henderson, Yount, Boggs, and Ripken are theleague’s best. Evans is never within ten percent of the league’s leader. I award him zero “Keltner Points.”

2) Could he ever have been considered the best player at his position?

My evaluation here is the same as above, only restricted to his position. The player need have played RF in the current season but can have played any other OF position in the two years before (figuring that RF and LF are reasonably close in skill level and that CF is more difficult but not so different as, say, 3B and SS). Playing any other position eliminates the player from consideration until he has amassed three consecutive years at the position. Evans’ primary position was RF from 1973 through 1986. He never again put in three seasons at one place to qualify for this analysis, though it’s close.

Our focus period, then is 1975-1986. During this time he had a legit claim to being the best RF in his league 1982–1984 and 1986. I award one point each for four points. In addition, Evans was within ten percent of the best RF in his league in 1985. I award him .1 points for this. He totals 4.1 points.

3) How many MVP-type seasons did he have?

The assessment is How often did he finish among the top-five position players in his league, or expanded equivalent? Evans finished in the money four times, 1981 (2nd), 1982 (2nd), 1984 (3rd), 1987 (9th). Evans gets 4.0 points, bringing his total to 8.1 points.

4) How many All-Star type seasons did he have?

I award one point for each instance where a player is one of the two best players at his position in his league, or expanded equivalent. Evans was a WS All-Star in 1989, 1988, 1987, 1986, 1984, 1982, 1981. Seven times for 7.0 points. That brings him to 15.1 points.

5) Is his career comparable to already enshrined players?

To be honest I use the HOF here. Only because I use this system to evaluate both HOF and HOM, and I’ve been too lazy or preoccupied to rejigger this for HOM-only use. But the system is the same, I figure out how many HOFs are at the player’s primary position, then take his career WS, and see how many HOFs are above him, subtract from the total HOFs at position, then divide that into the total number at position and multiply by 10. One wrinkle, I use adjusted WS here, so strikes, war-cred, MiL cred, MLEs for NgLs are all included. Evans scores at 359 adjusted WS, which and there’s 16 HOFs above him, and 23 RF HOFs total. 23-16 / 23 = 3.0 points. Adding to his current total, that’s 18.1 points.

6) Was he good enough to play past his prime?

I find the player’s ten best consecutive years. That’s his prime. After that, I count the total value of all his seasons of 15+ WS after his prime. The points are awarded based on a scale.

Evans confounds the system a bit in this regard due to his late prime. His primiest prime times were 1980 to 1989, 242 WS. After that, he had no years of 15+. So he gets zero. Let’s say he had the Dave Parker career arc, and all the garbage time came after the prime. Well, then he’d have scored up to six points in the category. He still wouldn’t be over the line, and there’d be a couple names ahead, but he’d be closer. On the other hand, once his prime was over, he was pretty much done, which is what the question is asking. James never defined prime, he never made clear whether he meant most-productive years (as I’ve defined it) or whether he merely meant youth. Usually the two go together, so he probably didn’t feel he had to. In this case it’s not the same, and that could be seen as a liability of how I’ve constructed this system. Still 18.1 points.

7) Does he meet the HOM’s standards?

Again I’m using the HOF, not HOM, same reasons. I take the 3, 5, 10, 15, career numbers of all HOFs. Then I split them up by position. Then I sort in each of these five categories and retain the adjusted WS total of the 25th percentile player at the position. That forms the basis of the standards I’ll observe. I compare the current player in each of those five categories, counting them equally. Anyone who scores 1.0 or higher gets 10. Anyone above 0 gets 5, anyone below 0 but within ten percent gets 1. RFs have the following for 3/5/10/15/career standards: 93/144/250/306/313. That’s the minimum standard. Evans goes 98/14/246/322/359. En toto that’s a .26 score and I award him 5 points. 18.1 points plus 5 is 23.1 points.

8) Would a team with Dwight Evans as its best player win the pennant?

As I’ve mentioned in other threads, I answer this question by taking a player’s fifth-best season. Then I compare it to the historical norms for the WS leaders of pennant-winning teams. Evans’s fifth-best season clicks in at 24 WS. Historically, teams with 6 percent of all pennants were won by a team led by a 24 WS player (that’s adjusted WS). That’s similar to Jake Beckley among others. Evans gets 1 point, bringing his total to 24.1 points.

