Baseball for the Thinking Fan

Login | Register | Feedback

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

Monday, February 20, 2006

Nellie Fox

Eligible in 1971.

John (You Can Call Me Grandma) Murphy Posted: February 20, 2006 at 12:53 AM | 113 comment(s) Login to Bookmark
  Related News:

Reader Comments and Retorts

Go to end of page

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

Page 1 of 2 pages  1 2 > 
   1. John (You Can Call Me Grandma) Murphy Posted: February 20, 2006 at 01:00 AM (#1868542)
Is there a picture of Fox without a 10-pound piece of chew popping out one of his cheeks?
   2. TomH Posted: February 20, 2006 at 05:49 PM (#1869317)
looks to me like he may suffer the same HoM fate as Mr. Schoendienst - a single hitting glvoe man, payed for a winner, all of which the HoF luvved, but in my estimation, it sure is tough to place him above a backlog of Childs/Doyle/Monroe/Gordon/Doerr
   3. sunnyday2 Posted: February 20, 2006 at 09:04 PM (#1869575)
It's gettin' to be a reg'lar glut around the keystone sack.

15. Fox 304/32-30-28/128/20.8/A
16. Gordon 242/31-28-26/134/21.1/A without 2 years WWII credit, 289+ with (depending on how you view '46)
18. Doerr 281/27-27-27/127/22.6/A without 1 year WWII credit, 308 with
19. Lazzeri 252/30-27-24/115/23.5/C
20. Doyle 289/33-29-28/130/26.5/C+, or 295 with WWI credit
24. Myer 258/33-24-23/115/21.7/B-
25. Evers 268/28-27-27/117/24.3/A-, 270 adj. to 154 games
26. Childs 238/32-31-27/127/26.5/B+, about 268 adj. to 154 games
28. Schoendienst 262/27-26-25/108/19.2/A, 261 adj. for 1945
89. Dunlap 165/38-17-17/100/27.7/A-, 229 adj. to 154 games and UA discount

I think James is close to right--i.e. Fox is no way comp to Schoendienst. He's got Lazzeri too high IMO and Childs (needless to say) too low. But first, Fox.

The Big 4

Fox 32-30-26*-25-22-22-22-21-21 = 9 seasons of 20+ (*TNBLHBA shows a 28, Win Shares does not)
Gordon 31-26-25-25*-25*-25*-24-24-24 = 9 seasons of 20+ (*WWII adj or XC)
Doerr 27-27-25.5*-24*-25-24-23-22-21 = 9 seasons of 20+ (*WWII adj or XC)
Doyle 33-29-28-27-25-22*-21-20 = 8 seasons of 20+(WWI adj to 154 games)

Defensive WS (*adj for WWI or WWII [discount and/or XC])

Fox 112
Gordon 84*
Doerr 92*
Doyle 50*

The fact that Fox was a slap-hitting singles guy doesn't show up directly in any of these numbers. Roy Sievers would be happy to have posted this kind of a career. If he had exactly these numbers, would he be better than Nellie?

On career you could argue Doerr if you add in a WWII year, but even then Fox is right there. With the glove he is right there for his letter grade and accumulated by far the most career DWS. For a WS peak he is right there--Gordon leads by 6 and Doyle by 2 for 5 years, but Nellie is tied for first for 3 years.

Finally on my totally idiosyncratic Reputation Monitor:

Fox 201
Doerr 182
Gordon 180
(Aparicio 174)
Doyle 167
Schoendienst 153
Dunlap 150
Lazzeri 143
Childs 141
Evers 136

Again, Fox is no Red Schoendienst.
   4. John (You Can Call Me Grandma) Murphy Posted: February 20, 2006 at 09:10 PM (#1869583)
Again, Fox is no Red Schoendienst.

Seconded.
   5. Chris Cobb Posted: February 20, 2006 at 09:21 PM (#1869602)
In win shares, Fox holds his own with Gordon and Doerr.

In WARP, he doesn't. There, he is closer to but better than Schoendienst, whom WARP likes better than WS.

How well he does on each ballot will depend upon the what systemic lens the voter uses for viewing, becuase they offer v. different images of Fox's merit.
   6. karlmagnus Posted: February 20, 2006 at 09:52 PM (#1869628)
Fox, with an OPS+ of 93, looks like Rabbit Maranville to me. WS doesn't seem to distinguish adequately those players who could actually hit.
   7. Mark Shirk (jsch) Posted: February 20, 2006 at 10:02 PM (#1869644)
I think that Fox will be a strong ballot candidate for me. I had Gordon at #10 and Doerr at #12 in 1970 and I dont' think he was as good as either of them. However, I can't say that there is much difference and he will likely fit between #12 and #20.

What seperates Schoendienst from Fox is that Fox has a few years where he was an MVP candidate and Schoendienst doesn't.
   8. Chris Cobb Posted: February 20, 2006 at 10:03 PM (#1869646)
OPS+ underrates Fox somewhat as a hitter, since he is of the "above average OBP, below averager SLG" type.

But I agree that Doerr and Gordon were both much better hitters: their OBP+ was pretty comparable to Fox's, and they had some pop in their bats, which he did not. (Childs and Doyle were better hitters, too.)
   9. Mark Shirk (jsch) Posted: February 20, 2006 at 10:05 PM (#1869649)
Honestly, unlike Tom, I am a little afraid of Fox turning into RIchie Ashburn. Both had fairly broad skills, both had decent peaks and a good amount of career value and it is my belief that most people will see him as ballot worthy. He may be HOM worthy, but I must admit that I would rather wait a few years on him.
   10. Chris Cobb Posted: February 20, 2006 at 10:11 PM (#1869652)
I think there's little danger of Fox going in like Ashburn. Ashburn needed a handful of #1 votes, which he got from his excellent WARP3 numbers, to slide in to HoM safely. Fox may land well in the backlog, but I don't see him getting the elect-me votes that Ashburn did. He doesn't have a case for that by any measure. He'll place, at best, in Doerr/Gordon territory.
   11. John (You Can Call Me Grandma) Murphy Posted: February 20, 2006 at 10:35 PM (#1869680)
Fox, with an OPS+ of 93, looks like Rabbit Maranville to me. WS doesn't seem to distinguish adequately those players who could actually hit.

No, it picks up that Maranville's OPS+ is 11 points below Nellie's just fine.
   12. AJMcCringleberry Posted: February 20, 2006 at 11:29 PM (#1869723)
A better version of Maranville was my first thought when looking at Fox. I think I'll have Fox about 15 spots ahead of Rabbit, but just off the ballot.
   13. Daryn Posted: February 21, 2006 at 12:55 AM (#1869807)
Does a lukewarm reception of Fox mean Aparicio is as good as gone? I like them both.
   14. John (You Can Call Me Grandma) Murphy Posted: February 21, 2006 at 12:59 AM (#1869815)
Does a lukewarm reception of Fox mean Aparicio is as good as gone? I like them both.

Let me put it this way, Daryn: I think Aparicio hopes he gets a lukewarm reception from us. :-)
   15. Mark Shirk (jsch) Posted: February 21, 2006 at 01:41 AM (#1869852)
Not only does WS pick up those 11 points of OPS+ very well but it also recognizes that Fox's OPS is heavily weighted toward OBP, a stat in which he was above average and that Maranvile was below average in both OBP and SLG. Taking a stat like Eqa, which does a better job of weighing OBP and SLG evenly, Fox is at .261 when adjusted for season (i.e. WARP1) and .257 when adjusted for all-time (i.e. WARP 2/3). Maranville is at .248/.240. .260 is average so if you tried to put make this look like OPS+ their numbers would be 100/99 for Fox and 95/92. Not sure how instructive it is to use a made up stat like Eqa+ but Fox is definitely the better hitter and in fact was an average offensive player to add to his stellar defense.

And this is only while looking at career rate stats, which I don't like much anyway. Fox had the higher peak as well (A reason why he was much better than both Schoendienst and Maranville).
   16. DavidFoss Posted: February 21, 2006 at 02:10 AM (#1869880)
(A reason why he was much better than both Schoendienst and Maranville)

Better than two guys that get almost no support here is not that helpful to me.

