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

Friday, January 05, 2018

Most Meritorious Player: 2011 Discussion

The Cardinals defeated the Rangers in seven games. Vote for 15.

Player			WS	bWAR
Ben Zobrist		26.5	8.7
Matt Kemp		37.0	8.2
Jacoby Ellsbury		32.2	8.1
Jose Bautista		38.5	8.1
Dustin Pedroia		27.3	7.9
Ryan Braun		36.0	7.8
Miguel Cabrera		36.0	7.5
Evan Longoria		23.7	7.4
Alex Gordon		24.5	7.2
Ian Kinsler		25.5	7.1
Adrian Gonzalez		28.3	6.9
Joey Votto 		30.1	6.3
Pablo Sandoval		23.6	6.1
Troy Tulowitzki		24.2	6.1
Justin Upton 		29.8	6.1
Adrian Beltre		18.8	5.8
Robinson Cano		31.1	5.7
Andrew McCutchen	26.8	5.7
Curtis Granderson	27.8	5.7
Hunter Pence		25.2	5.6
Mike Napoli		23.0	5.4
Shane Victorino		24.2	5.4
Albert Pujols		25.9	5.3
Peter Bourjos		17.3	5.1
Alex Avila		28.0	5.1
Asdrubal Cabrera	23.0	5.0
Chris Young		21.1	5.0
Austin Jackson		15.4	4.9
Matt Wieters		19.4	4.9
Brandon Phillips	22.5	4.7
Jose Reyes		28.4	4.7
Yunel Escobar		19.0	4.7
Carlos Beltran		25.6	4.6
Howie Kendrick		19.7	4.6
Prince Fielder		32.7	4.6
Cameron Maybin		17.8	4.5
Erick Aybar		20.6	4.4
Melky Cabrera		19.5	4.4
Carlos Gonzalez		17.5	4.3
Elvis Andrus		17.8	4.3
Miguel Montero		26.2	4.1
Giancarlo Stanton	20.6	4.1
JJ Hardy		21.7	4.1
Brett Gardner		15.2	4.1
Carlos Lee		18.5	4.0
Brendan Ryan		13.4	4.0
David Ortiz		19.8	3.9
Kevin Youkilis		15.6	3.9
Matt Holliday		21.3	3.9
Alex Rodriguez		15.1	3.9
Lance Berkman		28.7	3.5
Carlos Santana 		25.5	3.8
Paul Konerko		24.8	2.9
Mike Morse		24.2	3.4

Pitcher 		WS	WAR
Cliff Lee		22.3	9.2
Roy Halladay		23.3	8.7
Justin Verlander	27.0	8.4
CC Sabathia		18.7	7.4
Clayton Kershaw		22.1	7.2
Jered Weaver		24.0	7.0
Cole Hamels		18.1	6.5
Ricky Romero		18.9	6.3
Josh Beckett		15.6	5.7
Doug Fister		17.7	5.3
James Shields		20.2	5.1
CJ Wilson		18.2	5.1
Ian Kennedy		19.8	4.8
Scott Baker		11.7	4.4
Gio Gonzalez		14.7	4.4
Dan Haren		18.2	4.4
Jon Lester		13.4	4.3
Justin Masterson	15.1	4.1
Matt Cain		16.0	4.0
Tim Lincecum		16.8	4.0
Mark Buehrle		16.0	3.9
Matt Harrison		13.4	3.9
Jeremy Hellickson	15.0	3.8
Felix Hernandez		15.6

Craig Kimbrel		16.8	2.4
Jon Axford		15.5	2.3
Jose Valverde 		14.8	2.0
David Robertson		11.1	4.0
Tyler Clippard		12.6	3.4
Mariano Rivera		13.2	3.3

 

DL from MN Posted: January 05, 2018 at 05:52 PM | 90 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.