When I ballotize him (yes, made that one up), I ask two more questions.

Is he the best available player at his position?

No. Not according to this system. YMMV. My system sees him as falling behind Cravath, Bonds, Tiernan, Oliva, Cuyler, and Reggie Smith. It also suggests that Bobby Murcer is better—I don’t agree with the system there. Oliva and Cuyler are highly questionable as well, though Cuyler has enough CF time that it might be a less straightforward comparison. But Cravath, Bonds, Tiernan, and Smith are all formidable opponents for Evans, and I don’t think he’s better than all of them.

Is he the best available player?

No. See above. There are many other players who my system believes have obviously better credentials at a variety of positions. They’re all on my ballot. You agree with some, disagree with others, naturally. If the system is at fault, then his rating would have to be so undercooked and everyone else’s overcooked, that it would be difficult to prove that the whole thing was wrong just because Evans is off. (Of course, so is Parker.) I’ll concede that it’s not perfectly constructed, and that peak seasons like Murcer’s and Parker’s get double-counted, but Evans benefits by it as well, so I don’t think that necessarily means the whole thing is being unfair to Evans as an individual.

So where does that leave Evans? Well, 24.1 points is not great in my system. It’s between Hooper and Cuyler in RF. It’s about as good as Beckley and Konetchy, Elliott, Latham, Fielder Jones, and Lou Brock. Those aren’t guys I’m supporting, so I’m not feeling rosy about Evans. If he were a catcher, he’s in, but he’s not.

OK tear on in....

Obviously, there are a lot of points I disagree with, but I'll take just one.

Our focus period, then is 1975-1986. During this time he had a legit claim to being the best RF in his league 1982–1984 and 1986. I award one point each for four points. In addition, Evans was within ten percent of the best RF in his league in 1985. I award him .1 points for this. He totals 4.1 points.

Evans wasn't the best rightfielder in the American League in 1981? Who was better?

I'm not taking anything away from him, Sean. Whatever he did within those 28 games is a fact. Prorating his numbers, however, is not factual for HoM, since he he didn't actually play 154, 162 or whatever amount of games that you want to use. I'm 100% confident that, while he would have been the best hitter with terrific numbers, his numbers would have regressed with a larger schedule.

But that's irrelevant to his merit; it's an assertion of ability over actual value. A pennant that takes 28 games to decide is just as valuable as a pennant that takes 162 games to decide (league strength adjustment notwithstanding). 10 win shares in a 40 game season is identical to 40 win shares in a 160 game season.

Evans wasn't the best rightfielder in the American League in 1981? Who was better?

He wasn't the best RF for the 3-year period ending in 1981. I always forget to write it as three-year periods, and it's easy to get tripped up. My bad. So for the years cited, he's the AL's best as gauged over a three-year period. That being the shortest amount of time that I could imagine a player establishing a claim as the best player in league or at position.

Let's use a more contemporary example: what was Jeff Bagwell's value in the 1994 season?

He earned 30 win shares in 115 team games, and had the strike not happened, he would not have earned any more as he was going to miss the rest of the season with a broken hand (IIRC).

Is his value

A: 30 in 115

B: 30 in 162

C: 42 in 162?

I argue the answer is C. When we attempt to measure and compare players' value across disparate season lengths, we are not imagining a world in which the 1994 season had 162 games, nor are we assuming the strike would not have happened and extrapolating from there. Instead, we are attempting to measure the pennant-value of the player's actions in that specific pennant-season. We are not creating value that didn't exist, we are merely translating it into a single season length denominator.

The 1994 pennant took 115 games to decide, and 30 win shares in that season had exactly the same pennant-impact as 42 win shares did in the 1993 season.

The strike was different - it really was a 108 game season, and the pennant was decided based on those 108 games. There were no "additional games" for which to project. Evans had 26 win shares in a 108 game season, which has precisely the same value as 39 win shares in a 162 game season. This isn't my opinion - its a fact.

Would it be a fact if it were true? I don't believe it's true and I wonder whether the Pennants Added measures presume this.