Nellie's Win Shares number demonstrate that the defensive spectrum has definitely shifted -- especially at second base. He's a throwback to an inductee like Bobby Wallace. How great of a defensive player was Nellie Fox?
   17. karlmagnus Posted: February 21, 2006 at 02:20 AM (#1869888)
Wallace was a MUCH better hitter, with an OPS+ of 105. As I said, WS overvalues these long career defensive players who can't hit. Indeed, I'm not sure it's a useful metric at all.
   18. John (You Can Call Me Grandma) Murphy Posted: February 21, 2006 at 02:32 AM (#1869904)
karlmagnus, I'm pretty sure the HOF didn't use Win Shares when they elected Fox. ;-)
   19. sunnyday2 Posted: February 21, 2006 at 02:59 AM (#1869928)
Nellie Fox couldn't hit 2,663 times. Granted he couldn't hit but 502 XBH so there were a lousy 2,161 singles that prove he couldn't hit. His OBA wasn't much either at .348 though there were those .402, .380, .372 and .372 years. And the years he scored 111, 100, 109 and 110 runs prove something, I don't know what. Maybe he couldn't run either.

Also couldn't stay in the lineup, only played 150 games 11 times and led the league in games 5 times (4 in a row). BI also includes hits 4 times, 3B once, and 2B FA 6 times. He even slugged .400 3 times.

Fox .288/.348/.363/94 in about 10,000 PA
Aparicio .262/.311/.343/82 in about 11,000 PA
Wallace .268/.332/.358/106 in about 9400 PA (adj. to about 9,700)
Schoendienst .289/.337/.387/93 in 9100 PA
Maranville .258/.318/.340/82 in about 11,000 PA (adj. to about 11,600)

Wallace is in the HoM.
   20. jingoist Posted: February 21, 2006 at 03:52 AM (#1869966)
I concur sunnyday2; posters are giving Nellie short shrift.

Name me a better 2B-Man from 1950 to 1965? I don't claim he was the very best every single year running, but I'll bet he was solidly near the top every year. I don't have the reference manuals to look that up; perhaps someone else has the data.

There must be some imputed value to having a guy play 15 solid years, day in and day out (he missed 70 games from 1950-1964) be in the top ten in hits for 10 years; games and at bats for 10+ years and be a 12 time all-star.
I don't know what his fielding rating is (letter grade) but BBref shows him way above league average in fielding and range.

Was he Maz with a glove?
As close to Maz as they come I'd say.
Fox's fielding stats compare quite favorably and I can recall they were generally referred to as equals during their prime even though Fox was a much better hitter who drew substantially motre walks.

That said I don't see this electorate smiling on Nellie and maybe they're right but he sure doesn't belong on the slag heap somewhere.
   21. DavidFoss Posted: February 21, 2006 at 04:21 AM (#1869981)
I don't think that Fox was Maz with the glove, but 1619 DP is pretty darn good. Maz's obscene DP totals put him in the HOF, but we rank players on total value here and Fox's bat ranks him better than Maz, in my opinion.

The big question is whether Fox's fielding, in-season durability, and relative longevity ranks him above guys like Doerr and Gordon. Those guys could flat out rake compared to Nellie (and have favorable WS letter grades themselves).

I've always been a fan of earlier guys like Doyle & Childs as well, but I acknowledge that with the spectrum shift its tough to directly compare them to later players.
   22. Chris Cobb Posted: February 21, 2006 at 04:45 AM (#1869995)
On fielding: win shares sees Fox as one of the best defensive second basemen ever, somewhere between #2 and #6 all-time, depending on how you figure it.

He is #2 all time in defensive win shares at second base, trailing Bill Mazeroski by .7 win share (Maz has 112.2, Fox 111.5).

He is #3 all time in defensive innings at second base with 20,127, trailing only Eddie Collins and Joe Morgan.

Among the 18 second basemen with 15,000+ defensive innings, here is how he ranks by rate:

1) Bill Mazeroski, 6.13 ws/1000 innings (A+) in 18301 innings
2) Bobby Grich, 5.68 ws/1000 innings (A) in 15098 innings
3) Red Schoendienst, 5.61 ws/1000 innings (A) in 15782 innings
4) Frank White, 5.58 ws/1000 innings (A) in 17852.2 innings
5) Nellie Fox, 5.54 ws/1000 innings (A) in 20127 innings
6) Frankie Frisch, 5.42 ws/1000 innings (A) in 15483 innings
7) Bid McPhee, 5.25 ws/1000 innings (A) in 18789 innings
8) Willie Randolph, 5.23 ws/1000 innings (A) in 18648 innings
9) Bobby Doerr, 5.23 ws/1000 innings (A) in 16377 innings
10) Ryne Sandberg, 5.16 ws/1000 innings (A) in 17241

For purposes of comparison, Joe Gordon has a rate of 5.19 in 13396 innings.

WARP sees Fox as excellent, but not perhaps quite so high in the all time rankings. Here's the same top 10 from WS, with their WARP1 FRAA/FRAR and WARP2 FRAA/FRAR

Mazeroski has 224/690, 237/604
Grich has 129/531, 114/451
Schoendienst has 100/506, 101/442
White has 164/607, 150/528
Fox has 73/513, 64/467
Frisch has 189/700, 162/470 (+ play at 3rd base)
McPhee has 340/1153, 218/598
Randolph has 96/537, 83/455
Doerr has 134/544, 131/459
Sandberg has 72/508, 67/429

Gordon has 25/359, 24/294

I'd have to check the next 10 by win shares to see if WARP views any of them as better than Fox. He looks like he'd be 8-10 in this group, but he might be as low as 12-15. That's still a great fielder, but as (for his career) a major-league average hitter, he needs to be nearer to #1 among second basemen to have a strong HoM case, I think.
   23. Brent Posted: February 21, 2006 at 04:54 AM (#1870004)
His defensive reputation--he won the first AL Gold Glove award in 1957, his age 29 season, and won 2 additional awards before Bobby Richardson took his place as the top defensive 2B in the league.
   24. Brent Posted: February 21, 2006 at 04:58 AM (#1870011)
Oops - a correction. The first Gold Glove awards in 1957 covered both leagues. Thereafter separate awards were given in each league.
   25. Paul Wendt Posted: February 21, 2006 at 05:30 AM (#1870044)
His career is merely "longish" in this forum, about 15 full and one half-season, about the same as Bancroft(ss), Schoendienst, and Frisch.

Contrast Aparicio, 18 full. Maranville 17 full and two half. Wallace at least 15 full and one half at shortstop plus one and a half at pitcher; 148 games in his other seven seasons. Davis, 18 full. Dahlen 17 full and two half. McPhee (the only 2B in this paragraph) 17 full and one half. Lajoie 16 full and four half.

Wallace was a MUCH better hitter, with an OPS+ of 105.
about 114 in his 13 seasons playing 85% of team games. {92 66 89} in the other three shortstop seasons that I counted among 15 full and one half, 1287 PA. Really bad in 1894-96 and 1913-18, 703 PA.
   26. Mark Shirk (jsch) Posted: February 21, 2006 at 05:59 AM (#1870065)
It can be a little dangerous to simply say that Fox was better than any other 2Bman between 1950-1965. For instance, the same can be said of Joe Sewell and I guess that Frank Chance was the best 1B of 1900-1910. Neither are terribly close to induction and rightly so.
   27. TomH Posted: February 21, 2006 at 01:40 PM (#1870251)
since the WS rate figures in the NBJHA have been shown to be frequently in error, can someone post for me the no-kidding-real WSper162 ##s for Doyle, Childs, Fox, Gordon, and Doerr? Many thanks.
   28. Dr. Chaleeko Posted: February 21, 2006 at 05:02 PM (#1870394)
Here's a little miniessay I've been working on with regard to Fox. Most of what's in it has been covered, but I thought it might help the discussion:

Value comes in many packages, some big, some the size of Nellie Fox. Players like Fox often seem to be characterized by what they don’t do more than by what they do because their key skills defy direct measurement. Here’s what Nellie Fox didn’t do:
•hit for any kind of power at all—he exceeded the league’s slugging average only three times during his career
•walk even at the league average rate—he only walked about 50 times a year, or in 7% of plate appearances
•steal bases with anything resembling proficiency—76 out 150, ouch!

But Nellie Fox helped his teams an awful lot by not doing certain other things such as:
•missing the ball—he’s the toughest strikeout since Lloyd Waner retired, once every 42 at bats
•missing games—not counting his three token years, in sixteen years he averaged 152 games (in 162-game notation).
•avoiding pitched balls—he was hit almost three times as often as the average player in his day
•grounding into double plays—he hit into 20% fewer than the league average
•letting defenses get easy outs by him—by putting the ball in play all the time and taking extra bases at every opportunity, he stretched the defense and made them work to get him, leading him to a .288 career average.

