Baseball for the Thinking Fan

Login | Register | Feedback

btf_logo
You are here > Home > Baseball Newsstand > Baseball Primer Newsblog > Discussion
Baseball Primer Newsblog
— The Best News Links from the Baseball Newsstand

Tuesday, February 25, 2014

Ozzie Smith Talks Derek Jeter’s Career and Gives His Shortstop Mount Rushmore

Great Gutzons! Garry Templeton is on Ozzie Smith’s Mount Rushmore of shortstops!

For now, Jeter already has a special place in history in Smith’s eyes. When asked to name his shortstop Mount Rushmore, Smith was quick to include him.

“Well, of course Cal would be on there. Derek would be on there. Omar Vizquel would be on there. The guy I got traded for actually would be one of those guys because he was a true five-tool player,” said Smith.

Also, naturally: “And of course, I’d put myself on there.”

If you’re scoring at home, that’s Cal Ripken Jr., Jeter, Vizquel, Garry Templeton and Smith himself. You’ll have to visualize The Wizard’s shortstop Mount Rushmore on your own, but statistically it looks like this:

Of the numbers up there, only Jeter’s aren’t set in stone yet. For what it’s worth, he is within range of Smith’s career WAR. One last great season in 2014 will put him right there with The Wizard among the greats to ever play shortstop in WAR’s eyes.

...As far as Smith and Budweiser are concerned, Opening Day has gone long enough without being declared an official national holiday. Smith says it might as well be considering that many Americans already treat it like one.

“There are 22 million Americans who at some point in time have played hooky from work and school. So that in and of itself makes it an unofficial holiday,” said Smith of Opening Day. He added that he’s not asking for much, as merely getting Opening Day proclaimed “as some type of day of observance would really fit the bill.”

Smith will be on the campaign trail for the next 30 days as he and Budweiser attempt to collect 100,000 signatures on a petition at Budweiser.com/OpeningDay. And while only fans 21 and older can sign it, the White House is required to respond if the signature quota is met within the 30-day window.

Repoz Posted: February 25, 2014 at 08:16 PM | 177 comment(s) Login to Bookmark
  Tags: history, hof

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 2 of 2 pages  < 1 2
   101. Walt Davis Posted: February 26, 2014 at 04:56 PM (#4663076)
One thing we probably should think through more is the impact of the fact that the number of outs is essentially fixed.

For example, RF is essentially the average number of outs a player/position is responsible for per 9. But by definition the team can only record 27 outs. More outs recorded by the SS (whether due to being good or GB or positioning or pitcher effects or whatever) necessarily means fewer outs recorded at the other positions.

So unless I'm missing something, assuming a standard distribution of BIP, one good fielder means the others will, on average, have a below-average RF/9 regardless of their defensive ability. From an RF/9 perspective, it would be silly to have an IF of Brooks, Ozzie, Maz and Hernandez -- either they'll all look like average fielders or your OF will be "terrible."

Similarly I worry about the linear weight (or other) estimates of the impact of an out. At the team level, from whence the coefficients come, there is very little variation in the number of outs in a season -- due to some teams winning more often at home and random variation in extra-inning, rain-shortened and cancelled games. If we required the home team to bat in the 9th and allowed for ties, every team would record the same number of outs. So in any one season, the model has relatively little power to estimate that effect. Of course over 100+ years that power is probably large enough assuming the coefficient is stable enough.
   102. cardsfanboy Posted: February 26, 2014 at 04:59 PM (#4663079)
Yes, there are definitely is a large differences in possible chances between 1935 and today. But all of those things that affect the number of chances between eras also, to some smaller extent (except for things like GB ratio, which I think may definitely vary more within year than the average GB rate varies between era. 13% more BIPs were GB for Pirates pitchers than for Athletics pitchers (52.5% vs 39.5%). I would be fairly shocked if the average rate of GB had dropped by 13 percentage points since 1935). That is what I was trying to say by comparing between shortstops from 2013.


Just for the record. After adjusting for ONLY Strikeouts, here is the team level range factors.

Tm range (year, 1936)

BRO 23.11
CHC 23.11
PIT 23.39
STL 23.40
LgAvg 23.60
NYG 23.75
CIN 23.98
PHI 24.01
BSN 24.32


And that is only with one variable adjusted out of the equation. Haven't even looked at number of batters faced, adjusted out double plays.
range factor for the individual level is putouts+assists/games played(or per 9 innings played) You really need to adjust out double plays to get a more accurate accounting, since it's effectively being counted twice on the same play.

Rough numbers since I don't have innings played, but in 1936 by my account
Bartell Vaughan
put 317 327
ass 559 477
game 144 156
dp 100 86
rf 6.08 5.15
adj rf 5.38 4.60

And again, we still haven't tried to put team into context.
Vaughan moves from being .93 back in range factor to .78 by just adjusting out the double counting of double plays.
   103. AROM Posted: February 26, 2014 at 05:11 PM (#4663088)
But what you do see is a sort of limit that is reached and managers inherently understand that. Instead of people talking about a "Mendoza line" a better concept would be the "Vaughan line" ALthough that is unfair to Vaughan who was at least VG to excellent for 10 years or so.


This thread is the first place I've ever seen suggesting that Vaughan was a some sort of terrible fielder. Yeah, his career range factor is a bit below average (5.24 to 5.38) but I've got him at +21 runs as seen on BBref. Humphreys' data as posted above has him better than that. It looks like his range factor is a product of opportunities that can be adjusted for.
   104. cardsfanboy Posted: February 26, 2014 at 05:11 PM (#4663090)
(Note: I'm not arguing that Vaughan was as good defensively as Bartell... pointing out that it doesn't take much to move the needle on range factor)
   105. Sunday silence Posted: February 26, 2014 at 05:28 PM (#4663101)

So unless I'm missing something, assuming a standard distribution of BIP, one good fielder means the others will, on average, have a below-average RF/9 regardless of their defensive ability. From an RF/9 perspective, it would be silly to have an IF of Brooks, Ozzie, Maz and Hernandez -- either they'll all look like average fielders or your OF will be "terrible."


my response to this was that in situation with all bad fielders, or all really good fielders, yes this would probably be an effect and probably a large effect. But the reality of MLB is that most teams have a few good ones, a few bad ones and a few in between. And because of that, for the majority of teams this effect is probably going to wash out.

I havent attempted to put a numerical value on all this, only that for Bill James to make this sort of argument is a vast over simplification, as long as teams have varied talent at def. positions.


Vaughan moves from being .93 back in range factor to .78 by just adjusting out the double counting of double plays.


That sort of adjustment certainly seems very reasonable. I dont quite understand the methodology and wish you could explain it more. How did you get adjust range factor? WHere did that come from. Not the outs per game???

*****

Can we take a time out and visit double plays for a moment?

First of all I dont think they are being double counted. if it says "DP Bartell 100" that doesnt mean he got a PO and an ass't on 100 DPs. I think it means he participated in 100 DPs. He could have been back end front end, or in the middle.

So when you want to avoid double counting them, how do you propose to do that? We dont have the info for which plays the guy was in the middle and record two chances, or he was at the end and only recorded one chance. Am I correct in that assumption?