A player who contributes 4 extra wins in 108 games (single-season of that length with one pennant winner) does not in theory have the same effect on probability of winning the pennant as a player who contributes 6 extra wins in 162 games. That is because team margins in pennant races (in at least one simple probability model) are not proportional to the length of the season. On the field, not the blackboard, however, other things vary at the same time. The 1981 division pennants were famously decided by method other than division wins. And team margins in the two ~54-game half-seasons may not have been what the simple probability margin would predict, either.

(By the way, beside StL and Cin with the best records measured by division wins, the two AL mini-series featured 1 vs tie3/4 and 1 vs 4. Especially Texas and Chicago, also Baltimore Detroit and Boston earned nothing for good half-seasons. The teams did not all play 108 games, as someone said. That is merely a convenient (because 108/162=2/3) and close approximation. In the AL East, Baltimore finished 1 game ahead in the not-so-important loss column, with 105 games played to a decision, vs 109 for Milwaukee (the leader by wins, WL%, and GB) and 103 for Cleveland (sixth place at .505).

On my first run-through, my peak-heavy but prime-attentive system puts Evans and Parker in almost precisely the same place, and very close to Darrell Evans and Hernandez (neither of those two are in my pHOM yet, so that's not as high praise as it would be from many). They're all in the mid-20s, so just a bit off-ballot.

Since I do question how well WS credits great defense, I think Dewey gets the edge in the end, but it's not by a ton. I suspect when I'm finished with this "year," Evans will join the other Evans and Hernandez on the fringe of pHOM-dom (like them, he'll probably make it eventually, though it's possible one of the three won't), while Parker will trail by just enough to keep him out.

(By the way, I agree with Sean and Dizzypaco re 1981. I'm giving Evans—and everyone else—full prorated credit for that year, as I always have since we got to players who played that season.)

(Now I suspect that my system gets the data from thinkfactory slowly when someone else is hitting the database at essentially the same time.)

I don't say anyone should do any differently than Sean G and Mark D regarding mental or spreadsheet translations figures such as 26 WS (rounded to not-necessarily nearest integer by Bill James) and 108 games for Dwight Evans 1981. But sabrmetricians (sabrmetrical designers) of Pennants Added, for example, should consider the point more carefully. They want to get it right in a different sense and to another decimal point if not two.

During the second half-season, Cleveland gained slightly on the rest of the division in games played, barely beating New York for fifth place in WL%, .491 to .490. Two behind New York in the not-so-important loss column for the first half, one behind for the second half. Ah, the memories. A summer of overstudying for exams because there was no baseball.

St Louis and Cincinnati - both five games ahead in the season loss column. St Louis played 102 games to a decision, tie for the ML low with Pittsburgh. San Francisco played ML high 111.

Dandy Little Glove Man--what you are referring to by pointing out that the same OPS+ is not the same rank in all eras is the standard deviation of the league--exactly what my WARP system is designed to account for. 1969 was an expansion year, which means a high standard deviation, which means that lots of guys will have career years (McCovey, Jackson, Staub, Wynn etc.), all of which need to be docked. Conversely, low-scoring years far removed from expansion, the mid-80s in particular, need to be boosted.

My system loves Dewey even more than Darrell, although I do straight-line adjust 1981 and given my salary estimator's peak preference that obviously helps him--I have 22% of his value coming in 1981. With a straight-line adjustment, I have it as the 76th most valuable season since 1893 (at a level with Musial's 1948, Banks's '59, Hornsby's '25, and Mays's '65, to give you an idea of how much I like it). On OPS+ it's just a 163, but you have to add on 1. He was, according to James Click, an absolutely outstanding baserunner, adding 5.7 runs with his legs in just 108 games 2. He was a superlative outfielder, with +10 runs in just 108 games, and 3. He was playing in the lowest standard deviation era in major league history, a low-scoring league far removed from expansion with a very high population per team. Clearly there is a strong case it should be regresesd, but through those 108 games he was stunningly good.
Aside from the monster peak year, his 1982 and 84 were high All-Star seasons, 85 and 87 were strong All-Star seasons, and 75, 86, 88, and 89 were marginal All-Star seasons. He was further above-average from 76-80 thanks to his terrific fielding. So you have peak, prime, and career. What's not to like? He's a no-brainer for me, and will be in an elect-me spot on my 1997 ballot.

I would argue he was the best Red Sox position player between Yaz and Nomar/Manny. Boggs was overrated, IMHO and Evans was clearly better than Rice, Lynn or Vaughan.

