Looks like David defended himself before I could. ;-)

I like Averill just fine myself, but I wanted to note the possible objection for people with different priorities. Good player or not, we ARE basically talking about eleven seasons' worth of career.

Joe Dimino, Gwynn was much more valuable than Raines in '87!

WS does fairly emphatically not agree.

Well, as usual, I don't know what WS is smoking--I don't even see how Raines is in the top five. Gwynn's OPS+ was 9 points higher than Raines's, he played in 18 more games than Raines did, he was nearly Raines' equal on the basepaths, and according to Chris Dial, he was a *far* better fielder than Raines (Dial has Gwynn at +16 and Raines at -3). I mean, it's night and day...I love Tim Raines as much as anyone, he's a slam-dunk Hall member, but there's no way he was as valuable as Tony Gwynn in 1987.

I know that, David. . . it was just the mood I was in, and a general sort of thing, not directed at you. I actually expected it to be completely ignored!

Well, as usual, I don't know what WS is smoking--I don't even see how Raines is in the top five.

The 1987 Expos overperformed their Pythagorean record by eight wins. The Padres underperformed theirs by six. I suspect that is, in fact, what Win Shares is smoking.

Everybody's talking about Clark as a Giant and a Cardnal, and I don't blame you guys, but I remember him for his two years as a Padre. He was an absolute force, even hitting .242 for in '89 (he caught a lot of flak for the low BA, but he was a big part of that team chasing the Giants for the division title).

His 1990 was quite similar to his '87, except he missed even more time to injury. And most people here have negative memories of him, for calling out Tony Gwynn as selfish. Not that I agree with him, but he was a frickin beast in his SD years. Joe Carter got all the credit for driving in 115 runs, but he had a pretty bad season; it was all due to a top-4 of Bip Roberts, Robby Alomar, TGwynn and Jack Clark.

The 1987 Expos overperformed their Pythagorean record by eight wins. The Padres underperformed theirs by six. I suspect that is, in fact, what Win Shares is smoking.

Agreed. I love Raines, and think he was the best player in the NL at that time. But given the month he missed, it's hard to make the case for him over Gwynn in '87. 158 vs. 149 in OPS+; yeah, Raines was an all-time great base-stealer, but I don't believe 50-5 vs. 56-12 is enough to make up for it, especially when Gwynn had 53 more PA and was a GG RF.

Gwynn even had 20 points of OBP on Raines in '87. 9 RC/9 vs 8.6. Before adjusting for park (22-point difference in park-adjusted OPS).

Not that I'm advocating Gwynn for MVP -- but he did lead the league in VORP by 11 runs and win a GG.

.343 EQA for Gwynn, .335 for Raines. Both 9 FRAA, Raines in LF, Gwynn in RF.

While Gwynn's #'s were slightly better, the fact that Raines' team was efficient and Gwynn's was inefficient is enough for me to swing it to Raines. The win totals have nothing to do with it. If Gwynn's team was 65-97 but should have been 60-102 and Raines' team was 91-71 but should have been 100-62, I'd say give it to Gwynn. But it's quite obvious that the Expos (and therefore their individual players) were better than their individual stats would lead you to believe, and the Padres were worse.

And I'm not going to penalize Raines in the MVP vote because of collusion. I don't care at all that Gwynn played 17 more games, due to the circumstances.

BTW, according to RC in the All-Time Stats Handbook, Raines is at 9.01/27 outs in a 4.52 league, Gwynn is at 8.60.

The Expos PF was 103, San Diego's was 96.

So 9.01/4.6556 or 8.6/4.3392 it's very close. That accounts for the fact that Montreal must have had a much more efficient offense, if Gwynn could be ahead by so much in EQA, but basically equal once you account for having it all add up.

Raines OWP comes out to .789, Gwynn's to .797.

Then you throw in that with the runs the Expos scored they won even more games than they should have and it's very easy to see why WS would have Raines ahead, despite missing the extra games.

PythaganPatting the exponents makes it slightly closer, .806-.801.

Arrgh, let's do this properly.

Gwynn 132.2 XR in 390 outs, Raines 121.2 in 365 outs(including non-SB baserunning and double play avoidance)
Gwynn 96 PF > 137.8, Raines 103 PF > 117.7.
The 1987 NL scored .175 runs per batting out and had 4,187 batting outs per team. So Gwynn's teammates score 663 runs, and Raines's score 667 runs. That brings the Average Offense Plus Gwynn to 801 runs scored, and the Average Offense Plus Raines to 785 runs scored.
So Gwynn had 16 more BRAA than Raines in 1987. He also played more games, so his BRAR advantage should be even bigger. *And* he was a much better defender.