Secondly, I think starting a DP can result in underestimating the importance of a fielder. If say Mazeroski or Hernandez starts a DP, it is entirely possible that their quickness in starting the DP was the primary reason that the DP was made in the first place. If that is so, then in fact Maz or Keith was actually the primary reason they got 2 outs and not one, and in fact he is being under represented in his range factor. Because while Keith gets one assist on the play, in fact he might be the main reason they got two outs. I say "main" because obviously someone had to catch the ball etc, but we assume most of that is routine.

It's also possible that pivot guy was the reason that DP was made, and he did get credit for two chances. There's no real record on this, and I guess he have to use contemporary accounts to try to get an idea of this.

But the point is that DP participation can in fact under represent a good fielders effectiveness. This would be hard to quantify admittedly.

Maz. is a good example, because it seems quite likely that he was one of the main reasons the Pirates set that DP record back in the 60s when they did. He was also near the DP lead or on top for many seasons in the 60s. THere is no way to prove that he was the main factor, but we do think he was a super fielder and it is reasonable to suppose that he was more responsible for getting two than anyone else. Unless we are to believe that Gene Alley or Bob Bailey was the real reason. It's possible but not likely.

So basically DP participation is important, and it can underrepresent defensive ability. But this is not shown by the records we have.
   106. Sunday silence Posted: February 26, 2014 at 05:31 PM (#4663106)

This thread is the first place I've ever seen suggesting that Vaughan was a some sort of terrible fielder.



I dont think I said he was terrible. He's decent for most of his career and he really stepped it up for '38 to '40.

It is also clear from the Raw range, that Bartell is making a lot more plays than he is. It is also quite possible enuf to make up for the deficiency in the offensive department.

The only question is how do we quantify that? That seems a fair question. I dont see what your pt is at all.
   107. Sunday silence Posted: February 26, 2014 at 05:40 PM (#4663114)

Vaughan moves from being .93 back in range factor to .78 by just adjusting out the double counting of double plays.


OK that's how you did that, you adjusted out the DPs.

That's not going to work, unless I misunderstand you.

Participating in 100 DPs does not mean, Bartell recorded a PO and an ASST in 100 DPs, it only means he participated in 100 DPs.

Besides, it can just as well be argued that starting a DP is sometimes more important than being the pivot. And if that's true then DP participation is under representing someones value.

If the pivot man was the key, and many times he is, then recording an asst/PO in one play would at least be a decent reward.

It's at least a wash, but there's no reason to substract DPs, that's just not correct at all.

Unless you know how many DPs he was in the middle. But even then it's still arguable.
   108. snapper (history's 42nd greatest monster) Posted: February 26, 2014 at 05:40 PM (#4663115)
I dont think I said he was terrible. He's decent for most of his career and he really stepped it up for '38 to '40.

It is also clear from the Raw range, that Bartell is making a lot more plays than he is. It is also quite possible enuf to make up for the deficiency in the offensive department.

The only question is how do we quantify that? That seems a fair question. I dont see what your pt is at all.


I think his point is that every other source (objective or subjective) has Vaughan average or better. Since raw RF is basically useless, why should we pay any attention to it?
   109. Sunday silence Posted: February 26, 2014 at 05:45 PM (#4663120)
Adjusting for KOs, as Cardsfan touches on in his post, means that the NYG get about what 22 more balls in play than league average over a season?

That works out to say 5 or 6 balls more for the SS? Which is something like .02 or .03 of a range factor.

So I dont think the answer is going to lie there.

There could very well be something going on there but the KOs dont seem to be it.
   110. SoCalDemon Posted: February 26, 2014 at 05:46 PM (#4663121)
It is also quite possible enuf to make up for the deficiency in the offensive department.


I think people have given plenty of reasons why Bartell is not 50 runs/year better than the metrics say, which is the difference we are talking about for 1933-35.

The only question is how do we quantify that?


Which a lot of different people have done a lot of different ways, and pretty much all of them agree that:

1. Range factors overstates the difference between the two.

2. Bartell is a better fielder than Vaughn

3. Vaughn is an average to above average fielder

I think you should ask a different question. Instead of focusing in on Bartell vs. Vaughn, how about you look at some more recent fielders (whom we have a lot more data for), look at their range factors, look at the advanced metrics, look at their defensive reputations, and see if their is a 1 to 1 correlation between all of that. There isn't. Not even close.
   111. Sunday silence Posted: February 26, 2014 at 05:48 PM (#4663123)
Since raw RF is basically useless, why should we pay any attention to it?


OK this is the second time this has been asserted without a reference.

Is there a reference that you have for this? or what is the reasoning.

I understood what was said about a team of bad fielders or a team of all good fielders. But on a team where fielding varied across the spectrum is this still a valid argument?




   112. Sunday silence Posted: February 26, 2014 at 05:51 PM (#4663127)

I think people have given plenty of reasons why Bartell is not 50 runs/year better than the metrics say, which is the difference we are talking about for 1933-35.


Wait a second! Where did we come to that conclusion?

The team strike out data?

Some sort of GB/FB split?

Some park factor?

Can you at least show your work here, this seems to be a pretty big point if you're just glossing it over.
.
   113. SoCalDemon Posted: February 26, 2014 at 05:51 PM (#4663128)
Re 109: Doing the math, .36 plays per game is 58 plays more, which is really close to the 59 SOs more for Vaughns team I cited earlier. And again, NOBODY is saying strikeouts explain the difference. They are saying that you need to take a whole basket of things into account, which, even for players from 1936, we either CAN do (SOs, caught stealings, double plays, etc), or can estimate (estimated palys in zone based on estimated groundballs from the fielding data we do have and pitcher handedness [and I am sure other stuff too]).
   114. cardsfanboy Posted: February 26, 2014 at 05:51 PM (#4663129)
That sort of adjustment certainly seems very reasonable. I dont quite understand the methodology and wish you could explain it more. How did you get adjust range factor? WHere did that come from. Not the outs per game???


Adjusted range factor was (po+asst-dp)/games.

Agree with you on that it's possible he didn't get an assist and a putout on each dp, but we don't have that data. Which is why they often regress(in this case you could argue half the dps should be removed) And about the pivot, a second baseman can obviously be a big deal on that, not really sure that a shortstop figures that heavily in the number of dps other than their own range.

And agree about there are different players capable of making a difference in the likelihood of a dp occurring, but honestly the single biggest factor in double plays is always going to be number of runners on base. My entire point is that it's not that hard to move the needle on an individual .10 .20 or .30 even in regards to range factor. There is a reason nobody uses it any more, it's been thoroughly trashed.

Most of the advance stats do all this legwork. If you read Bill James win shares book, he did a fantastic job of explain how he came up with his defensive shares (which in my opinion is pretty good stuff) The advance stats basically remove homeruns, walks and strikeouts, they estimate flyball vs groundball, you would also remove caught stealing from the equation if you have that information, you adjust double plays since it's almost always being counted in both putouts and assists.


To create your own defensive system from scratch you have to do a lot of legwork. You need to figure out handedness of the pitching staff. You need to figure out flyball/groundball rate. You need to adjust for park. You need to adjust for double plays. Strikeouts. Homeruns. (Defense efficiency does a lot of that) Adjust for scoring era. Bill James book spent probably over 100 pages explaining his methodology.
   115. Sunday silence Posted: February 26, 2014 at 06:02 PM (#4663136)

I think his point is that every other source (objective or subjective) has Vaughan average or better. Since raw RF is basically useless, why should we pay any attention to it?


let me try to answer this again.