Of course Fox’s most memorable attribute was his defense. The Gold Glove was established in 1957, and Fox won three of the first four awarded. Good choices, he had been the best defensive second baseman in the AL since 1955 and either he or Bobby Avila had been the best since 1951. Fox’s defensive excellence lasted into the 1960s, and as late as 1963 he was the second-best defender in the league. In all Fox had a marvelous defensive prime from 1951 to 1960 and declined more gracefully in the field than at the plate. He straddles the in/out line, and depending on how you see Ross Barnes and Pete Rose, he’s on the good side of the line. Or depending on your view of the reliability of fielding statistics, he’s on the bad side. Not bad for a mighty-mite who couldn’t hit it past the right fielder.
   29. DavidFoss Posted: February 21, 2006 at 05:24 PM (#1870415)
steal bases with anything resembling proficiency—76 out 150, ouch!

I saw this. The attempts for most seasons are quite low. These look like botched hit and runs to me. (both the successes and the failures).
   30. TomH Posted: February 21, 2006 at 05:35 PM (#1870434)
a fair summary, doc. It reminds me of another player in our backlog, so riddle me this, Batman--

a player who is very similar to Nellie, in that he was a very good defender at a key position, didn't have mcuh power or speed, and was tough to strike out.

however, he has one very big advantage over Nellie - his bat. His career was a little shorter, but even if you eliminate Nellie's worst years, this guy's OBA and SLG relative to the league clearly exceed Mr. Fox's.

League strength? Don't see a huge difference.
Defense? Great 2Bman vs very good shortstop. Similar.

OPS+, RCAP, WARP, WS, all of the stars are aligned the same way.

So if anyone intends to place Nellie Fox above Joe Sewell on their ballot, I'll be real interested to hear them 'splain why.
   31. Dr. Chaleeko Posted: February 21, 2006 at 05:54 PM (#1870451)
FYI, Fox is rated by WS as the top 2B in the AL five times, 1955-1957 and 1959-1960.

TomH, I think your question gets to a larger question that you might call HOM replacement level. If equal representation is a goal, then there's a line to be drawn at each position at about #18 (assuming that pitchers end up around 30% of the HOM). Sewell falls further below that 18th slot at SS than Fox does the 18th slot at 2B. I see Fox between 20th and 25th at his position and Sewell around 25th-30th. Of course the closer you are to this line (on the in or the out side of it), the more likely your selection is to be iffy, or the more likely that individual voter preference is to decide your case. The further from it (whether in Ruth's direction or Enzo Hernandez's) the less subject your case is to rethinking.

Anyway, the bigger point is that there's more excellent shortstops than 2Bs in my opinion, so I can understand why someone believing in equal representation for all positions could put Fox as few as 1 or as many as 15/16 ballot slots ahead of Sewell.
   32. Daryn Posted: February 21, 2006 at 07:02 PM (#1870517)
I will have Fox at 8 and Sewell at 24 -- not that far apart really. The defense makes the difference for me -- and the MVP and all-star games which seem to indicate he was very well thought of at the time despite his slap hitting limitation. Ozzie Smith's OPS+ was 87 and he is also considered to be one of the five best fielders at his position all-time.
   33. Daryn Posted: February 21, 2006 at 07:17 PM (#1870533)
or as many as 15/16 ballot slots ahead of Sewell.

Phew, I just made it within the realm of reasonableness.
   34. Dr. Chaleeko Posted: February 21, 2006 at 07:35 PM (#1870549)
Daryn,

A close call for you! The penalty for not being within my realm of reasonableness is extraordinarily stiff! You'll be forced to comb through all of Karlmagnus's HOM posts to find out how many times he uses the phrase normalized hits!

; )
   35. Trevor P. Posted: February 21, 2006 at 07:43 PM (#1870557)
There's an issue I've been grappling with for the past few elections: does trying to determine the relative merit of Player A in comparison to player B on the basis of how they rank in an all-time context at their position not somewhat violate the spirit of the HOM? In order to make those determinations, you'd have to look at players not yet eligible, no? Maybe it's been my misunderstanding - and I admit I can't find any precise language in the constitution to this effect - but I've been striving hard not to make such comparisons as I thought we were encouraged not to do so.

To say (and I'm not directing this at you specifically, Dr. Chaleeko, but your comparison is the one I have in front of me) that Fox is one of the 20 best 2B, and that Sewell is only top 25 among shortstops, and to then use that as a reason to rank Fox ahead of Sewell - that sort of rationale presumes knowledge of players yet to come, no? Or am I being overly obtuse in my understanding of how we should be voting?

It's not a major sticking point for me, but one I thought might raise some interesting debate in terms of voting philosophies.
   36. Daryn Posted: February 21, 2006 at 08:35 PM (#1870631)
There's an issue I've been grappling with for the past few elections: does trying to determine the relative merit of Player A in comparison to player B on the basis of how they rank in an all-time context at their position not somewhat violate the spirit of the HOM? In order to make those determinations, you'd have to look at players not yet eligible, no?

Yes, you are definitely right and I don't believe the good doctor uses that type of comparison in his actual rankings -- it is just a shorthand for generally identifying where the borderline is if you employ any type of quota system.

That said, the same calculation can be done up to 1965, where the number 10 or 12 is the cutoff rather than 18 at each position. It has the same effect on Fox and Sewell (though I guess there have been a few more great ss in the last 40 years than great 2bs).
   37. DL from MN Posted: February 21, 2006 at 09:08 PM (#1870680)
Part of my problem with great fielding 2B is the fact that they played 2B. If they were such outstanding fielders, why didn't they play SS where they would have more value? If your arm is too weak to play the more important defensive position, then your bat had better make up for the weaker arm. I don't see much bat at all from Fox. He's the only guy in my consideration set with a negative BRAA. As such I have him buried in the 60s.

What's interesting is I have him slotted right next to Larry Doyle. I can't think of two more different ways to put up the same value.

My guess is FRAA if off low since it only gives him 65. However, the way I have things set up he would need 300 additional FRAA to make my top 15.
   38. jingoist Posted: February 21, 2006 at 09:12 PM (#1870688)
Good question: just what exactly is a normalized hit?
   39. Dr. Chaleeko Posted: February 21, 2006 at 09:32 PM (#1870716)
Trevor

I agree with your interpretation of my Fox v. Sewell rankings would be unconstitutional, but as Daryn suggested, I've used that idea (this ad other times) to try to offer a sense to the group of where I see the candidate in the whole of things and sometimes to offer some ideas of players with comparable careers, value, or skills.

When it comes to my ballot I look at where each candidate ranks among the eligibles and HOMed only.
   40. Daryn Posted: February 21, 2006 at 09:45 PM (#1870728)
If someone has 154 hits in a 154 game season, he has achieved 162 normalized hits.

Jack Beckley has more than 3200 normalized hits and more than 2900 real ones.

Don't tell Roberto Alomar that great fielders don't play second base. That guy could pick and throw. I assume, given his accolades, that Fox could too.
   41. Mark Shirk (jsch) Posted: February 21, 2006 at 11:11 PM (#1870836)
There are some 2B that played 2B because their team already had a great SS (Robinson/Reese comes to mind). Also, 2B does have one unique feature to it and that is the pivot on the double play. For that reason I would posit that Mazeroski was more valuable as a 2B than as a SS, even if he would have been a decent SS. SS's don't have to pivot on the double play.

I will have Fox somewhere between 12 and 20 and Sewell is at #40, so I may need to take a look at this. I have Fox so high because I have him just under Gordon Doerr, who will be #'s 10 and 11. Maybe I should look into dropping all three. Sewell is so low because his WS #'s dont' say that he was much better than Bancroft, Long, or Tinker. His WARP #'s (along with OPS+ and other stats) are what seperates him 20+ spots from that group.
   42. Chris Cobb Posted: February 21, 2006 at 11:21 PM (#1870849)
Sewell is so low because his WS #'s dont' say that he was much better than Bancroft, Long, or Tinker.

Do Gordon Doerr's WS #'s, as you see them, say that they were much better than Bancroft, Long, or Tinker?

That's not meant to be a challenging question, but a curious one.
   43. Chris Cobb Posted: February 22, 2006 at 12:02 AM (#1870890)
since the WS rate figures in the NBJHA have been shown to be frequently in error, can someone post for me the no-kidding-real WSper162 ##s for Doyle, Childs, Fox, Gordon, and Doerr? Many thanks.

Alphabetically:

Childs 238 in 1456 = 26.48/162
Doerr 281 in 1865 = 24.41/162
Doyle 289 in 1766 = 26.51/162
Fox 304 in 2367 = 20.81/162
Gordon 242 in 1566 = 25.03/162
   44. Chris Cobb Posted: February 22, 2006 at 04:31 AM (#1871128)
In post 22 above, I did a first study of Fox’s fielding. A brief inspection of win shares shows that it views him as one of the top half-dozen defensive second basemen of all time.