   1. DL from MN Posted: January 08, 2018 at 11:06 AM (#5602633)
Looks like another good election this season
   2. Carl Goetz Posted: January 08, 2018 at 02:58 PM (#5602836)
Wow, this is definitely a season where it matters to which uber-stat you are partial.
   3. DL from MN Posted: January 08, 2018 at 06:46 PM (#5603044)
2017 Prelim

1) Cliff Lee - some of his value is bat
1) Roy Halladay - in a dead heat at the moment
3) Ben Zobrist - within rounding error of Lee and Halladay, terrific glove
4) Matt Kemp - big bat and just slightly below average defense in CF
5) Jacoby Ellsbury - good year for centerfielders
6) Justin Verlander - top AL pitcher
7) Dustin Pedroia - good glove
8) Jose Bautista - 2nd best bat
9) Ryan Braun
10) Evan Longoria - good glove
11) CC Sabathia
12) Ian Kinsler - good baserunning
13) Miguel Cabrera - best bat
14) Alex Gordon - well rounded player
15) Jered Weaver
   4. caiman Posted: January 11, 2018 at 09:41 AM (#5604507)
Here's the RPA top players for 2011, in terms of runs produced above the median:

1. Jose Bautista 49.61 runs
2. Ryan Braun 49.48 runs
3. Jacoby Ellsbury 46.46 runs
4. Matt Kemp 46.23 runs
5. Miguel Cabrera 42.53 runs
6. Alex Gordon 39.94 runs
7. Lance Berkman 37.23 runs
8. Andrew McCutchen 32.79 runs
9. Clayton Kershaw 32.32 runs
10. Joey Votto 32.28 runs
11. Shane Victorino 32.06 runs
12. Justin Verlander 31.31 runs
13. James Shields 31.12 runs
14. Justin Upton 31.01 runs
15. Prince Fielder 29.95 runs
16. Mike Napoli 29.95 runs
17. Curtis Granderson 28.40 runs
18. Dustin Pedroia 27.87 runs
19. Giancarlo Stanton 27.78 runs
20. Jose Reyes 27.57 runs
21. Carlos Beltran 27.27 runs
22. Roy Halladay 27.23 runs
23. Jered Weaver 27.04 runs
   5. Carl Goetz Posted: January 11, 2018 at 10:47 AM (#5604574)
RPA does not seem to like 2B or the Tampa Bay Rays. Or Cliff Lee for that matter. 9 of the top 10 (and 15 of the top 23 and Napoli played a lot of 1B too, though I categorize him as a catcher) are OFs & 1B. Only 2 other infielders in the top 23.
2010 ballot looks similar. 9 of top 10 are OF/1B and 16 out of 20 (I counted Thome in there even though he was all DH).
Either the defensive metrics are skewed or there's no positional adjustment being made. I suspect the latter is true and that the former is true to a small degree.
   6. Carl Goetz Posted: January 11, 2018 at 11:04 AM (#5604597)
Duh, I just realized there's no comparison to Average or Replacement in RPA. It appears to be a straight rate stat. That's the source of the skew.
   7. Carl Goetz Posted: January 11, 2018 at 11:53 AM (#5604652)
He does say above median in the above ranking actually. Is that median of all players? Or median at a player's specific position?
I think what's happening here is that on defense, a SS is compared to other SS (though I'm unclear if an average defensive SS would have a positive defensive rPA or a 0 defensive RPA) while on offense RPA is a straight rate stat. He's then converting the rate to runs and comparing all players to the median of all players.
caiman, please correct me if I'm wrong on your methodology. And please let me know what an average defenders defensive RPA would be.
   8. ThickieDon Posted: January 12, 2018 at 01:02 PM (#5605484)
Preliminary Ballot

1. Bautista
2. Kemp
3. Ellsbury
4. Braun - post-season bump
5. M. Cabrera - post-season bump

6. Halladay
7. Zobrist
8. Pedroia
9. Votto
10. Cliff Lee

11. Granderson
12. A. Gonzalez
13. Kinsler - post-season bump
14. Verlander - league-leading 251 IP (seventh most since 2000)
15. Napoli - 2nd in wRC+; post-season bump
   9. caiman Posted: January 12, 2018 at 10:37 PM (#5605824)
Carl: While RPA does rate players position-by-position, and I do that every year, as my primary analysis, in this matter I use RPA against the median of all players.

When I rate players on this MVP list, I do so, not from the position-by-position comparison that I do each year, but on an absolute run production value vs. the median.

I take the MVP value, pitching, hitting and defense from the team run production.

Jose Bautista, when I did the Toronto, 2011 team run production, he contributed +51.78 runs as a hitter (above the median) and was -2.17 runs on defense for the aggregate of +49.61 runs of value.

The overall offense of Toronto in 2011 was -19.07 runs (which was -17.78 hitting and -1.29 runs on defense).