If Vaughan is average and Bartell is above average, isnt there some sort of quantification that we have to make? to see if Bartells def. is enuf to equal Vaughans' offense?

Or are you saying that they are both average? or they are both equal? is that what you are saying? I admit I have trouble following what exactly your pt is.

If you assume Vaughan is as good as Bartell defensively, then how do we explain the difference in range factor.

It cant be that the NYG are letting like 200 balls a year dribble through their infield so that Bartell is getting to 100 more balls a year than Vaughan.

That's not possible if the NYG are a competitive team which we know they are.

It could certainly be possible if the NYG were a little leaguers that couldnt field. But that's not reality.

It might also be possible if most of their middle infielders were sub average. This also doesnt seem realistic because we've seen Bartell put up very good numbers in PIT years before this. THere's no reason to think that.
   116. Sunday silence Posted: February 26, 2014 at 06:09 PM (#4663137)
My entire point is that it's not that hard to move the needle on an individual .10 .20 or .30 even in regards to range factor. There is a reason nobody uses it any more, it's been thoroughly trashed.


Yes but how do you go from moving the needle 0.2 to "thoroughly trashed" as if it doesnt matter at all?

That's a huge leap isnt it?

For one thing, if we could quantify how a 0.2 difference might have happend (say from GB pitching staff) then we could have even better idea of what FieldRange means. Yes?


Can you again explain what is the basis for this "thoroughly trashed" belief?

is it entirely in the James example of 8 bad fielders? That's an over simplification as I've said.
   117. cardsfanboy Posted: February 26, 2014 at 06:09 PM (#4663138)
Other variables that might have factor into the differences.
The Giants had 513 innings thrown by left handed pitchers. The Pirates had 109 innings.
The Giants on averaged faced an extra right handed bat every other game.

Again there is a lot of involved. By rfield Bartell was worth 30 more runs than Vaughan that year. (1936)
   118. Mefisto Posted: February 26, 2014 at 06:10 PM (#4663139)
So unless I'm missing something, assuming a standard distribution of BIP, one good fielder means the others will, on average, have a below-average RF/9 regardless of their defensive ability. From an RF/9 perspective, it would be silly to have an IF of Brooks, Ozzie, Maz and Hernandez -- either they'll all look like average fielders or your OF will be "terrible."


But from a team perspective, we don't care how the individual fielders look. That team would have a great DER, so any manager who could put those players together should do so.
   119. Sunday silence Posted: February 26, 2014 at 06:15 PM (#4663144)
And about the pivot, a second baseman can obviously be a big deal on that, not really sure that a shortstop figures that heavily in the number of dps other than their own range.


I havent really thought much about this, but I was thinking that shouldnt the SS be the pivot in maybe 40% of the DPs? Thinking there are a lot of left hand batters in MLB but then again not quite that many in the 1930s. Again havent really thought about it much.

But arent you making my pt? If Bartell is not the pivot in most of those plays, then most of those plays he's not getting an asst/PO. And so you cant penalize him 100 chances. If say he's the pivot in say 30 of those, then at most you can subtract 30.

But it works both ways as I've said. THe pivot guy might be key to one DP, the guy starting the DP could be key in another.

I have to say it appears to be a wash and I would just leave those untouched. I am not sure why you are so certain there needs to be a correction here.
   120. cardsfanboy Posted: February 26, 2014 at 06:15 PM (#4663145)
Can you again explain what is the basis for this "thoroughly trashed" belief?


Most of the articles are so old, I haven't been able to find them. Here is a methodology for adjusting range factor.

Basically range factor has been thoroughly and completely replaced with the different modified range factors that it's dropped out of use. Uzr and other systems use the very basic concept of range factor but removes the simplifcation out of it to make it more analytical.

From this article.

On another point of great interest, James does a big mea culpa on his prior advocacy of range factors. While that groundbreaking work set the stage for many of the more recent developments in fielding statistics, by focusing attention away from errors and onto a fielder’s ability to make plays, James now concedes that the statistical illusions that plague range factors make them too suspect to use as a benchmark for defense. He specifically argues that Total Baseball’s rating of Nap Lajoie as one of the top handful of players all time, based on his defensive statistics, is deeply misguided.


   121. SoSHially Unacceptable Posted: February 26, 2014 at 06:18 PM (#4663148)
But from a team perspective, we don't care how the individual fielders look. That team would have a great DER, so any manager who could put those players together should do so.


Walt's not disagreeing with that. He's only noting the inherent weakness in RF. Judging a team's players that way will ultimately result in some guys looking good/others bad or everybody looking average.

It's like trying to judge league quality, pre-interleague play, by looking at W-L records.

   122. cardsfanboy Posted: February 26, 2014 at 06:20 PM (#4663150)
I havent really thought much about this, but I was thinking that shouldnt the SS be the pivot in maybe 40% of the DPs?


But it's not that difficult of a play from the shortstop prospective. They are moving towards first base, they don't have to make a rotation. Basically every shortstop, even Jeter and Eckstein, if they get the ball in time, will make the dp.

So unless I'm missing something, assuming a standard distribution of BIP, one good fielder means the others will, on average, have a below-average RF/9 regardless of their defensive ability. From an RF/9 perspective, it would be silly to have an IF of Brooks, Ozzie, Maz and Hernandez -- either they'll all look like average fielders or your OF will be "terrible."


Yes. That is true (it's been examined with regards to Andruw Jones etc.) That has been the problem with range factor on the individual level. If you have one good guy who is stealing an extra play every other game, the rest of the individuals are now missing out on a chance roughly every two weeks(on average obviously there is a different distribution on the play.)

That is partially why to create a modified range factor, you have to start with the team level first and include balls missed in the discussion. Individual range factor without modification is pretty useless.
   123. Sunday silence Posted: February 26, 2014 at 06:24 PM (#4663152)
but honestly the single biggest factor in double plays is always going to be number of runners on base.


OK I am pretty sure you're missing my pt here; and I dont blame you; Im prolly not being clear.

I am not talking about the rate of DP during the season. I am saying for one DP:

If Hernandez does not make the quick throw to start the DP, that DP might never happen. He might have been the single most important factor among those defenders. THe 2baseman had an easy pivot and the pitcher made an easy catch at first.

SO Keith gets one asst, but he was actually the most important reason they got 2 outs.
2bman, made a routine pivot, he got two chances.
pitcher made a routine catch he got one chance.

It seems to me much of the time, the guy starting the play might be most important. In other times, the pivot guy might be key. I think it's a wash, and I dont you can penalize players for participating in more DPs.



This is the example I gave of Maz. There's a reason they set a DP seasonal record, that reason could very well have been more about Maz then Gene Alley or whoever.
   124. Sunday silence Posted: February 26, 2014 at 06:32 PM (#4663158)
But it's not that difficult of a play from the shortstop prospective. They are moving towards first base, they don't have to make a rotation. Basically every shortstop, even Jeter and Eckstein, if they get the ball in time, will make the dp.


Yeah but I'm thinking their is a physical ability to avoid the slide, to throw in the air, or just the sheer physical courage of standing there. Right? Maybe Jeter is not so athletic but does he credit for courage to stand there and make that play?

Maybe its' replaceable ability. I dont know. Once in while you do see someone short arm it over there.

OH thanks for that reference cardfan. I will have to check it out. I have to take a break now.