WARP’s assessment was less clear. In the group that WS shows as top 10 all time, Fox looks to be about 9th by WARP, trailing Mazeroski, McPhee, White, Frisch, Grich, Doerr, Schoendienst, and Randolph, more or less in that order. I have now expanded my study to include all the players with over 15,000 defensive innings at second base by win shares’ count and all the players with ws/1000 defensive innings rates of 5.0 ws/1000 or better and with 10,000+ defensive innings at second.

Here is Fox’s Win-Share career fielding data again, along with that data for these additional 17 players.

Player Innings – fws – ws/1000 ** W1 FRAR/FRAA, W2 FRAR/FRAA

Nellie Fox 20127 – 111.5 – 5.54 ** 72/511, 63/465

Eddie Collins 23637 – 107.5 – 4.55 ** 180/924, 105/575
Joe Morgan 21521 – 90.8 – 4.22 ** 19/579, 20/449
Charlie Gehringer 19343 – 86.8 – 4.49 ** 100/663, 81/471
Lou Whitaker 19061 – 87.2 – 4.57 ** 64/494, 50/432
Nap Lajoie 18263 – 85.7 – 4.69 ** 208/852, 142/506
Roberto Alomar** 17162.1 – 86.8 – 5.06 ** 0/393, -7/391
Billy Herman 16029 – 74.9 – 4.67 ** 67/578, 71/390
Johnny Evers 15061 – 68.6 – 4.55 ** 94/615, 28/329
Joe Gordon 13396 – 69.5 – 5.19 ** 26/358, 25/293
Hughie Critz 12847 – 69.6 – 5.41 ** 36/457, 18/274
Manny Trillo 12796.2 – 67.8 – 5.30 ** 8/344, 6/260
Jim Gantner 12220.1 – 61.5 – 5.03 ** 114/397, 104/349
Ski Melillo 11340 – 56.7 – 5.00 ** 107/449, 89/318
Dave Cash 11285 – 58.4 – 5.18 ** 102/399, 101/326
Glenn Hubbard 11195 – 66.3 – 5.92 ** 109/394, 107/332
Ted Sizemore 10902 – 57.7 – 5.30 ** 50/327, 49/265
Max Bishop 10366 – 53.8 -- 5.19 ** 97/416, 79/286

** WS through 2001, WARP through 2005

By WARP, Collins, Gehringer, and Lajoie all appear superior defensively to Fox; Whitaker and Herman appear similar but slightly worse; Evers, Gordon, and Critz are a notch below; Alomar and Trillo are strictly average fielders; Gantner, Melillo, Cash, Hubbard, Sizemore, and Bishop appear better, but all in significantly shorter careers.

My conclusion of this survey of fielding is that win shares sees Fox as about #5 all time, but WARP sees him as around #15 all time.
   45. Mark Shirk (jsch) Posted: February 22, 2006 at 06:13 AM (#1871186)
With war credit, yes they are. I have Gordon at 300 career vs. 282, 111 vs. 83 in prime and 15 v. 12 in peak. 111 vs. 83 in prime is a big deal and Gordon wins in the other two measures.

oh, my prime score is WS>15 and my peak score is WS>25.
   46. John (You Can Call Me Grandma) Murphy Posted: February 22, 2006 at 02:06 PM (#1871307)
jschmeagol, who are you comparing Gordon to in your post?
   47. TomH Posted: February 22, 2006 at 02:26 PM (#1871310)
WSper162 ##s alphabetically:

Childs 238 in 1456 = 26.48/162
Doerr 281 in 1865 = 24.41/162
Doyle 289 in 1766 = 26.51/162
Fox 304 in 2367 = 20.81/162
Gordon 242 in 1566 = 25.03/162

"One of these things is not like the other, one of these things just doesn't belong. Can you tell me which thing is not like the others before I finish my song."
   48. Mark Shirk (jsch) Posted: February 22, 2006 at 06:31 PM (#1871396)
Sorry John, I was responding to Chris' coment about Gordon/Doerr and Sewell, so I was comparing Gordon to Sewell. Should have specified that. However, I will say that ever since reading this thread I am beginning to think that I have been overrating Gordon/Doerr a bit and they may move down my ballot a little ways.

It does seem that 2Bman would end up a little underrepresented vs. SS if we relied solely on WS (which is my metric of choice). It isn't a quota system, just an observation that WS treats SS a little better than 2B, maybe it is because post-shift 2B seem to have shorter careers due to having their back to the runner so often in DP's. This would extend the careers of players like Sewell and Rizzuto in comparison to Gordon and Doerr.

I am also thinking of a move up into the top 20 for Willard Brown, but I have a bit of a problem with him. While it seems that he could have been an absolute monster in MLB, his walk rates are so low that sometimes I wonder if he may not have been a flop, a kind of Glenallen Hill type of player (not saying that Brown was only as good as Hill here so dont' jump on me). I guess I just see Brown as probably the biggest question mark when it comes to translating NeL performance into MLB performance. However, Gadfly has recently given us some walk data that suggests that Brown could at least walk every once in a while when he absolutely had to (like after a bad season). Is this evidence that maybe Brown would have tried to control the strike zone at least a little had he played in MLB? What do our NeL guys think about this?
   49. Paul Wendt Posted: February 22, 2006 at 07:26 PM (#1871477)
Here is Fox’s Win-Share career fielding data again, along with that data for these additional 17 players.

Player Innings – fws – ws/1000 ** <u>W1 FRAR/FRAA, W2 FRAR/FRAA</u>

Nellie Fox 20127 – 111.5 – 5.54 ** 72/511, 63/465


Make that
<u>W1 FRAA/FRAR, W2 FRAA/FRAR</u>

According to WARP1, Roberto Alomar 0/393 is precisely average in the field, not precisely replacement-level.
   50. John (You Can Call Me Grandma) Murphy Posted: February 22, 2006 at 07:31 PM (#1871484)
"One of these things is not like the other, one of these things just doesn't belong. Can you tell me which thing is not like the others before I finish my song."

Yeah, Fox has by far the most games played of that group, Sesame Street boy. ;-)
   51. John (You Can Call Me Grandma) Murphy Posted: February 22, 2006 at 07:33 PM (#1871488)
Sorry John, I was responding to Chris' coment about Gordon/Doerr and Sewell, so I was comparing Gordon to Sewell.

Then I agree with your post, Mark.
   52. Mark Shirk (jsch) Posted: February 22, 2006 at 07:33 PM (#1871490)
Sorry I thought I was on the ballot discussion thread, I am going to move that last paragraph over there right now. Don't answer it here.
   53. John (You Can Call Me Grandma) Murphy Posted: February 22, 2006 at 07:40 PM (#1871501)
Don't answer it here.

Okay, I wont. ;-)
   54. Chris Cobb Posted: February 22, 2006 at 09:06 PM (#1871615)
Childs 238 in 1456 = 26.48/162
Doerr 281 in 1865 = 24.41/162
Doyle 289 in 1766 = 26.51/162
Fox 304 in 2367 = 20.81/162
Gordon 242 in 1566 = 25.03/162

One of these things is not like the other


One way to make Fox more like the others, and to more fairly evaluate him by rate, is to remove his decline phase, since it is a feature of the others' careers not to have them.

Nellie Fox through 1960: 251 in 1760 = 23.10/162

He still trails the others by this measure, but not nearly so badly as he does when his 4+ decline seasons are included.
   55. DavidFoss Posted: February 22, 2006 at 09:33 PM (#1871663)
Another way to possibly help Fox is to try and somehow incorporate his in-season durability. That becomes a much trickier calculation.
   56. jingoist Posted: February 22, 2006 at 11:24 PM (#1871839)
Determining one players relative value versus anothers seems to be a vey tricky thing to get right.

Tom H in post #47 gives us the "who doesn't fit" equation with the idea that because of the way WS are calculated and he chose high peak/short career players, his 4 candidates with much higher WS/162: Childs, Doerr, Doyle and Gordon are somehow superior to Fox.
Grandma called him on it with the obvious fact that Fox played significantly more games AND innings that any of those 4 players over his career.

Does the fact that you show up every day ready to play, both games of a double-header and almost every inning of every game at a level significantly above average impute a certain degree of value?

Is a player with a high batting peak and league average fielding pct. over a short career, Gordon, more valuable to your team in its quest to win championships or are you better off with a player of more modest hitting accomplishments who almost never misses a day for 15 years, plays spectatcular defense and allows your GM and manager to concentrate on assembling/managing other parts of the team to reach the championship?