As I said, it's not even close--Gwynn was a full two wins better.

Except that you aren't making sure all of the Expos and Padres individually add up to team RS. How do the numbers work once you do that?

*And* he was a much better defender.

I'm not buying that. Raines and Gwynn were both 9 FRAA and Gwynn played in more games. What exactly was the difference between a LF/RF, it can't be all that much.

Gwynn had 4 more assists, but also made 5 more errors. Raines made as many putouts as Gwynn despite playing in 17 fewer games.

I have no idea how the pitching staffs were constructed in terms of LH/RH IP, or GB/FB, but the Expos had 1577 IF assists, the Padres 1676. But the Padres whiffed 897 while the Expos whiffed 1012., so it's probably a wash there. The Expos allowed 1428 H in 1450 IP, the Padres 1402 in 1433. So I don't think either's overall D was much better or worse.

Looking LH/RH IP, the Expos had 274.3 LHIP, 1176 RHIP The Padres 329/1104.

Actually, the Expos D was a little worse.

If you look at non-strikeout IP, the Expos D allowed 1428 H in 1113 IP, the Padres 1402 in 1134. So the Padres were a little more efficient on D (BPro backs this up, the average SD pitcher gets a .03 NRA-DERA demerit, the average Montreal pitcher gets a .09 NRA-DERA advantage). This would give Gwynn slightly fewer opportunities. I don't think this is a huge edge though.

My favorite description of the 1987 Cardinals: "Before the national anthem is over, the bases are loaded and Jack Clark is up."

Mine was "Jack Clark and Jack Rabbitts"

Dan the big difference is that RC see Raines at 124 RC, in 372 outs, Gwynn at 128 RC in 402 outs.

> won even more games than they should have

Stolen bases are leveraged and that leverage should be taken into account when calculating their value. Bullpen leverage could be another reason why the Pythags aren't predicting correctly on these teams.

But it's quite obvious that the Expos (and therefore their individual players) were better than their individual stats would lead you to believe, and the Padres were worse.

I don't think this is 'obvious' at all! The Expos may have been luckier, or more 'clutch', or had a more efficient distribution of runs scored, but certainly not obvious that the players were better than their individual stats would lead you to believe, or that the Padres were worse.

I guess that's the point. There's nothing obvious about the '87 MVP thing at all...well, except that Dawson, well, you know. But please, don't tell me that what is not obvious is obvious.

One thing that effected Gwynn's win shares was that Padres played two different seasons. Larry Bowa was in his first season managing the team. They started 12-42. Then finished the year 53-55. This threw the various pythag measures out of whack. They were very inefficient for the first two months of the year - especially May.

Wow, Joe Dimino--you *really* buy James's approach in Win Shares. I think that in *extreme* cases, such as the 1890's Boston Braves, where teams *consistently* outperform their component stats in run scoring far beyond the level of statistical significance, some adjustment to a standard run estimator should be made. But the overwhelming majority of the time, over/underperformance of RS relative to component stats is pure luck. You'd have to show me *extremely* compelling evidence that the 1987 Expos didn't just get lucky at stringing their hits together before I'd concede any advantage to Raines for that. My runs created estimations are done by using XR so that the whole league-season adds up, rather than any given team. You will be using 12 different run estimation formulas for 12 different teams, and, well, that just seems to me to be a terribly misguided approach.

As for the defense, why are you relying on FRAA when we have play-by-play data for 1987? I'm not arguing that Dial's Zone Rating numbers are the Holy Grail, but they do show an *extremely* high correlation to UZR. And they have Gwynn at +16, Raines at -3. I don't know why you'd even pay much attention to things like handedness of the pitching staff when actual PBP metrics are available.

I don't know why you'd even pay much attention to things like handedness of the pitching staff when actual PBP metrics are available.

This is true for zone rating in later years, but 1987 was the first year they recorded it and the data quality is not trustworthy. I'll have to check when I get home, but if you total the ZR for all the players on the 87 Padres, they come out as one of the best defensive teams of the last 20 years.

1989 on looks pretty good, but I wouldn't play much attention to ZR based fielding for 87 or 88.

Well, as usual, I don't know what WS is smoking--I don't even see how Raines is in the top five. Gwynn's OPS+ was 9 points higher than Raines's...I love Tim Raines as much as anyone, he's a slam-dunk Hall member, but there's no way he was as valuable as Tony Gwynn in 1987.