When, in the final analysis, we are judging the players value, we cannot ignore the fact that 1B and the OF is where most of the big hitters congregate, and the middle infield is where a higher percentage of the weak hitters congregate.

We cannot ignore the fact that it is the OF where the greatest impact takes place on defense as well - both positive and negative.

RPA allows me to look at the player BOTH WAYS - as against players at the same position and as producers within an entire team.

As for Lorenzo Cain, the Mets would have been better off trying to sign Cain, than Jay Bruce. I have Bruce a the 15th best RF in MLB, which places him, barely, in the 'premier group' of RFs, as the last one 'in'.

I have Cain rated as the 6th best, overall, CF in MLB, which places him in the upper half of the "Premier group", and solves the biggest problem for the Mets, because they have no qualitiy CF on the roster. I rated Conforto as the top LF in MLB, but he appears to be a poor defensive CF. With Bruce signed, it will force the Mets to play Conforto in CF, which lowers, in effect, the value of Conforto and the signing of Bruce. The signing was a lose-lose, whereas a Cain signing would have been a win-win.
   10. caiman Posted: January 12, 2018 at 11:41 PM (#5605843)
On more thing that complicates the Mets situation in LF:

Yoenis Cespedes is rated the third best LF in MLB and - Get this: Brandon Nimmo, in too few computed plate appearances, just 318 CPA's over two years, I have rated the 8th best LF!!!

An incredible logjam in LF! However, none of the three appear to be able to play CF and the only one who would be able to play RF with plus defensive ability is Cespedes, but he wants to stay in LF. Conforto would have to move to RF, in that situation, where he'd still be a defensive deficit, but not as much as in CF.

Both Cespedes and Conforto have injury histories, so if the Mets had signed Cain, Nimmo would likely have gotten lots of playing opportunities. With Cain, they would have had 4 premier OFs, and the 4th would be the perfect protection for the Mets from an injury to Cespedes and Conforto.

Bruce is not a bad player, but he does not fit the Mets need.
   11. Carl Goetz Posted: January 13, 2018 at 12:48 AM (#5605855)
"We cannot ignore the fact that it is the OF where the greatest impact takes place on defense as well - both positive and negative."

It seems odd that teams would be willing to accept poor hitters at positions that are also unimportant defensively. Don't you think?

Absolutely no one manages a team on the theory that the defensive spectrum is C-LF-RF-CF-3B-SS-1B-2B.

Who knew that the Phillies should have "hidden" Greg Luzinski's glove at 2B?
   12. Carl Goetz Posted: January 13, 2018 at 10:01 AM (#5605886)
It sounds like you are fully charging outfielders for doubles in their 'area'. Most doubles though are hard hit balls in a gap or down the line that the fielder didn't have a great opportunity to make an out on. There is a lot higher volume of balls that a good SS or 2B can make a difference on by turning a single into an out (or an out into a DP) or a bad SS or 2B can turn an out into an error or single.
   13. Leroy Kincaid Posted: January 13, 2018 at 11:23 AM (#5605898)
"We cannot ignore the fact that it is the OF where the greatest impact takes place on defense as well - both positive and negative."


Maybe he means that OFers "prevent" or "allow" more extra base hits(?)
   14. caiman Posted: January 13, 2018 at 02:27 PM (#5605952)
Carl: You accept the theory that the old way is the proper way? Yikes!

Stop with the 'straw man' argument about Luzinski! I do not know how good or bad his defense was in LF. The likely result, if he was awful in LF, was to move him to first base.

In any case, I have a much better example: Kevin Mitchell playing short for the 1986 NY Mets.

Rafael Santana, NY Mets SS, was the worst hitter on the 1986 team, costing his team an amazing -18.63 runs on offense alone.

I was a huge Mets fan at the time, having watch almost every game since their birth in 1962.

I was appalled, not just by Santana's pathetic offense (.065 RPA!!!), but by his pathetic defense as well.

I have not done the analysis of the play-by-play data for 1986, so my opinion is purely based upon my visual response at the time to what I saw as Santana's lack of range and lack of athleticism.

Manager Davey Johnson was clearly concerned about Santana's deficiencies and decided to often play HoJo and Kevin Mitchell at SS.

I especially loved it when Mitchell played SS. NO RUNNER TRIED TO TAKE HIM OUT! He was just too intimidating a figure when trying to break up a DP! And he could hit - big time!