Gordon was a key componenet in enabling the Yankees to win 5 pennants in the 7 years he played for them.
Can we say that's value received? I guess we should.
Would the Yankees still have won those pennants with Joe Average playing 2B versus Gordon? Maybe someone smarter than me can manipulate the numbers and state yea/nay. I know that after they traded him to Cleveland and inserted Stirnweiss/Coleman/McDougald at 2B for the next 15 years they still won a ton of World Series without a Joe Gordon in their lineup.

Now you are the White Sox and you discover by 1951 you have a guy who plays terrific defense, is a decent hitter, goes on to miss only 7 games over the next 10 years and you suddenly begin to contend for pennants (six 3rd place finishes; two 2nds and a pennant from 1951-1960.
Are you deriving little/some/most of this value from Fox playing 2B every day or are Minoso, Pierce, Kell and Lollar the guys that are getting you to the series by 1959?

I'd like to think that the Sox benefited a bit more from having Fox in their lineup - began to contend for pennants - that the Yanks got from having Gordon in their lineup.
   57. Mark Shirk (jsch) Posted: February 23, 2006 at 06:48 AM (#1872179)
Um, since when did Gordon only play 'average' defense. Even WARP has his peak defense to be quite spectacular.
   58. Chris Cobb Posted: February 23, 2006 at 07:08 AM (#1872191)
Another way to possibly help Fox is to try and somehow incorporate his in-season durability. That becomes a much trickier calculation.

Well, if you want to incorporate in-season durability, just use seasonal win-share totals rather than rates. The trick is to settle on an appropriate replacement level.
   59. TomH Posted: February 23, 2006 at 02:27 PM (#1872304)
whoa, Nellie. Back up the truck.

Maybe I didn't make myself clear; I wasn't trying to denigrate Mr. Fox in post 47; merely point out that he is different; his rate is lwoer, his career is lnger. I did not cherry-pick my favorite players to compare Nellie with - I used the second basemen our backlog; who all HAPPEN to have similar stats. As anyone can see from my ballots, I'm no enemy of long-career players.

As Chris Cobb pointed out, there are ways to attmept to make the comparisons more similar - one is to remove Fox's delcine years, wich then shows that in careers of like length, his rate stats were not quite with the others. And this doesn't even account for the guys who are missing WWII years.
Another point in Fox's favor was his durability. As Chris said, we can look at totals per season to check this For example, still using Win Shares, each of the 5 above 2Bmen's top 3 seasons:
Childs 32 31 27 - 90
Doyle. 33 29 28 - 92
Gordon31 28 26 - 85
Doerr. 27 27 27 - 81
Fox.... 32 30 26 - 88
BUT; WS has a very low baseline, which rewards guys for showing up, even if they aren't very good. Some have suggested that something like 8 WS/year be subtracted out if you want to compare players to a true 'replacement level' (below-avg MLB starter, or decent bench player). This is about 1 WS per 90 PAs. Nellie Fox averaged 90 more PA per year than these other 5 (his top 3 yrs had PAs of 706 718 716 - he really WAS durable), so he gets about a 1 WS advantage each year that is mostly "fluff"; kind of like the pitcher who gets a WS for tossing 20 IP of 5.00 ERA. Take away 1 WS/yr, and Fox's peak is ditinctly less than Childs or Doyle, and even with Gordon's. By this one measure.

Yes, the Yankees won pennants after Gordon left. And they also won after Ruth retired, after Gehrig retired, after DiMaggio retired....these guys were all bums. Obviously only Paul O'Neill and Soctt Brosius were True Yankees :)
   60. TomH Posted: February 23, 2006 at 02:29 PM (#1872306)
and I really need to make fewer typos...
   61. sunnyday2 Posted: February 23, 2006 at 02:50 PM (#1872316)
If we use WS rates, the old-timers will always (always) beat up on the newer guys because it was (indeed) easier to dominate in the old days.

If we use raw WS, whether seasonal or career, the new guys will almost always win (unless you're comparing an inner-circle versus a backlogger) because of longer seasons and longer careers. Everybody knows that Bill James already had a built-in timeline by neglecting to normalize seasons and then put a double whammy on old-timers by adding a second more explicit timeline.

And of course if you normalize season length, then you're now using a hybrid--not exactly a rate stat, exactly a raw stat--but the old-timers will still generally dominate here for the same reason mentioned above.

So when using WS I think you have to do what James did--integrate your career and peak and rates. How you do that may vary.

And what Tom has shown in #59 is only too obvious--that we've got a number of 2Bs who are pretty darn comparable. How they rate will depend more on your analysis than on the numbers themselves. i.e. are you a peaker or a careerist, a timeliner or not, raw or rate, offense versus defense, and so on. You can make any case you want with numbers like theirs.
   62. Trevor P. Posted: February 24, 2006 at 01:24 AM (#1873125)
I wanted to take another look at my 2B rankings, so I re-did a WARP-based analysis of eight 2B that are in my consideration set: Fox, Gordon, Doerr, Cupid Childs, Tony Lazzeri, Buddy Myer, Larry Doyle, and Red Schoendienst.

I made an adjustment for the four pre-war players (Doyle, Lazzeri, Childs, and Myer) by assuming their true WARP values laid halfway between their WARP1 and WARP3. I solely used WARP3 for post-war players.

I then looked at each player's four-year peak, nine-year prime, and career scores. I added up those scores, planning to weight prime at 50%, career at 35%, and peak at 15%.

Everything then got dumped into my superdupermagicnumeralysis machine, and these are the values it spit back:

Prime:
Childs (39.5)
Doerr (38.2)
Lazzeri (37.9)
Gordon (37.0)
Fox (33.1)
Schoendienst (32.0)
Doyle (31.9)
Myer (30.48)

Career:
Doerr (34.6)
Gordon (32.8)
Lazzeri (31.8)
Childs (30.5)
Fox (29.2)
Schoendienst (28.7)
Myer (27.3)
Doyle (26.3)

Peak:
Childs (7.8)
Lazzeri (6.9)
Doyle (6.3)
Myer (5.9)
Gordon (5.8)
Doerr (5.5)
Schoendienst (5.3)
Fox (5.1)
   63. Trevor P. Posted: February 24, 2006 at 01:34 AM (#1873132)
Overall:
Doerr (78.6)
Childs (77.8)
Lazzeri (76.6)
Gordon (75.7)
Fox (67.4)
Schoendienst (66.0)
Doyle (64.5)
Myer (63.4)

It's important to note that due to rounding, some of the final figures may be off by a tenth, but not enough to affect those rankings. Also, there are no war discounts for Doerr or Gordon, and two war years have been added for Gordon.

As to what I think of this system I've concocted: first, the adjustments to the pre-war players suggest that they're all superior in peak to the war/post-war players. To me, that suggests a quirk in my analysis - however, Doyle and Lazzeri did have pretty decent peaks, in all fairness.

Second, I'm shocked at how well Tony Lazzeri does on this measure. I'd always had him beneath Doerr and Gordon, but perhaps he deserves to be up there with them. Perhaps he's been unfairly forgotten?

Third, Cupid Childs packed quite a lot of value into quite the short career - enough to nip Nellie Fox in this ranking system.

Finally, it seems Fox is not that far off from Schoendienst, which to me is a bit of a surprise.
   64. Chris Cobb Posted: February 24, 2006 at 02:02 AM (#1873165)
Second, I'm shocked at how well Tony Lazzeri does on this measure. I'd always had him beneath Doerr and Gordon, but perhaps he deserves to be up there with them. Perhaps he's been unfairly forgotten?

Hmm. . .

I think we have here a case of WARP warping with time. Back when Lazzeri was first eligible, WARP had him, at 76.7 WARP3 (according to DanG's eligibles list). Now he's at 80.8, a non-trivial gain, and there may well have been a larger gain in WARP1, which is now at 101.1. I certainly don't recollect him as a 100 WARP1 player back in the day . . .
   65. Brent Posted: February 24, 2006 at 04:39 AM (#1873324)
I think the problem is Warp's "reverse timeline" for its fielding replacement level. WARP1 (yes, the one that is supposedly not adjusted for "all time") shifts the relative weights given to fielders and pitchers by adjusting the replacement levels over time. Thus, an average pitcher from a more recent time period will have a significantly higher WARP1 score (due to a lower replacement level for PRAR, "pitching runs above replacement"), and a more recent position player will have a lower WARP1 score (due to a high replacement level for FRAR, "fielding runs above replacement"). Thus, WARP1 has a reverse timeline for position players, favoring the average players of earlier eras.

I've tried to figure what BP's WARP1 considers an "average" second basemen to have been worth (in runs per 162 games) during each player's career. I've calculated it as 162*(FRAR - FRAA) / G.