But that's irrelevant to his merit; it's an assertion of ability over actual value. A pennant that takes 28 games to decide is just as valuable as a pennant that takes 162 games to decide (league strength adjustment notwithstanding). 10 win shares in a 40 game season is identical to 40 win shares in a 160 game season.

Okay, but I think you have to acknowledge that top players dealing with shorter schedules have an unfair advantage over players with longer schedules.

That's something standard deviations would measure, is it not?

On OPS+ it's just a 163, but you have to add on ...

How much one weights this depends on how you adjust for season length, but Evans also had a ridiculous number of plate appearances in 1981: 504 in 108 games, which would work out to 756 in 162 games. That would be the thirteenth-most of all time, the second-most of the 1980s, and I believe the most ever for a player not primarily used as a leadoff hitter.

I think most people agree that short seasons--I mean, really short seasons, not the 154 game variety, but the 100 game seasons and less--can't really be extrapolated directly to 162. This is especially true for pitchers and catchers but to a lesser extent for everybody. But how much you discount/regress, etc., is pretty much a matter of taste.

Oh whoops, I forgot to mention Evans' insane # of plate appearances in the strike year. He played every game and was hitting 2nd on a team that led its league in OBP by like 10 points. I did mention that in the 1996 ballot discussion thread. That's another big factor in his giant WARP score.

Sean Gilman, I put season length in my regression of standard deviation--it definitely does show a correlation, but it wasn't statistically significant so I didn't include it in my final equation. I may try to re-do the regression and tweak it to see if I can get that in.

Taking Evans and the other mostly-offense recent candidates, cutting off their years at beginning or end of career where they had essenetially no value, measurin gby RCAP, which adjusts for position played but not for defensive quality:

player years RCAP Games
D Evans 72-89 267 2382
T Perez. 67-78 221 1815
R Smith 68-82 286 1823
B Bonds 68-79 224 1718

Agai, RCAP measures above AVERAGE, so if you want above REPLACEMENT, career length is an important factor. Isn't it obvious that Dwight is a clear #1 on this list, even BEFORE accounting for his fine defense? Really, the only argument for not putting Reggie Smith #2 is that he spread his value in the same # of games across more seasons (less in-seaosn durability).

Oh, and karlm, you really have Evans as a bigger star than Boggs? I can't wrap my arms around that at all.

That's something standard deviations would measure, is it not?

Yes, it would, Sean.

Oh, and karlm, you really have Evans as a bigger star than Boggs? I can't wrap my arms around that at all.

I don't understand that either, Tom. Boggs is near the top of the greatest third basemen ever, while Evans can't remotely have that said about him.

IS that when Hriniak got ahold of him, kev?

I agree that Boggs is oevrrated by those who use sbaermetric stats only, since he was a poor baserunner, and this weakness was magnified by the difficulty of placing him in a lineup to match his extreme on-base skills.

However, Boggs is credited with 1703 runs created in his career; 143 more than Evans. And he used up 380 fewer outs in doing so. Take away 100 runs, and he was still a bigger offensive force than Dwight, and probably about even on defensive value.

As someone else stated, while I have Wade as nowhere near as valuable as the BIG 3 3Bmen, he is clearly MUCH higher in the overall 3B rank than Evans is among RFers.

Adding up the SB-CS and James Click's non-SB baserunning runs, Boggs cost his teams 3.3 wins with his legs over his career. Not good, but not a big enough effect to meaningfully change his overall value.

"Boggs was basicallly station to station."

He really was. And as a person, he's turned out to be almost comically doucheb@gian.

Adding up the SB-CS and James Click's non-SB baserunning runs, Boggs cost his teams 3.3 wins with his legs over his career. Not good, but not a big enough effect to meaningfully change his overall value.

[pokes up head, grins]

Anyway, he'd have to be a better baserunner than Max Carey to avoid all of those DP balls off Rice's bat!

[ducks!]

Put me in the camp to doesn't regress one iota for 1981. The teams played 103-111 games. I simply do a straight line adjustment. A pennant is a pennant. Even if Cincinnati and St. Louis didn't get one that they should have :-)

Sean Gilman, I put season length in my regression of standard deviation--it definitely does show a correlation, but it wasn't statistically significant so I didn't include it in my final equation. I may try to re-do the regression and tweak it to see if I can get that in.