Thanks for all the debate, this has helped me think about this a lot more. SOme of these threads come and go real fast and I never get to the bottom of things.
   125. alilisd Posted: February 26, 2014 at 06:33 PM (#4663160)
Yeah I'm going Everest, Kilimanjaro, Olympus, Fuji


Olympus, Fuji, Nikon, and Canon for me.
   126. GuyM Posted: February 26, 2014 at 06:50 PM (#4663166)
I don't have a dog in this fight, but looking at the 1936 data it's amazing how many more putouts SS had at that time:
2013 SS: 483 assists, 229 putouts
1936 SS: 494 assists, 340 putouts

A 1936 SS had about the same number of assists as today's SS (in slightly fewer games), but made 110 more putouts. That's a huge difference. While there must have been somewhat more caught linedrives and popups (because there were so many more BIP), this must mean there were a ton more forceouts at 2B. That suggests to me that maybe middle infielders in 1936 were going for the force on a lot of plays that today would result in an out at 1B. That could be because their arms were weaker, and maybe the slower speed of the runners played a role too (although the batter and lead runner would both have been slower).

In general, it's the assists that are a much better measure of fielding ability. I would focus on that rather than range factor. Of course, you still need to adjust for BIP, handedness, etc. But it's hard to imagine any adjustments that will result in Bartell's 559 assists (vs. 477 for Vaughn) not looking impressive.
   127. Mefisto Posted: February 26, 2014 at 06:51 PM (#4663167)
Walt's not disagreeing with that. He's only noting the inherent weakness in RF. Judging a team's players that way will ultimately result in some guys looking good/others bad or everybody looking average.


Ok, then we agree.
   128. PepTech Posted: February 26, 2014 at 07:37 PM (#4663201)
Arod was a fiery and restive horse, but Legolas rode him easily without saddle or rein.


There's just no way *not* to quote that when given an opportunity. You may now go back about your business.
   129. Walt Davis Posted: February 27, 2014 at 02:49 AM (#4663340)
Something got into the water in NY in 1936.

Bartell,Schmartell ... Burgess Whitehead, the 2B, had 552 assists himself and a RF/G of 6.50. That's insane.

Prior to that year, Whitehead never had a RF/G of 5 or better. He maintained his insane rates for a few more years.

Bartell had a similar peak/fluke with his assists jumping by over 100 in a year (and his age 28 season at that). Both of them see massive jumps in their Rfield as well.

To put that in perspective, Maz never quite got to Whitehead's total and the best RF/G of his career was 6.20.

Anyway, see two guys peak at the same time, with fairly extreme numbers and not a massive amount outside those peak years -- that's most likely something contextual, not true talent.

By Rfield

34-35, Bartell was +9 (+5 in NY), Whitehead was +3 in about 100 starts at 2B and 30 at 3B in StL
36-37, Bartell was +48 and Whitehead was +33
38, Bartell was +14 and Whitehead didn't play
39, Bartell was on the Cubs and -8 (assists down by about .5 per start), Whitehead still in NY was +9 in just 89 starts
40, Bartell was in Det and -10, Whitehead in NY and +2 in a full season

Whiteheas's (main) replacements on the 38 Giants were Chiozza, Cissell and Kampouris. The first two spent some time at other positions too, all told they were +11. In Cincy, Kampouris had been average at best (+6 in a half season in 38), Cissell was in his last season and had been average or worse the years before (+4 in 32 starts at 2B, 4 at 3B), Chiozza was in his next-to-last season and played all over before this and was average in 38. The three combined for 491 assists at 2B, 23 fewer than Whitehead the year before.

In 39, the Giants used Billy Jurges at SS. He is credited as an oustanding SS early in his career but had been just +3 in Chicago for 37-38. At age 30 in NY, he put up a +11, and he maintained that level for several years with the Giants.

Two solid but unspectacular IF head to NY in their mid-20s and put up 3 outstanding years of fielding each then do nothing special. When Bartell was traded and when Whitehead was hurt (I assume), their replacements (emergency ones in Whitehead's case) also put up very good numbers. It's like the Polo Grounds were the Coors Field of easy ground balls those few years.
   130. Sunday silence Posted: February 27, 2014 at 03:57 AM (#4663345)
I have a question: the Rfield that you all are referring to that is based on Total Zone ratings, correct?

If yes, then does anyone have an idea of how these are computed for years prior to 1956? 1956 was mentioned as the cut off year for certain play by play data, I think from project retrosheet.

Here are two articles by Sean Smith explaining his Total Zone rating for those who are interested:


http://www.hardballtimes.com/measuring-defense-for-players-back-to-1956/

http://www.hardballtimes.com/measuring-defense-for-players-back-to-1956-part-2/
   131. AROM Posted: February 27, 2014 at 10:33 AM (#4663413)
The system used before 1950 or so is called JAARF. Just another adjusted range factor

It was a painful process that started with the need to estimate innings in the field. Say you've got 5 players who played 140, 25, 15, 10, and 5 games at shortstop in a 154 game season. Obviously they didn't all play a complete game, but one thing we know is it's likely that the starter played more complete games, or at least more innings per game, than the backups.

One thing you can do is use plate appearances to estimate innings, but I reject that because it works really poorly considering defensive replacements and strategic pinch hitting. I prefer to use defensive chances to estimate innings. But that runs into problems, say the guy who played 10 games had a lot of chances, and the system says he had 13 equivalent games. That's not possible, so I cap him at 10, and then do another iteration for the rest of the players. I never published the estimate innings because I don't trust them at all. It's just a stand in and I hope some day retrosheet can give us more accurate actual innings played (they've done this for much of the 1940's, though some games are missing).

Then there's an adjustment for LHP and GB/FB based on the team assist to putout ratio.

Looking at 1936, NYG was second in DER for the league at .693. Pittsburgh was third at .691. Both very good defensive teams. Giants had 2068 assists, most in the league. Pittsburgh had 1672, 7th place. That strongly suggests that New York was getting many more groundballs than Pittsburgh.
   132. Sunday silence Posted: February 27, 2014 at 10:48 AM (#4663427)

The system used before 1950 or so is called JAARF. Just another adjusted range factor


DOes it use the raw Range Factor in some way? I am not sure what Sean Smith was describing in those two articles but it sounded like he was talking about estimating how many hits were falling into areas assigned to various fielders.

Still didnt get an answer to the first question: is Rfield based on Total Zone?

is any of this based on any raw range factor?
   133. AROM Posted: February 27, 2014 at 11:02 AM (#4663451)
No, not based on raw range factor. I don't think I did much with putouts, focused on assists. That's an infielder's plays made. Opportunities is plays made + errors + an estimate of hits responsible for.

The last part is the tough one. It's not like in total zone where I know a ground ball fielded by the center fielder must have been the responsibility of the shortstop or second baseman. Could be anywhere. I make estimates based on GB/FB ratio (an estimate itself) and pitcher handedness (which mostly affects 1B and 3B).
   134. GuyM Posted: February 27, 2014 at 12:03 PM (#4663518)
SS: Just so's you know, "AROM" = "Sean Smith."

AROM: do you have any thoughts on why SS recorded so many more PO back in the 1930s?
   135. AROM Posted: February 27, 2014 at 12:40 PM (#4663544)
No idea. Nobody ran back then so it's not caught stealing.

If it's more force plays, I don't think it's elective. I mean, teams will go for the force out today if the option is there, because taking a runner off second base is more valuable than taking one off first. In other words there aren't enough force plays that SS could take now but elect not to, but were taken 80 years ago.