Player-Career Fielding runs / 162 G (of average 2nd baseman)
Childs 52.3
Doyle 47.0
Lazzeri 41.7
Myer 41.3
Gordon 35.4
Doerr 35.9
Schoendienst 33.1
Fox 31.0

So we see that an average 2B of Childs' time was considered to be worth 21 runs / 162 G more than an average 2B of Fox's time.

Some of this effect goes away in moving from WARP1 to WARP3. Users of WARP need to be aware that there are important timeline effects embedded in both WARP1 and WARP3.
   66. Chris Cobb Posted: February 24, 2006 at 05:05 AM (#1873354)
Brent,

That is great data! Thank you!

Some of this effect goes away in moving from WARP1 to WARP3.

All of this effect should go away, since one of the major things WARP2 does is make the difference between replacement level and average constant for each position.
   67. Joey Numbaz (Scruff) Posted: February 24, 2006 at 05:40 AM (#1873389)
Brent, is that saying that an AVG 2B was worth more defensively in 1890-1920 period than in the 1950s and 60s? I that next to impossible to believe. Am I missing something, or misreading or something?
   68. Chris Cobb Posted: February 24, 2006 at 03:55 PM (#1873611)
Brent should correct me if I am wrong, but that is exactly what his data is saying.

There are two important factors to note here:

1) WARP at present does not adhere to the win-share 2B/3B defensive spectrum shift. In 1882, for example, a quick study indicates that WARP views an average second baseman as worth about 42 FRAR in 84 games, while an average third baseman is worth 29 FRAR in 84 games. If these two positions have switched on the defensive spectrum, WARP is systematically overrating early 2B defensively and underrating early 3B defensively.

2) WARP generally finds that the shift in defensive value from fielders to pitchers as the game moved towards its modern configuration is much greater than any shift in defensive value from one fielding position to another, so even if WARP is acknowledging (which I don't thinki it is), that 2B is a more important defensive position relative to other defensive positions in 1950 than it was in 1895, all fielding positions are so much more important that any shift of value to second from other positions is hardly noticeable when the FRAR from the later era are compared to the FRAR from an earlier one.
   69. Guapo Posted: February 24, 2006 at 03:58 PM (#1873615)
As long as you are talking second basemen, I just wanted to say: don't forget Bobby Avila.
   70. John (You Can Call Me Grandma) Murphy Posted: February 24, 2006 at 05:18 PM (#1873711)
...and don't forget the Maine or the Alamo! ;-)
   71. jingoist Posted: February 24, 2006 at 05:23 PM (#1873720)
I think I follow the drift of this conversation......that said, should the electorate revisit the Tommy Leachs and Pie Traynors abd adjust or has that already been done in previuos ballot preparations?
Does that also mean that early 2b-men like Childs should have their numbers reduced accordingly if Warp is used to determine value?
   72. Dr. Chaleeko Posted: February 24, 2006 at 05:25 PM (#1873725)
From these posts and others, I understand why WARP does a lot (not all) of what it does. And I have an understand about why WS does much of what it does. WS is more conservative and probably, therefore, better for making comparisons across eras. But, overall, someone's got to be able to come up with a better way than both systems of evaluating and contextualizing historical fielding value. It won't be me, I'm not that smart or SQL-enabled, but someone out there is brainy and savvy enough to do it.
   73. KJOK Posted: February 24, 2006 at 05:32 PM (#1873736)

I think we have here a case of WARP warping with time. Back when Lazzeri was first eligible, WARP had him, at 76.7 WARP3 (according to DanG's eligibles list). Now he's at 80.8, a non-trivial gain, and there may well have been a larger gain in WARP1, which is now at 101.1. I certainly don't recollect him as a 100 WARP1 player back in the day .


Lazzeri WAS at 84.8 WARP1 when we started this project. This is one of the frustrating things in using WARP, as it changes EVERY year when BP does a re-calculation...
   74. Mark Shirk (jsch) Posted: February 24, 2006 at 06:03 PM (#1873784)
So Lazzeri has seen his WARP1 grow by 17 wins in the span of three years after hsi death? Damn, that man is good.
   75. Chris Cobb Posted: February 24, 2006 at 06:05 PM (#1873789)
17 wins!?! DANG!
   76. Mark Donelson Posted: February 24, 2006 at 06:48 PM (#1873858)
This is the kind of thing that tempts me to dump WARP out the window. (I won't, but it really makes you wonder...)
   77. DL from MN Posted: February 24, 2006 at 07:16 PM (#1873917)
2B WARP3 BRAR BRAA FRAR FRAA G
Fox 83.5 276 -38 467 65 2367
Red 82.0 284 2 442 101 2216

What's the difference? Fox has slightly higher replacement value but Red has the better peak numbers. I can't justify having either in my top 15. Can someone run a Win Shares top 10 seasons for both?
   78. sunnyday2 Posted: February 24, 2006 at 07:26 PM (#1873944)
I'm sure somebody has tried to enlighten me on this before, but I just can't seem to get it.

How does Fox' 467 FRAR (versus Red's 442) translate into 36 fewer FRAA? Is it league strength? That would imply that the R is within the league but the A isn't? Or is it that NL hitters hit the ball harder? Or what?
   79. DavidFoss Posted: February 24, 2006 at 08:07 PM (#1874035)
How does Fox' 467 FRAR (versus Red's 442) translate into 36 fewer FRAA? Is it league strength? That would imply that the R is within the league but the A isn't? Or is it that NL hitters hit the ball harder? Or what?

Those are WARP3 numbers so there is some translation in there. If you check the WARP1 numbers, then I can see what's happening.

Using 2B play only:
Fox is 511 FRAR / 72 FRAA in 2247 Adjusted Games
Red is 469 FRAR / 94 FRAA in 1767 Adjusted Games

That puts the value of an average fielder above replacement at 0.195 R/G for Fox's context and 0.212 R/G for Red's context. (Still a 10% difference, which is a bit odd for contemporaries)

Anyhow, the *main* cause for the swap in FRAR & FRAA rankings is that BP thinks that Red was a better fielder, but that Fox played more.
   80. sunnyday2 Posted: February 24, 2006 at 08:24 PM (#1874068)
>That puts the value of an average fielder above replacement at 0.195 R/G for Fox's context and 0.212 R/G for Red's context. (Still a 10% difference, which is a bit odd for contemporaries)

But why? Because the NL had better fielders? Or because the NL had worse replacement fielders? I mean, we're talking about essentially the same seasons, not like Fox versus Evers or something.
   81. Chris Cobb Posted: February 24, 2006 at 08:59 PM (#1874119)
Re Fox & Schoendienst & WARP:

1) Value above average shifts from WARP1 to WARP2 in response to changes in league quality.

Schoendienst's league is viewed as essentially at the all time standard in terms of fielding quality:
He has 100 FRAA in WARP1 and 101 FRAA in WARP2.

Fox's league is viewed as a bit below the all-time standard in terms of fielding quality:
He has 73 FRAA in WARP1 and 64 FRAA in WARP2.

So fielding quality is viewed as being slightly higher in Red's league.

2) Replacement level for fielding in WARP shifts in relation to the division of responsibility between pitchers and fielders. WARP sees that division shifting steadily towards pitchers from 1945 to 1965. Schoendienst, by virtue of playing a few years earlier than Fox, gets slightly more value above replacement level in WARP1 and therefore loses slightly more per game when he shifts to WARP2.

Not including the 9 FRAA above average that Fox loses in the transition from WARP1 to WARP2, he loses
37 FRAR in 2247 adj. games at second base, a rate of 2.54 RAR/154 games
In WARP1 in Fox's career, an average second basemen was worth (as Brent posted above) 30.1 RAR/154 g

Not including the 1 FRAA that Schoendienst gains in the transition from WARP1 to WARP2, he loses
58 FRAR in 1767.1 adj. games at second base, a rate of 5.05 RAR/154 games
in WARP1 in Schoendienst's career, an average second baseman was worth (see Brent) 33.1 RAR/154 g

In the all-time context, an average second baseman is worth about 27.5 RAR/154 games

It may be the case that this difference is due wholly to Schoendienst playing 1946-1950 before Fox became a regular and Fox playing 1959-64 after Schoendienst became a regular, but it may also be that, because the AL was stronger in pitching in the 1950s than the NL was, value above replacement was shifted to pitchers away from fielders faster in the AL than in the NL. More study would be needed to identify this effect, but as you can see from Brent's chart above, Gordon and Doerr, playing in the 1940s AL, earn 35.5-9 RAR for 154 games of average fielding at second, so this value seems to be shifting at 2-3 FRAR every 5 years, 1940-1965

1940-50 35.5 G & D
1946-60 33.1 Schoen
1951-64 30.1 Fox
   82. Trevor P. Posted: February 24, 2006 at 09:21 PM (#1874159)
So we see that an average 2B of Childs' time was considered to be worth 21 runs / 162 G more than an average 2B of Fox's time.