I think people may be discounting/regressing for league strength and then doing it again for schedule length, thus doubly penalizing short-season players.

If the advantage that short season players have is an easier time putting up higher deviations, the players should be given a straight line adjustment to 162 games, and then discounted for league strength.

If what you're saying is that schedule length isn't a significant factor in the size of the standard deviations, then do you agree that short season players should be translated directly to 162 games?

"Okay, but I think you have to acknowledge that top players dealing with shorter schedules have an unfair advantage over players with longer schedules."

No they don't. They are just as unlikely to have a bad 40 games (relative to their ability) in the next 40-game season (see Start, Joe or Sutton, Ezra).

If the season is 40 games, and a player performs amazingly well and leads his team to a pennant that's just as valuable as if he does it in a 162 game season. Degree of difficulty should not be an issue.

If the season is 40 games, and a player performs amazingly well and leads his team to a pennant that's just as valuable as if he does it in a 162 game season. Degree of difficulty should not be an issue

I'm all for a "pennant is a pennant" and my ballots attest to that, Joe, but I can't subscribe to your position here. The SD for a shorter scheduled season is always going to be greater than a longer scheduled season. That to me is just unfair to the 154- and 162-game schedule players.

Well, looks like Kevin beat me to the punch. :-) Thanks.

Dewey is a key member of the HoVG; his career defines my expectation level of what a member of the HoVG should have attained for enshrinement.
He, Rice and Lynn are perhaps the very definition of a HoVG outfield; each one a member of the Hall and together they attained a level of "VeryGoodness" not reached by many other outfields.

That said, the 1889 Chicago White Stockings outfield of Duffy, Van Haltren and Ryan practically screams to recognized as one of the truly VG outfields of all-time.

I apologize for going a bit off topic there but I think a robust Hall of the Very Good might be an excellant exercise once this hall is almost complete.

John Murphy, if you are sensitive to SD's for short seasons (as well you should be), what about for post-expansion and high-scoring years? And are you properly crediting for their opposites (long seasons, many years since expansion, low-scoring)? Take a look at the graph in the file I've posted to the Hall of Merit Yahoo! group, it shows you how much to adjust for stdevs for each league-season.

Sean Gilman, difficulty and standard deviation are not the same thing. This has nothing to do with league strength. There is definitely a correlation between season length and stdev, it's just not robust enough for me to include in my regression yet. Hopefully I'll be able to tweak it enough to get it in for version 2.0.

But he should have had more runs scored and RBI's than he did. His runs scored are low for someone who had such great OBPs and his RBI totals are pretty pathetic for someone with so many doubles and who hit at such a high average. S

RBIs- you did notice that he batted leadoff didn't you? :-)*

RUNs- you're right they look low for his times on base, and intuitively it seems that he's somewhat to blame for the obscene GDP totals put up by the men batting after him, but...

Ichiro Suzuki was n base 311 times in 2004 and scored just 101 runs, he was on base 254 times in 2005 and scored 111 runs - the difference was the offense coming up behind him.

Suzuki's run scored performance in 2004 looks alot like Boggs during his 200 hit days- and we can all agree that Ichiro is not a base clogger.

Still most of the time (The Buckner years excluded) the Sox coming uop behind Boggs looked a lot more formidable than the Mariners coming up after Ichiro in 2004.

* Looking at BBREF- Boggs' splits are just odd looking- he had 2327 PA batting 3rd, he hit .329/.422/.470 batting third and yet had only 246 rbis batting third. Granted that when Boggs batted third Remy frequently lead off- but that's still odd.
In 1987 he slugged over .600 and batted 3rd most of the year and drove in only 89 runs- the men he was batting behind, Burks and Barrett were not good at getting on (.351 and .324) but not Dusty Baker Cubs' terrible either... He had only 127 at bats with RISP a shockingly low # for a #3 hitter- in fact Dusty Baker Cubs terrible- Derrick Lee had only 124 at bats with RISP in 2005.
In fact Boggs did have more baserunners to drive in 1987 than Lee had in 2005- but they were on FIRST not in scoring position, and I'm not sure if a double off the Monster scores a runner from first...