It could be opportunities for force plays - higher batting averages but not as many power hitters means many more runners on first.

More infield popups?

Lots of runners trying to stretch an extra base? That would show up in OF assists, and the difference isn't that huge, 40 average in 1936, 25 average last year.

There was a lot of bunting in the 30's, maybe part of it is unsuccessful sacrifice attempts.

Retrosheet now has half of the seasons in the 1930's, so the answer is there for someone who wants to look into it.
   136. SoSHially Unacceptable Posted: February 27, 2014 at 12:41 PM (#4663547)
AROM: do you have any thoughts on why SS recorded so many more PO back in the 1930s?


And to follow up: SS = Shortstops, not Sunday Silence or Sean Smith. (-:

   137. jingoist Posted: February 27, 2014 at 01:18 PM (#4663582)
GuyM @ #126:
Could the number of switch hitters, late innings platooning or more LH hitters in todays game be a partial cause for the numerical difference in SS plays made?
   138. Sunday silence Posted: February 27, 2014 at 01:20 PM (#4663585)



SS: Just so's you know, "AROM" = "Sean Smith."


That helps. I was going to ask, but just felt sheepish for not knowing or whatever.

Hopefully Sean can stay and answer a few more questions:

Q 1. Would it be possible to use Team Defensive efficiency along with player As range in order to calculate some sort of defensive rating? Would it be roughly similar to what you are doing with Total Zone rating?

Q 2. Do you attempt to reconcile your defensive rating, along with whatever the pitcher gave up (HRs, walks, etc) with the total runs that team gave up? I.e. is there some sort of double check to make sure we are not over or under counting the defensive rating?

This part is probably what concerns me most about all these methods.

Q 3 What is the state of PBP data in the pre war era? For instance do we have complete PBP back to 1920?

Q 4 What is up with caught stealing % in the 1930s? The NL didnt keep this record but when I go to individual catchers in Baseball Reference I do see it. How is that?
   139. Sunday silence Posted: February 27, 2014 at 01:38 PM (#4663601)

AROM: do you have any thoughts on why SS recorded so many more PO back in the 1930s?


I am not sure the issue is with SS but with certain teams and their high number of assists. Hughie Critz '33-34 has like 6.4 range factor which is out of line with his career or anybody elses. It seems like the NYG have unusually high number of assists in those days.

maybe they produced lots of 6-3-6-5-6-3 run downs or something?
   140. bjhanke Posted: February 27, 2014 at 01:39 PM (#4663603)
For what this is worth, Bill James, back in the New Historical, concluded, doubtless by computer counting, that SS are the pivot man on 83% as many DPs as the 2B is. Don't know if this was true in Vaughan and Bartell's time. - Brock Hanke
   141. AROM Posted: February 27, 2014 at 02:36 PM (#4663672)
Q1: Probably. The TZ ratings for all the players on the team generally correspond to the team DER. Sometimes there's a discrepancy, and park factors play into it (though I didn't have any park factors for the pre 1950s teams).

Q2: Can't be sure if the defensive rating is over or under. But generally closer to zero than to any positive or negative extreme at the individual level. At team level, it's probably about right since it's on the same scale as DER. At the team level pitcher WAR is based on runs allowed - defensive support, so pitchers + fielders will by definition add up to park adjusted runs allowed.

Q3: See retrosheet. I did look at the games missing from 1934, and see that they have PBP data for about 60% of games in that year. I haven't had the time to do much with the older data retrosheet keeps adding.

Q4: I think the source of those numbers is Pete Palmer, and I have no idea how he got it. You'd have to ask another Sean about that.

maybe they produced lots of 6-3-6-5-6-3 run downs or something?


Those kind of plays are too rare to affect the numbers that much. There's an easy explanation here: Their pitchers got a ton of ground balls. I wonder if the Neyer/James guide to pitchers could shed light on this. Everyone knows Hubbell threw a screwball, other than that I'm not sure what they threw.
   142. GuyM Posted: February 27, 2014 at 02:42 PM (#4663674)
My guess is that elective plays is part of the explanation, but probably a number of things have changed. It's true that you want to get the lead runner if possible, but not with 2 outs. So 1/3 of the time, getting the lead runner or going to 1B is discretionary. In today's game, I often see SS and 3B throw across the diamond with two outs on plays where I would go to 2B (in a softball game). It may be that in 1936, with weaker infielder arms (and sometimes very bad fielders at 1B), the default was to take the force out more often with two outs.

But I doubt that can account for the whole change. Here's some more data for anyone interested in pursuing this. Part of the change is increased Ks of course: catchers had 1,234 PO per team last year, compared to just 656 in 1936. Another change is more flyballs today: Outfielders made 33% of the non-catcher PO last year, compared to 30.6% in 1936.

If we look just at PO made by infielders and pitchers, this is what we see:

POS 1936% 2013% Diff
1B 60.4% / 65.5% / +5.1%
2B 16.3% / 14.1% / -2.2
SS 14.1% / 11.1 / -3.0
3B 6.8% / 5.1% / -1.7%
P 2.4% / 4.2% / +1.8%

So 1B and P are making more putouts, while 2B, SS, and 3B are all making fewer. It's just a gain/loss of 2% or 3% of PO at each position (except 1B, a gain of 5%), put proportionally these are pretty big changes: 13% drop at 2B, 22% drop at SS, 25% decline at 3B, and a 76% increase at pitcher. Pitchers are clearly handling a lot more plays today. And it does seem that there are many fewer force outs being made at 2B than in the 1930s.
   143. AROM Posted: February 27, 2014 at 03:07 PM (#4663699)
Maybe more popups too.

If it's more forceouts, it's probably a matter of opportunity.

On-first-base-PCT:
2013 .241
1936 .275

That's OBP with extra base hits removed, since you can't get a forceout if the previous batter hit a double. Compared to 1936, we have much lower OBPs, but also much great extrabase hit rates, particularly the home run.
   144. Sunday silence Posted: February 27, 2014 at 03:12 PM (#4663702)

Those kind of plays are too rare to affect the numbers that much.


I was being facetious.

I did read something, was it your explanation, where it says the only take the initial fielder on the assists, so they would have taken these plays out of the equation in any event. Or at least that's the way I read it, not sure.
   145. AROM Posted: February 27, 2014 at 03:15 PM (#4663706)
If it's popups I'd expect that to show up more at 3B than the other positions. Sometimes a 3B fields a ball and steps on the bag for an out, but other than bunts they rarely take a forceout throw from another infielder.

3B averaged 163 putouts in 1936, and just over 100 last year. Oakland (with their large foul territory) with 151 led MLB by a wide margin. Of those, 132 were catches, 15 were tags, and only 4 were forceouts. Average MLB team had 10 forceouts by 3B last year.

Bill James once wrote about catchers back then having a lot more non-K putouts, and I think popups were the answer then.

So that's my theory - combination of more runners on first to force out, and more popups (maybe old parks had more foul territory?).
   146. AROM Posted: February 27, 2014 at 03:17 PM (#4663708)
I did read something, was it your explanation, where it says the only take the initial fielder on the assists, so they would have taken these plays out of the equation in any event. Or at least that's the way I read it, not sure.