Okay, I just want to make sure I've got this right: the movement from WARP1 to WARP3 tries to address that defense was considered more important in Childs' time than in Fox'sRight?think I've got that.

But, here's what I don't get: if that's true, why does Larry Doyle go from a 93.1 WARP1 to a 57.3 WARP3, and why does his FRAA go from -154 to -213? If WARP3 is putting things into an all-time context, and Doyle was a horrible fielder during an era when fielding was much more important, shouldn't his atrocious horrible fielding be negated somewhat when making that adjustment?

Sometimes, WARP makes my head hurt.
   83. Mark Shirk (jsch) Posted: February 24, 2006 at 09:52 PM (#1874199)
Trevor,

WARP doesn't try to make Childs defense more orless important in comparison to 3B at all. It only adjusts for era, which negates Childs replacement level advantage.

DL,

Fox ..............33,31,27,26,23,23,23,22,22,20 316
Schoendienst 28,27,26,22,21,20,19,18,17,17 275

Here are your top 10 WS values for Fox and Schoendienst, they aare adjusted to 162 games (5% increase) so others will have slightly different (about 1 WS a year lower) numbers. However, that shouldn't really obscure the fact that by this measure Nellie and Red aren't really in the same ballpark.
   84. Chris Cobb Posted: February 24, 2006 at 09:53 PM (#1874200)
Trevor,

Changes in fielding runs above average from W1 to W2 show league quality. WARP sees Doyle as being a terrible defensive second baseman in a weak league.

An average second baseman in Doyle's league would be 47 fielding runs above replacement per 154-game season, as Brent has calculated. For Doyle, that would be 509 FRAR for his career in WARP1, but he is 154 FRAA below average, so he ends up being only 347 FRAR in WARP1 (some fractional runs are being lost somewhere, but the totals are close).

Going from W1 to W2, Doyle loses 59 runs above average because his league was bad, dropping him to -213 FRAA. He loses those runs off his FRAR total in W2 _and_ he also loses about 19.5 FRAR/154 games for the shift from his defensive context (avg. = 47.0 RAR) to the all-time context (avg. = appx. 27.5 RAR), which shaves off appx. 211 more runs. So his 347 FRAR in WARP1 is dropped by 59 for low quality of play and 211 for the context shift, leaving him with a mere 77 FRAR by my calculation for his career. A check of his player card indicates that WARP2 awards him 85 FRAR, so he actually lost about 18.9 FRAR/154 games rather than 19.5. But (hopefully), you get the idea.
   85. Jim Sp Posted: February 24, 2006 at 10:42 PM (#1874274)
warp trying to improve the numbers is a good thing. it's the complete lack of explanation that is a bad thing.
   86. Dr. Chaleeko Posted: February 24, 2006 at 11:03 PM (#1874305)
warp trying to improve the numbers is a good thing. it's the complete lack of explanation that is a bad thing.

Agreed. And they've got more changing or 'splaining to do before I can buy it: Barry Bonds' value being 17% defensive....something's still not smelling right about that one.
   87. Mark Shirk (jsch) Posted: February 24, 2006 at 11:15 PM (#1874329)
Well, Bonds was a terrific defensive outfielder in his younger days AND they are counting 17% of his value above replacement. If you used average as your baseline I would expect that number to be closer to 5%, if not lower.
   88. Dr. Chaleeko Posted: February 24, 2006 at 11:26 PM (#1874351)
That's assuming that average shouldn't = replacement.... ; )

Sorry I was feeling like a scamp tonight, and opened up the whole Warp replacement level fielding thing.
   89. jimd Posted: February 25, 2006 at 12:18 AM (#1874415)
1895NL 5101 BB - 3602 K - 484 HR - 4603 E
2005NL 5153 BB - 10360 K - 1583 HR - 997 E

I've normalized the 2005 NL numbers to the number of games played in 1895 to make them easily compared.

Clemens is much more in control of Outs (via K's) and Runs (via HR's) than Cy Young ever was. Errors in Young's day were just about as common as Walks were then and now; I'm sure he had innings here and there where the fielders surrendered multiple runs without a hit.

(Just to demonstrate again that fielding was much more important then than now.)
   90. Chris Cobb Posted: February 25, 2006 at 03:10 AM (#1874515)
To follow up on jimd's data, I've tried to go a little deeper into WARP's division of value in 1895

I studied Cleveland, 1895 (the team of Cupid Childs and Cy Young), to see how WARP divided up the difference between replacement level and average. Cleveland played 131 games.

The team was
78 PRAA, 236 PRAR
-6 BRAA, 148 BRAR
41 FRAR, 303 FRAR

By subtraction, an average team would be
158 PRAR (27.5% of Value above replacement)
154 BRAR (26.8% of Value above replacement)
262 FRAR (45.6% of Value above replacement)

Win Shares gave Cleveland
88.7 pWS (20.1 above avg.)
120.6 bWS (26.3 above avg.)
42.7 fWS (9.1 above avg.)

So, an average team had
68.6 pws (34.9% of value)
94.3 bws (48% of value)
33.6 fws (17.1% of value)

At the team level, the systems are dividing value in radically different proportions: it's not surprising that they cast some players in this era in very different lights. What's somewhat surprising is that they evince considerable agreement!

One more tidbit: here's the division of FRAR for an average team in 1895 and in the all-time context in WARP:

Pos 1895/All-time
P 9/8
C 45/32
1B 14/8
2B 44/24
3B 30/17
SS 46/27
LF 29/12
CF 31/20
RF 15/10
   91. Brent Posted: February 25, 2006 at 05:32 AM (#1874588)
Since they're relevant to the discussion of WARP1's timelining of pitchers (and "reverse timelining" of position players), I thought I'd repost a couple of items I posted on the 1967 Ballot Discussion Thread:

Post # 152

Going back to the treatment of pitching in the BP statistics that I discussed in posts # 122 and 123, I decided to add up the PRAR and FRAR, which are the weights for pitching and fielding in WARP1, for all the teams in the 1925 NL and the 1949 AL. I also have figures for fielding and pitching WS:

Year Lg PRAR %_def FRAR %_def
1925 NL 2031 50.2% 2016 49.8%
1949 AL 2933 70.1% 1254 29.9%

Year Lg pit_ws %_def fld_ws %_def
1925 NL 638.3 66.9% 316.5 33.1%
1949 AL 640.1 66.7% 319.1 33.3%

We see that for WARP1 the weight given to pitching increased tremendously between 1925 and '49 (and the weight given to fielding dropped), whereas for WS there was little change in relative importance of the two contributors to defense. Although I think BP is correct that the importance of pitching increased between 1925 and '49 with the increases in fielding-independent components of baseball such as walks, home runs, and strikeouts, I also think they've greatly overstated the magnitude of the shift. On the other hand, the shift doesn't even register in the WS statistics.


Post # 171:

Returning to discussion in # 152-55 on changes over time in BP replacement levels for pitchers...

The question is how much should the relative weight of pitchers increase as their role in defense grew-- due to more strikeouts, walks, and home runs, and therefore fewer balls in play. I'll use the following simplified linear weights formula:

.47 1B + .78 2B + 1.09 3B + 1.40 HR + .33 BB - .25(AB-H)

I'll ignore the contributions of baserunning (defensive responsibility for which would need to be split between pitchers and fielders) and also ignore hit by pitch. I give strikeouts the same weight (-.25) as other outs.

Since components of this formula have both negative and positive signs, I'll weight each component by the absolute value of its coefficient. Using these weights, the purely defense-independent components account for 17.7% of the total in the 1925 NL, and 24.9% in the 1949 AL.

It's surely unrealistic, however, to give fielders 100% of the responsibility for the outcomes of balls in play. If we assign 30% of the responsibility for balls in play to pitchers, the weighted share of the pitcher components increases from 42.4% in 1925 to 47.4% in 1949. If we assign 50% of responsibility for balls in play to pitchers, the weighted share of the pitchers components increases from 58.9% in 1925 to 62.4% in 1949.

As TomH observed in # 155, errors are a bigger problem - they don't figure into the linear weights formula, so I haven't accounted for them in these calculations, although there was a substantial decrease in errors over this period. (Errors and unearned runs also should figure into the discussion of Joss, but I'll hold off on entering into that discussion.) The consensus of researchers is that responsibility for errors also needs to be shared between fielders and pitchers, so I'm not sure that the overall picture will change.