Sean Gilman, difficulty and standard deviation are not the same thing. This has nothing to do with league strength. There is definitely a correlation between season length and stdev, it's just not robust enough for me to include in my regression yet. Hopefully I'll be able to tweak it enough to get it in for version 2.0.

Yes, but "league strength" is one of the things you're attempting to correct for with standard deviations, isn't it? There are lots of factors that contribute to differences in standard deviation across eras, it's asserted that league difficulty is one, and season length is another, correct?

If you just adjust for standard deviation (which I believe you do) there's no problem. But, not all voters are as scientific, adjusting short season players to some less-than-162 number and then reducing those numbers for league strength issues.

John Murphy, if you are sensitive to SD's for short seasons (as well you should be), what about for post-expansion and high-scoring years? And are you properly crediting for their opposites (long seasons, many years since expansion, low-scoring)? Take a look at the graph in the file I've posted to the Hall of Merit Yahoo! group, it shows you how much to adjust for stdevs for each league-season.

Don't know if I do it properly, Dan :-), but I try to take into account all of that.

I will check out that file of yours when I get the chance. Thanks!

Well, there's no explicit "quality of play" independent variable in my standard deviation regression. Two very important variables are population and expansion, both of which probably also have an effect on league quality, but I prefer to simply call them population and expansion and be content with their statistically significant relationships to standard deviation. I hope to tackle quality of play at a later date, for now, I have no view on it and make no adjustment for it in my system. Again, stdev and quality of play are not the same thing--the stdev in the AL went *down* substantially during World War II, probably because all of the offensive superstars (Williams, DiMaggio etc.) went to fight while many scrubs stayed.

My only adjustment to raw wins above replacement is for regression-projected standard deviation.

League strength and season length simply have nothing to do with one another. It's perfectly possible that a league could both be weaker and have a shorter season, in which case two separate adjustments would indeed be in order.

Since kevin and JPWF13 have brought the subject up, time for me to trot one of my favorite cheap-toy stat comparisons:

Tim Raines and Wade Boggs both batted some leadoff, and some third. The mix of batting order positions is different, but it's not night-and-day different.

Boggs generally played for higher scoring teams (mostly because of higher offensive circumstances) than Raines. Hence Boggs' teammates contributed more to his R and RBI than they did to Raines.

Including GDP and CS, they made essentially the same number of outs in their careers: 6527 for Raines, 6566 for Boggs.

Raines scored 1548 runs, and drove in 946, for 2512 R+RBI.

Boggs scored 1513 runs and drove in 1014, for 2527 R+RBI.

(Subtracting HR to make it "runs produced" tilts a little towards Boggs, 2409 to 2342.)

So Raines had essentially the same R+RBI as Boggs in the same number of outs, despite the rather large difference in basic batting statistics that points in Boggs's favor.

Since when is being compared to Tim Raines a point *against* a player? If Boggs had similar offensive value to Raines but played a tougher position (3B), that simply confirms his spot comfortably within the pantheon.

Well, there's no explicit "quality of play" independent variable in my standard deviation regression. Two very important variables are population and expansion, both of which probably also have an effect on league quality, but I prefer to simply call them population and expansion and be content with their statistically significant relationships to standard deviation. I hope to tackle quality of play at a later date, for now, I have no view on it and make no adjustment for it in my system. Again, stdev and quality of play are not the same thing--the stdev in the AL went *down* substantially during World War II, probably because all of the offensive superstars (Williams, DiMaggio etc.) went to fight while many scrubs stayed.

My only adjustment to raw wins above replacement is for regression-projected standard deviation.

League strength and season length simply have nothing to do with one another. It's perfectly possible that a league could both be weaker and have a shorter season, in which case two separate adjustments would indeed be in order.

Now I'm confused. I don't believe I any longer have any idea what you are talking about. But I'm pretty sure we aren't talking about the same things.

How many win shares was Jeff Bagwell worth in 1994?

Rest assured that (looking ahead to make sure) Boggs will be #1 on my 2005 ballot. That was really an attempt to note how much difference baserunning might make.