That's not it. You can have 5 assists on a single play that results in one out.
   147. bjhanke Posted: February 27, 2014 at 03:58 PM (#4663733)
AROM (#141) - I had an old memory of something from Bill James' New Historical, so I dug it out. In his section on Gus Mancuso, Bill says that the ground ball tendency of the NYG staff made a huge leap in 1933. In 1932, their catcher was the huge Shanty Hogan, who could not give a low target, and so made low strikes look like balls because he had to stab down to get them. In 1933, Mancuso took over the job, easily gave low targets, and the # of assists went through the roof. Bill's essay specifically mentions that this is because all the Giant starters, including Hubbell, throw low pitches as their main feature. - Brock Hanke (who, like AROM mentioned in another thread, is more or less perpetually in the process of reading the New Historical).
   148. Sunday silence Posted: February 27, 2014 at 04:01 PM (#4663735)
Wait a second, I'm confused!

WHere are you guys getting this Rfield data for the 1930s. I look at Baseball reference.com and under Fielding-Advanced it has Rtotal only going back to 1954. Prior to that date I do not see any advanced fielding metrics like Runs Above average or anything.

Walt: where did you get that Rfield data you posted in no. 129??
   149. Sunday silence Posted: February 27, 2014 at 04:03 PM (#4663736)

That's not it. You can have 5 assists on a single play that results in one out.


Arom: I am not sure I made myself clear. I was referring to the methodology used in one of the advanced fielding metrics and not official MLB assist numbers. Are we on the same page?

I guess I should dig out this reference again...
   150. AROM Posted: February 27, 2014 at 05:08 PM (#4663777)
Arom: I am not sure I made myself clear. I was referring to the methodology used in one of the advanced fielding metrics and not official MLB assist numbers. Are we on the same page?


Ok, yeah. TZ would treat this as groundball fielded by shortstop, results in out. One play made, one opportunity.

WHere are you guys getting this Rfield data for the 1930s. I look at Baseball reference.com and under Fielding-Advanced it has Rtotal only going back to 1954. Prior to that date I do not see any advanced fielding metrics like Runs Above average or anything.


Look in the player value section, it's there but not in the fielding section.

Brock - awesome find. I see Hubbell's HR total dropped from 20 to 6 those 2 years.

   151. Sunday silence Posted: February 27, 2014 at 06:17 PM (#4663816)
ARom: where is WAR for pitching for the 1930s pitchers buried in baseballreference pages??
   152. AROM Posted: February 27, 2014 at 09:03 PM (#4663877)
Buried?

It's in the player value tab, just like for every other pitcher.
   153. Mefisto Posted: February 27, 2014 at 09:48 PM (#4663891)
Would it be possible to use Team Defensive efficiency along with player As range in order to calculate some sort of defensive rating?


Heh. I asked Bill James this question at the SABR convention in 1985. He said he'd tried to work it out, but was unable to. Still seems like there should be something there.
   154. Eric J can SABER all he wants to Posted: February 27, 2014 at 10:44 PM (#4663901)
One thing you can do is use plate appearances to estimate innings, but I reject that because it works really poorly considering defensive replacements and strategic pinch hitting. I prefer to use defensive chances to estimate innings. But that runs into problems, say the guy who played 10 games had a lot of chances, and the system says he had 13 equivalent games. That's not possible, so I cap him at 10, and then do another iteration for the rest of the players. I never published the estimate innings because I don't trust them at all. It's just a stand in and I hope some day retrosheet can give us more accurate actual innings played (they've done this for much of the 1940's, though some games are missing).

For quite a few seasons back (further than the PBP years), Retrosheet has numbers for games, games started, and complete games at a given position. I played around with turning those numbers into innings estimates a while back, albeit over a pretty small number of years. It worked out to be approximately:

Innings played (as a fraction of team innings) = {2/9*(games played but not started) + 7/9*(games started but not completed) + 1*(complete games)}/(team games)

There was some slight variation in the coefficients across positions, but not more than a couple of percentage points, if memory serves.
   155. Sunday silence Posted: February 28, 2014 at 05:18 AM (#4663948)
the value that's bugging me at the moment is Rpos; that's the value (Runs above average) added due to positional scarcity.

I understand vaguely that LF is less scarce than CF; so that's OK. What I'm not getting is why there is anything other than 0 for the team value here. Since every team plays all 9 positions equally why should there be any adjustment needed for positional scarcity.

I am looking at the 1933 NYG right now, they get 48 RUns here for RPos. Why? It's not like the NYG were playing CF more than any other team.

So I guess it's that they had better players in key defensive positions.???
   156. Sunday silence Posted: February 28, 2014 at 06:04 AM (#4663952)
I'm also trying to get these WAA for batters and pitching to make sense but I dont get it.

Under the team stats>Team Player Value Batters they have Rbat; Rbaser; Rfield; and Rpos; I add all of those to get Wins above average for positional guys (WAA)

ANd then under Team Player Value for the Pitchers I look up WAA for pitchers.

So if add them both together shouldnt I get something close to their wins, or more likely their Pythagorean wins?

For NL pennant winners in the 1930s, this is very close or dead on for six of those teams. For the other four it seems a bit off.

THe four that seem a bit off:

Year...Team...actual wins...pyth wins...WAA

1930 STL 92 94 11.5
1931 STL 101 97 11.6
1935 CHI 100 101 19.0
1939 CIN 97 95 13.9

for an average prediction error of 5.75 games. all four teams did play 154 games.

That seems too much of error for this type of system, maybe I am missing a park factor but I think that has already been taken into account.

For the other six teams, I think the method works very well though, but here....


Also why is Rbat so low? I think for almost every team but the Cubs it is in negative territory what is going on with that?
   157. AROM Posted: February 28, 2014 at 10:10 AM (#4663997)
I am looking at the 1933 NYG right now, they get 48 RUns here for RPos. Why? It's not like the NYG were playing CF more than any other team.


Pitchers.

Also why is Rbat so low? I think for almost every team but the Cubs it is in negative territory what is going on with that?


Again, pitchers.

The league average that everyone is compared to is batting with pitchers excluded. That is important after 1973 to make sure Cal Ripken and Mike Schmidt are on equal footing. But then I end up with pitchers being some big negative number for rbat. It's balanced out by crediting pitchers in the rpos column.
   158. AROM Posted: February 28, 2014 at 10:14 AM (#4663999)
for an average prediction error of 5.75 games. all four teams did play 154 games.


I used to have the batting portion reconcile to actual runs scored. That is no longer used on the site, at some point Sean Forman switched to Pete Palmer's batting runs. So I'm guessing those teams that are off scored more runs than would be expected from their offensive inputs.
   159. Sunday silence Posted: March 01, 2014 at 06:49 AM (#4664412)
OK I am going to make a few more posts,not to get the last word in, but to collect the data I have found and put it here, so that when we revisit this discussion again, we can look it up. I'm not trying to get the last word in.

QUESTION: Are the Rfield factors based on the Total zone method or not? I think I've asked this a couple times and just not sure where this has been answered. This is doubly confusing because in the last post AROM says that they are using Pete Palmer's batting runs, which is fine, I suppose because the problem I am having is with the defensive numbers. So just to be clear. It is Sean Smith's Total Zone method that is used for the Rfield. Yes?

It looks very much like there is a systemic error in the relationship between the defensive runs that Rfield is assigning to the various fielders. It is not just random but rather at the high end of the curve it is pushing defensive numbers down and at the low end, it is pushing them up. So great fielders are being diminished and bad fielders are being helped up. More or less the bell curve is being pushed in at both ends.