The bottom line still seems to be that win shares fails to adjust for the growing role of pitchers, while Warp over-adjusts. But I don't think we know enough to say what the ideal adjustment ought to be.
   92. TomH Posted: February 25, 2006 at 07:18 AM (#1874662)
on a different track:

we have honored more SS than 2B. Some have posited that maybe it's easier to be above avg (or replacement) at SS, so if we measure all players against avg play for their position, we will wind up with more SS.

Could be. But....

If I take the leaders in RCAP thru 1965 (current for us), the top second basemen beat the top shortstops all the way through - by that I mean, Rogers Hornsby (2B #1) has more RCAP than Honus Wagner (SS #1), etc. I did not expect that.

Maybe WS and WARP somehow adjust for position average differently than RCAP does? Maybe more of us should use RCAP (the sound you hear in the background is KJOK applauding)? Maybe the top SS have longer careers, or are better defensively?

lotsa questions, too tired to create good answers.
   93. TomH Posted: February 25, 2006 at 07:25 AM (#1874669)
for ref, top RCAP (position by career, NOT by individual season)
shortstops
rank player RCAP
1 Honus Wagner 1060
2 Arky Vaughan 598
3 George Davis 452
4 Joe Cronin 431
5 Luke Appling 375
6 Joe Sewell 346
7 Bill Dahlen 291
8 Jack Glasscock 289
9 Lou Boudreau 274
10 Hughie Jennings 263
..
20 Jack Rowe 142
(top 10 all HoM except Sewell hint hint)

second base
rank player RCAP
1 Rogers Hornsby 1094
2 Eddie Collins 822
3 Nap Lajoie 766
4 Charlie Gehringer 581
5 Cupid Childs 354
6 Tony Lazzeri 325
7 Jackie Robinson 323
8 Fred Dunlap 311
9 Billy Herman 298
10 Frankie Frisch 291
..
20 Del Pratt 151
Cupid Childs the highest non-HoMer

third base
rank player RCAP
1 Eddie Mathews 621
2 John McGraw 459
3 Stan Hack 370
4 Bill Joyce 367
5 Home Run Baker 364
6 Denny Lyons 326
7 Bob Elliott 241
T8 Heine Groh 216
T8 Deacon White 216
Mugsy sticks out like a sore thumb

third basemen had a harder time than other throwing infielders putting up the great RCAP totals
   94. OCF Posted: February 25, 2006 at 08:47 AM (#1874694)
What I've got is adjusted RCAA - I am adjusting for the differing value of a run in high scoring versus low scoring times; what I'm not adjusting for is era-to-era changes in the baseline expectations for the position. I don't have all players systematically, but I do have most of the players in TomH's post. I'll list two tables. One is adjusted career RCAA, (in units vaguely akin to wins); the second is a patched-together system that includes a nonlinear bonus for big seasons, and also incorparates RC above 75% of averages, a measure that has a little bit of award for longevity.

The people RCAA (or RCAP) will overrate the most are short-career players with in-season durability problems but terrific offensive rates stats - e.g., McGraw and Chance.

Career adjusted RCAA:

1. Wagner 109
2. Vaughan 53
3. Appling 33
4. Cronin 32
5. Pesky 24
6. Stephens 23
7. Boudreau 23
8. Reese 22 (includes war credit)
9. Sewell 19
10. Wallace 18 (not his whole career; lopped off some negatives)
11. Jennings 18
12. Rizzuto

(Note: I don't have this number for Davis, Dahlen, or Glasscock, elected before I started using this system)

Now the same system with all the extra bells and whistles:

1. Wagner 431
2. Vaughan 187
3. Appling 120
4. Cronin 114
5. Reese 91 (with war credit)
6. Pesky 90
7. Boudreau 85
8. Stephens 84
9. Jennings 77
10. Sewell 75
11. Wallace 69 (truncated)
12. Bancroft 47
13. Rizzuto 47
14. Joost 45

Now the second basemen. First simple career adj. RCAA:

1. Hornsby 101
2. Collins 94
3. Lajoie 78
4. Gehringer 48
5. Doyle 38
6. Robinson 38
7. Evers 30 (yes, Johnny Evers!)
8. Frisch 29
9. Lazzeri 29
10. Childs 28
11. Gordon 27
12. Herman 27
13. Doerr 22

And the full system:

1. Hornsby 405
2. Collins 350
3. Lajoie 297
4. Gehringer 175
5. Robinson 134
6. Doyle 128
7. Frisch 108
8. Childs 104
9. Evers 100
10. Lazzeri 99
11. Herman 98
12. Gordon 96
13. Doerr 84

Gordon and Doerr both have some war credit added in there. I never worked up Dunlap.

And the third basemen; this thim, I never worked up Joyce, Lyons, or White, and I haven't gotten to Mathews yet.

1. Baker 42
2. Hack 40 (includes war deduction)
3. McGraw 37
4. Elliott 34
5. Groh 27
6. Rosen 26
7. Leach 24
8. Clift 18
9. Kell 18
10. Traynor 16
11. Collins 16

1. Baker 155
2. Hack 138
3. McGraw 126
4. Elliott 117
5. Rosen 100
6. Groh 98
7. Leach 91
8. Collins 67
9. Traynor 66
10. Kell 66
11. Clift 66
   95. Brent Posted: February 25, 2006 at 04:38 PM (#1874803)
Mugsy sticks out like a sore thumb

I've commented on the RCAP numbers before that they just don't make sense to me. According to bbref, McGraw is credited with 728 runs created during his career; does that imply that the average third baseman in 1100 games would have been credited with only 269 runs created? That hardly seems possible. And how does one explain McGraw's lead in RCAP over Home Run Baker, who was credited with 953 runs created? Something strange is going on in the calculations - these can't just simply be runs created above the average player at the position.
   96. TomH Posted: February 25, 2006 at 09:31 PM (#1874989)
I agree, Brent - this does require review. I would think using OWP and PA ought to shed light on this.
   97. Mark Shirk (jsch) Posted: February 25, 2006 at 09:47 PM (#1875000)
Isn't the difference between McGraw and Baker BECAUSE Baker played so many more games. Sure he had 200+ more RC, but it was over a longer period of time, so teh average player is going to post more RC over that time.

Same with anything above replacment, the more you play the more runs you need to be above replacement so in 100g the number may be 12 in 162 it may be 20 or thereabouts. S0 two plaeyrs with 21 RC, one playing 100g and the other 162g would then be then have 9 and 1 RCAA, respectively. In fact the latter player could have 28 RC and still not have as many RCAA as the formerplayer because the threshold is higher.
   98. OCF Posted: February 25, 2006 at 11:01 PM (#1875034)
Just a reminder about the one thing I did correct for with my list - that the value of a run differs depending on the overall offensive context. McGraw played in a very high offensive context, in which each run was cheap. Baker played in a much lower offensive context. Even though his peak does correspond roughy with the 1911-1913 lively-ball outburst, that's still nowhere close to the middle '90's.
   99. TomH Posted: February 25, 2006 at 11:09 PM (#1875041)
From Sinins BBencyc, McGraw has 1101 RC, Baker 1027, so there's a big difference there. But I'll use a different method.

player OWP posOWP PA ...RC/game above pos .. 'games' modRCAP
McGraw .727 .495 ...4892 .......2.87 ....................122.3 ......351
Baker.. .659 .462 ...6508 .......2.08 .....................162.7 ......338

McGraw's 2.87 runs per 'game' above avg at his position is calculated by inserting his OWP into the pythagorean formula, assuming a 4.5 r/g environ, and solving for how many runs per game he created, then subtracting this figure from the league avg at his position. I'l send the algebra if anyone wants it.

"Games" is PA / 40.
modRCAP is RC/game above avg pos times games.

By this method, instead of being 85 or so RCAP above Baker, they are now almost even. This is mostly, I believe, due to the high run environment of McGraw's day.

Of course, if you tred to calculate 'replacement' as an OWP somewhere below position average, Bakes comes out ahead. As he should :)
   100. OCF Posted: February 25, 2006 at 11:21 PM (#1875050)
What TomH just did is close to what I've done.

From Sinins BBencyc, McGraw has 1101 RC, Baker 1027

My source was a Stats Major League Handbook. From that, I get McGraw 1026 RC, Baker 1021. So the difference between the outcome of Tom's calculation and mine mostly reflects a different method of computing RC for the 1890's.
Page 1 of 2 pages  1 2 > 

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
Brian
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 1.0493 seconds
49 querie(s) executed