Another illustration of the stdevs-aren't-league-quality point: stdevs in the 1910's AL were extremely high, while those in the 1910's NL were extremely low. That was simply because all the superstars were in the AL, which made it a stronger league with a higher stdev. (My WARP, which adjust based on a regression-projected stdev, regress both leagues more or less to the same degree, correcting for the run scoring level and population etc. but not for the "star glut" in the AL or "star drought" in the NL.)

Well, I have him at 8.3 WARP1 when the strike hit. First, I'd straight-line adjust that to 11.8 in a 162-game season. Next, because the 1994 NL was a high-scoring league just one year removed from expansion, I'd reduce it to 11.3. Finally, I'd want to make a further adjustment for a higher expected standard deviation due to season length--but I'm not sure how large that should be yet, since I can't get a statistically significant correlation between season length and standard deviation. I'd make no adjustment for league quality whatsoever.

What I was referring to would be a year in which, say, the AA had a shorter schedule than the NL in the 1880's. Then you'd have to make two adjustments--one for season length, and another for quality of play.

OK, that makes sense to me. I think we're in agreement.

JPWF13 #93
In fact Boggs did have more baserunners to drive in 1987 than Lee had in 2005- but they were on FIRST not in scoring position, and I'm not sure if a double off the Monster scores a runner from first...

Bill James wrote (in the 1988 Abstract?) that Boggs was very quick getting out of the batter's box and up to speed toward first base. That would boost his batting average and his doubles rate, relative to someone with his speed, his baserunning skills (judgment, I think that means), and average quickness. And it would proportionately the runner-advancement run-scoring effects of his singles and doubles.

Were several of his doubles off the wall the same hits as the single-and runouts that compose the assist records of Yaz, Rice, Greenwell, and Ramirez --when batted by visitors? Anyway, they weren't scoring runners from first base with one or none out.

--
DanR #97
Well, <u>there's no explicit "quality of play"</u> independent variable in my standard deviation regression. Two very important variables are population and expansion, both of which probably also have an effect on league quality, but I prefer to <u>simply call them population and expansion and be content with their statistically significant relationships to standard deviation</u>.

Thanks, at last I understand clearly what you are talking about.
And Sean Gilman comes to agreement a few articles after saying the opposite.

My coming to agreement often follows someone explaining to me what the hell I'm talking about.

My coming to agreement often follows someone explaining to me what the hell I'm talking about.

Another reason why we're best-friend voters!

30

30

Next time use W.A.S.T.E.

Quoting myself:
[DavidFoss:] I'm just wondering if "what to do with 1981" has been fully debated here before.

Mike Schmidt, Steve Carlton, and Tom Seaver were in another league and Eddie Murray will be.
Grich is in. Valenzuela is moot, too.

That leaves Dwight Evans, Andre Dawson, and Rollie Fingers with career years in 1981 and open questions about their Hall of Merit cases. (I guess that Evans will go in this year without needing serious 1981 strike talk but I've been wrong before.)

DavidFoss #50
Nolan Ryan's got a major outlier season in 1981. Ryan was overrated and if the HOM were a bit smaller, it might matter, but there's just too much career value there and he'll sail in easily.</i>

It doesn't stand out like a textbook outlier but 1981 is Dave Concepcion's best season at bat measured by OPS+ (116) and his only high finish in the MVP election (4).

On the downside, Dale Murphy stands out. But he isn't getting any support here.

Of course, any pilgrim's progress to an understanding of 1981 will pay off regarding 1994 too.

Actually, understanding 1981 is pretty complex, and quite different from 1994 due to the split season in 1981. Teams that won the very abbreviated first half were essentially playing for nothing in the second half. So, for instance, Goose Gossage had a great first half in 1981 and pitched in extremely high leverage situations as the Yanks won the first half title. After the strike, the Yanks used the second half mostly as tune-up for the playoffs. Goose did not pitch as well nor in situations which were as important.

If someone develops a way of quantifying this, my hat is off to him/her. Probably one needs to separate out the teams that won first half titles...Frankly, I suspect that subjective tweaking will be much more satisfying than any formula.

On the downside, Dale Murphy stands out. But he isn't getting any support here.

He's not eligible yet. I'm guessing his prime wasn't quite long enough and the decline came too soon, but I suppose we'll find out when the more thorough analysis comes in a couple of years.

After the strike, the Yanks used the second half mostly as tune-up for the playoffs.

What was MLB going to do if a team won both halfs? Anyone remember?