I am doing this by hand, so I'll post what I come up with in a day or two.
   160. Eric J can SABER all he wants to Posted: March 01, 2014 at 08:37 AM (#4664414)
Are the Rfield factors based on the Total zone method or not? I think I've asked this a couple times and just not sure where this has been answered. This is doubly confusing because in the last post AROM says that they are using Pete Palmer's batting runs, which is fine, I suppose because the problem I am having is with the defensive numbers. So just to be clear. It is Sean Smith's Total Zone method that is used for the Rfield. Yes?

TotalZone is used for the seasons in which we have play-by-play data but not video-based numbers, so roughly from the mid-'50s to the mid-2000's. Before the PBP data starts, the numbers are an adjusted range factor and will be considerably less reliable, which is probably why they're regressed to the mean more heavily.
   161. CFiJ Posted: March 01, 2014 at 09:42 AM (#4664415)
See, I think every Mt. Rushmore needs its Teddy Roosevelt. The decent enough guy who is by no means in the same league as the other three.
   162. AROM Posted: March 01, 2014 at 10:08 AM (#4664421)
Eric's answer is correct. Total zone covers something like 1951 to 2002 in bbref. Exact dates might be wrong, going by memory. From 2003 on or so, rifled is John Dewan's defensive runs.
   163. cercopithecus aethiops Posted: March 01, 2014 at 10:31 AM (#4664424)
See, I think every Mt. Rushmore needs its Teddy Roosevelt. The decent enough guy who is by no means in the same league as the other three.


Teddy is on Mt. Rushmore basically because if it wasn't for Teddy, there wouldn't be a Mt. Rushmore. So that analogy isn't always going to work for the "Mt. Rushmore of ..." but I suppose the Mt. Rushmore of shortstops should have whoever it was who figured out that putting an infielder between 2B and 3B was a good idea.
   164. Sunday silence Posted: March 01, 2014 at 06:05 PM (#4664637)
Thanks for all those answers, really appreciate that.

A related issue: has anyone done any comparison of Defensive efficiency vs "luck" (+ or - in actual wins vs pythagorean runs)?

Looking at the 1930s and then briefly at the millenium, I think there might be some relationship. Not 1:1 but there seem to be some interesting numbers there.
   165. The District Attorney Posted: March 01, 2014 at 06:20 PM (#4664642)
t I suppose the Mt. Rushmore of shortstops should have whoever it was who figured out that putting an infielder between 2B and 3B was a good idea.
George Wright is usually credited with "inventing the position", being "the first modern shortstop", etc. It's not entirely clear to me what specifically that stuff means, but in between whatever pioneer status he has and being a really great player, he'd surely be a very logical choice.
   166. Sunday silence Posted: March 02, 2014 at 01:36 PM (#4664867)
OK here is the data I calculated to measure how much this regression to the mean (if that's what going on) is suppressing both ends of the bell curve of performances. Basic conclusion is that very good teams and very bad teams are doing better/worse than the metrics are suggesting. About 5-7 games per team/season. That could amount to say 10-15 def. runs for an outstanding middle infielder, if the error is coming down on Rfield, which I think is quite possible.


KEY for each column: Year/Team/Pythag Wins/Total Wins Above average/Difference between Pythag and WAA valu

NOTE: under pythag wins if team played less than 154 games this is noted as "less X games", I then added one or two
wins to the total wins to account for this e.g instead if 2 less games, I might add one to go from 89 to 90 wins. No attempt was made to turn these into fractions. Total WAA is found by adding: Rfield, Rbaser, Rbat, and pitching WAA for each team. Under the total: positive numbers means the team did better than the value suggests; and - means their pythag record is worse than the WAA value suggests.

1st place teams year by year, by pythag wins

39 CIN 95 13.9 4.1
38 CHI 88 (less 2) 11.5 0.5
37 NYG 89 (less 2) 12.4 0.6
36 CHI 93 14.7 1.3
35 CHI 101 19.0 5
34 NYG 95 less 1 13.7 5.3
33 NYG 90 less 2 11.6 2.4
32 CHI 86 7.6 1.4
31 STL 97 11.6 8.4
30 STL 94 11.5 5.5

avg error 4.8 games per team/season
I.e. this is how much the team out performed (by pythag wins) the prediction via the metric

2nd place teams

30 BRO 89 14.6 -2.6
31 NYG 93 (less 2) 13.7 3.3
32 NYG 82 -1.6 6.6
33 CHI 90 12.9 0.1
34 STL 90 less 1 12 2
35 STL 96 11.5 7.5
36 NYG 89 11.8 0.2
37 CHI 89 12.4 -0.4
38 CIN 84 (less 4) 8.6 0.4
39 STL 91 (1 less) 16.7 -1.7

avg error: 1.5 games...

last place teams
30 CIN 59 -8.2 -10
31 BOS 60 -8.4 -8.6
32 CIN 62 -3.3 -11.7
33 CIN 58 (2 less) -11.7 -6.3
34 CIN 55 (3 less) -15 -6
35 BOS 50 (1 less) -20.6 -6.4
36 PHI 64 -6.6 -6.4
37 BRO 61 less 1 -11.4 -5.6
38 phi 47 (less 4) 23.7 -5.3
39 PHI 47 (less 3) 23.8 -5,2

avg error -7.1 per team/season; i.e. they performed worse than the metric suggests

next to last place

30 PHI 60 -21 +4
31 CIN 61 -10.8 -5
32 STL 74 -3.6 0.6
33 PHI 61 less 2 -12.1 -2.9
34 PHI 64 less 5 -6.6 -4.4
35 PHI 60 less 1 -13.7 -3.3
36 BRO/BOS 68 -3.9 -5.1 (average of both teams)
37 PHIL 64 less 1 -12.4 -0.6
38 BOS 69 (less 2) -1.5 -5.5
39 BOS 66 (less 3) -5.9 -1.1


avg -2.3
   167. Sunday silence Posted: March 02, 2014 at 01:48 PM (#4664878)
Again, pitchers.

The league average that everyone is compared to is batting with pitchers excluded. That is important after 1973 to make sure Cal Ripken and Mike Schmidt are on equal footing. But then I end up with pitchers being some big negative number for rbat. It's balanced out by crediting pitchers in the rpos column.


This is Sean's response to my question in 155 why is Rbat in negative territory for nearly every team.

With all due respect to Sean, this can't be entirely what is going on because the Cub numbers are too far out of whack. Here are the Cubs Rbat for the 1930s:

Year/Rbat
30 44
31 62
32 -48
33 4
34 -40
35 27
36 -21
37 44
38 -41
39 -54

Keep in mind the league is about -40 (going from memory and didnt check all) whereas the Cubs here have 4 seasons where they are like 70 to 100 runs above the league average. How can that possibly be coming from pitcher slot? They only bat 1/9 times, so if the team averaged say 80 runs above the league Rbat, then I guess every pitcher would have to be like 9 runs better than the average MLB hitter?? It would be off the charts...

SOmething else is going on here.

   168. cardsfanboy Posted: March 02, 2014 at 02:30 PM (#4664903)
Keep in mind the league is about -40 (going from memory and didnt check all) whereas the Cubs here have 4 seasons where they are like 70 to 100 runs above the league average. How can that possibly be coming from pitcher slot? They only bat 1/9 times, so if the team averaged say 80 runs above the league Rbat, then I guess every pitcher would have to be like 9 runs better than the average MLB hitter?? It would be off the charts...