Gwynn hit .302/.468/.389 in 174 PA with RISP and .366/.451/.496 "late and close", Raines hit .336/.500/.454 in 164 PA with RISP and .355/.459/.618 "late and close".

Wins shares takes situational hitting into account as well as actual wins.

Anyway- quick and dirty winshares style analysis:

The 1987 Padres won 65 games- meaning that 195 winshares are divided among the Pads players
The 1987 Expos won 91 games- meaning that 273 winshares are divided among the Pads players.

Gwynn provided 19.7% of the Pad's offense pursuant to runs created (assuming that winshares allocates 50% of wins to offense) that gives Gwynn 19.7% times 195 divided by 2: which gives approximately 19 offensive winshares for Gwynn

Raines provided 16.5% of the Expos's offense pursuant to runs created, that gives Raines 16.5% times 273 divided by 2: which gives approximately 22.5 offensive winshares for Raines

very roughly that's how winshares works

But the overwhelming majority of the time, over/underperformance of RS relative to component stats is pure luck.

Not surprisingly, Dan, I disagree with your logic. Lots of things that happen in baseball are pure luck. Whether a player hits .310 or .320 in a given season can be pure luck. Whether a player has 400 assists or 415 can be pure luck. It doesn't matter whether luck caused something to happen - all that matters is that it happened. The logic behind win shares is that regardless of what caused it, the Expos won 91 games, and someone deserves credit for those 91 games, which always struck me as a reasonable approach. I don't agree with all aspects of win shares (particularly on the defensive end), but I generally believe in offensive win shares.

I'm also much more skeptical of defensive statistics in general, and I do not believe that Gwynn was really 19 runs better defensively than Raines in 1987.

"Wow, Joe Dimino--you *really* buy James's approach in Win Shares"

Absolutely - it's all about value. I don't care too much about ability.

"As for the defense, why are you relying on FRAA when we have play-by-play data for 1987? I'm not arguing that Dial's Zone Rating numbers are the Holy Grail, but they do show an *extremely* high correlation to UZR. And they have Gwynn at +16, Raines at -3."

I don't trust the scorers much, considering I do scoring, I know how subjective it is. Especially on things like how hard the ball was hit. I think both metrics are important. UZR also doesn't account for positioning, which is very important (especially with outfielders). It's not close to the Holy Grail, IMO.

Whether a player hits .310 or .320 in a given season can be pure luck.

Whether a .300 hitter hits between .280 or .320 in a given season can be pure luck.
That's a one-standard deviation band around .300 (assuming a binomial distribution).

Amazing how little Jack Clark material there is in this thread, given that he's neither:
- a slam dunk, or
- not worth considering.

Then again, maybe everyone else is struggling like me.

Some time before the next election I'll be back with a comparison of Jack Clark to Ken Singleton. I'm not ready with it quite yet. But the upshot: even granting Singleton's one obvious advantage over Clark, which is in-season durability, I really don't understand why Singleton has so much more voter support than Clark does.

Is Ryan Klesko the modern day Jack Clark? Albeit slightly worse.

Player BRAR BRAA FRAR FRAA
JClark 694 458 128 -34
Cepeda 707 446 123 -49

That's about as close as you get in similarity.

I have JClark as better than Singleton and Cepeda, and voted for Clark only this year, OCF.

When all is said and done, Clark is too much like John McGraw and Frank Chance. There haven't been a lot of moderns you can say that of. But the missed time ends up really diluting his pennant impact, at least at a glance.

Even so, I should add that J. Clark is HoVG. According to the Reputation Monitor, he scores about a 137 which is bracketed by HoFers Ross Youngs and Harry Hooper, and by candidates Rocky Colavito and Mike Tiernan (immediately above) and Roger Maris, Bobby Murcer, Rusty Staub, Gavy Cravath (w/o MLE credit) and Ken Singleton. Not bad company and about 30th among ML RFers. And again, this is the RM which is only for assembling consideration sets, not my final answer.

since sunny brought it up... :)

player.. career WS car G WS/162gm
Jack Clark..... 316 .. 1994 .. 25.67
Frank Chance 237 .. 1287 .. 28.83
- diff - .......... 79 .... 707 ... 18.10

And of course Win Shares underrates Chance's superior defense at 1B (the Cubbies maxed out on their def score many years), and in fact undervaleues 1B def in general in the deadball era. Win Shares also doesn't take into account 'peerless leader on a truly great club' vs 'surly disposition, bounced from team to team'.

My choice between the two of these is clear.