Unless you look at the entire league, you can't really say that what Sean said is wrong. I'm not sure what you are trying to say here, but the sum of rbat for a season in the 30's for all the teams routinely finished around 800-900. Roughly on average -50 runs per team. Positive rBat is coming from the other slots, not the pitching slot.
   169. Sunday silence Posted: March 02, 2014 at 04:52 PM (#4664959)
I am not sure what you dont understand.

I am referring to the team Rbat; it appears to be under the Team Batting Value or some such. YOu yourself state that the average team Rbat for the 1930s is -50. The cubs routinely seem to be 70-100 pts better in half those seasons. I do not understand how that is possible since the pitching staff can only provide 1/9 ABs, it seems there is not enuf AB to create so large a differential.

Sean had previously stated that the large negative was due to comparing team Rbat (which includes pitchers AB) to position player AB and that was the reason.

If you still dont get it, ask again I will try to explain, I thought I was clear.

Unless maybe I am looking at the wrong category?? Was I looking at the pitchers RBat only? But I dont think so. You seem to agree that the average Rbat for the entire team is -50 so we seem to agree....
   170. cardsfanboy Posted: March 02, 2014 at 05:05 PM (#4664968)
I don't get at all what you are trying to say.

He set up a standard across the league that compares all(non-pitcher) hitters to league average. This is the baseline he uses for rbat. If you remove every pitcher's at bat from the team totals for team rbat, you would end up with a league wide score of right around zero. (rounding issues, and other things could play into it not being exactly zero) then he figures out pitchers hitting relative to league average. Obviously they are going to be pretty poor and create a negative number.

On average they seem to create around -50 runs over the course of the season below a league average hitter. This doesn't seem unreasonable.

Look at a league average hitter, which would be someone with around 0 rbat, and compare to an elite level hitter and you get around 50-100 rbat. It's not unreasonable at all to think that a pitcher is 50 runs worse than a league average hitter over 154 games.
   171. Sunday silence Posted: March 02, 2014 at 06:13 PM (#4664988)
wait is the Rbat I am looking for just for the pitchers? WHere the hell is the team Rbat? I was told by Sean upthread that Rpos would then account for the negative Rbat score, which it appeared to do. I really think the Rbat is for the team; let me check again...
   172. Sunday silence Posted: March 02, 2014 at 06:15 PM (#4664990)
See e.g. this page for the 35 cubs;

http://www.baseball-reference.com/teams/CHC/1935.shtml

At the bottom is Team Player Value Batters; that includes all the batters Phil Cavaretta, Tex Carleton, CUyler, everyone not just pitchers. THere is a total for everyone (I presume thats why its at the bottom of the colum) scored as 35. This value fluctuates wildly for the Cubs from year to year I dont get it.

I dont know what I am missing, I am confused again...
   173. cardsfanboy Posted: March 02, 2014 at 06:23 PM (#4664993)
At the bottom is Team Player Value Batters; that includes all the batters Phil Cavaretta, Tex Carleton, CUyler, everyone not just pitchers. THere is a total for everyone (I presume thats why its at the bottom of the colum) scored as 35. This value fluctuates wildly for the Cubs from year to year I dont get it.

I dont know what I am missing, I am confused again...


Because some years they have very good hitters and some they don't? Why would it be stable?

Give me a specific season that you are questioning. The rbat is based upon league average hitter(with pitchers offense removed) it's not based upon team level, but the entire league so some teams, will have positive numbers if they are very good offensive teams.
   174. Sunday silence Posted: March 02, 2014 at 06:27 PM (#4664998)
EDIT: Ok maybe that is correct. I didnt look at all the teams, the Cubs varied wildly maybe it does make sense now
   175. Sunday silence Posted: March 02, 2014 at 06:34 PM (#4665004)
I guess its OK, it tends to be negative cause its set vs position players not all,so yeah it's OK.

I didnt mess up the calc in post 166 then so good.
   176. cardsfanboy Posted: March 02, 2014 at 06:36 PM (#4665007)
Basically what they do is
1. Remove all pitchers plate appearances from the data
2. Figure out what "average" is.
3. Create a formula that rates average on a runs scale where average equals 0.
4. Apply that formula to each individual player to create a run value relative to average of zero. At this point in time they also apply it to individual pitchers.

This is why on the overall team level the rbat of all players add up to negative numbers. But it doesn't necessarily have to. The 1935 Cubs led the league in scoring so you would expect that they would have a plus team rbat.

Click here for 1935 league data for player value to see how all the teams do on rbat.
   177. Sunday silence Posted: March 02, 2014 at 07:08 PM (#4665016)
Yes that sounds correct. I should note that the tables I put in post 166 also use Rpos in order to calculate the final WAA. I wasnt able to edit that.
Page 2 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
The Ghost's Tryin' to Reason with Hurricane Season
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Hot Topics

NewsblogMike Scioscia, Matt Williams voted top managers
(26 - 5:07am, Oct 22)
Last: Dr. Vaux

Newsblog2014 WORLD SERIES GAME 1 OMNICHATTER
(603 - 5:01am, Oct 22)
Last: Harveys Wallbangers

NewsblogSielski: A friend fights for ex-Phillie Dick Allen's Hall of Fame induction
(116 - 4:59am, Oct 22)
Last: Harveys Wallbangers

NewsblogOT: Monthly NBA Thread - October 2014
(317 - 4:33am, Oct 22)
Last: HMS Moses Taylor

Hall of MeritMost Meritorious Player: 2014 Discussion
(20 - 3:04am, Oct 22)
Last: bjhanke

NewsblogFan Returns Home Run Ball to Ishikawa; Receives World Series tickets
(57 - 2:07am, Oct 22)
Last: PreservedFish

NewsblogAs Focus Faded and Losses Piled Up, Royals Change Their Game
(7 - 1:16am, Oct 22)
Last: boteman

NewsblogRoyals’ James Shields passed kidney stone during ALCS but is ready for World Series | The Kansas City Star
(40 - 1:00am, Oct 22)
Last: Roger Freed Is Ready

NewsblogDealing or dueling – what’s a manager to do? | MGL on Baseball
(19 - 12:52am, Oct 22)
Last: Mike Emeigh

NewsblogOT: Politics, October 2014: Sunshine, Baseball, and Etch A Sketch: How Politicians Use Analogies
(2898 - 11:11pm, Oct 21)
Last: The Yankee Clapper

NewsblogDombrowski told that Iglesias 'will be fine' for 2015
(21 - 10:22pm, Oct 21)
Last: fra paolo

NewsblogOT: The Soccer Thread, September 2014
(852 - 8:40pm, Oct 21)
Last: Biff, highly-regarded young guy

NewsblogBaseball's hardest throwing bullpen - Beyond the Box Score
(10 - 8:02pm, Oct 21)
Last: ReggieThomasLives

NewsblogMorosi: Could Cain’s story make baseball king of sports world again?
(107 - 7:04pm, Oct 21)
Last: Spahn Insane

NewsblogBaseball Prospectus | Pebble Hunting: An Illustrated Guide to the People of Kauffman Stadium
(10 - 6:00pm, Oct 21)
Last: Perry

Page rendered in 0.5793 seconds
52 querie(s) executed