At the time, I truly thought Mitchell was every bit as good, if not better on defense, at SS, than was Santana.

Most ss, up to that time were little guys, which was a prejudice not based upon anything other than belief.

I believe, correct me if I am wrong, that it was 6'4", 240# Cal Ripken who revolutionized the perception of what a SS looks like, by being both a terrific hitter as well as a terrific defender at SS.

What a science does, is deal with facts, not tradition.

The fact is that the OF who is slow and lumbering cannot be 'hidden' in LF, unless he has a huge bat, but even in that case, the team is probably better off trying to get him to play first base, where the penalty for his defensive liabilities would be far less.

The outfield is where extra-base hits are commonplace. Grounders up-the-middle are only singles. Fleet footed and quick OFs save an enormous number of runs. No infielder, in MLB, came remotely close to Byron Buxton, in terms of saving runs for his team.
   15. Carl Goetz Posted: January 13, 2018 at 05:26 PM (#5606026)
Cal Ripken was a good defensive SS. Fail to see the relevance. The Orioles weren't hiding Cal, they were playing their best defensive SS at SS.

"The outfield is where extra-base hits are commonplace. Grounders up-the-middle are only singles. Fleet footed and quick OFs save an enormous number of runs. No infielder, in MLB, came remotely close to Byron Buxton, in terms of saving runs for his team."

On defense alone, Andrelton Simmons is far more valuable than any OF. There are a few players in 2017 who made up that difference with their bats though.

"Maybe he means that OFers "prevent" or "allow" more extra base hits(?)"

Clearly, but I fail to see why a LF should be credited with a 2B prevented every time he catches a routine fly ball. What I'm trying to decide is if this guy is for real or is simply trolling us.

Question: I'm relatively new to the MMP voting. Is there any rule that says you have to treat players fairly across positions? Or is a ballot with all OF/1B considered legit? I know there's a rule in HoM about being fair across eras, but couldn't find anything about positions. It certainly seems against the spirit of this exercise to only consider OF/1B/P.
   16. caiman Posted: January 14, 2018 at 12:26 PM (#5606219)
Utterly ridiculous! Removing Ripken from SS for Mike Bordick?

Yes, Bordick was a terrific defensive SS. So was Ripken. The difference was that, in 1997, Bordick cost the O's 23.00 on offense - an astoundingly horrible number! What a disaster! What a terrible player in 1997! OIne of they very worst player performances in MLB in 1997! The O's destroyed Ripken's value by sending him to a position where both his offense and his defense were not positive factors,as they were at SS. A ridiculous decision.

Andrelton Simmons WAS a terrific SS, when he first came up -- NOT ANYMORE and not for a while! On what basis do you make the claim that Simmons was a great defensive SS in 1997? Almost every decent OF was a better player, in terms of run value, than Simmons in 2017. Simmons was NOT a superior defensive player at SS in 2017. While he has been improving a bit on offense over the years, his defense has badly declined.

Get over it! Positional value is nonsense! Games are decided, for the most part, by the big hitters and the big pitchers. The big bats carry the team by carrying their team scoring. It is not about 'fairness'!!! It is about the reality on the field. Those big bats are located at 1B and the OF and sometimes at DH and 3B. A few are located at 2B and SS, such as the Houston combo, but that is less common. Kris Bryant and and, to a lesser extent, Anthony Rizzo carried the Cubs in 2017. Without those two, the Cubs were a very ordinary team in 2017.

Defense matters, of course, but it has to be put into perspective and based on actual measure of runs affected, as compared to all other players at that same position on defense, against that same team at that same stadium.

Stop with the nonsense about fly balls. Defense is just the reverse of offense. Almost as many balls are hit into each of the three OF positions as are hit into 2B and SS. Only 1B and 3B get a significantly smaller number of chances.

With 500+ balls being hit into ea. OF position , each year, you get a 'hitting line' vs. that OF. Yes, individual seasons, just as with the hitter, can be 'lucky seasons' for a hitter or the defender, but we accept them for the hitter. Why not for the defender? If one OF is letting lots of doubles drop, and another doesn't, what do we account the difference to? For the hitter and for the defender? Luck? If so, then statistical analysis and all science is thrown out the window and we can go back to the 'good old days' of neanderthal baseball analysis.

   17. caiman Posted: January 14, 2018 at 12:38 PM (#5606223)
corrected my typo in the Ripken/Bordick from year 2017 to 1997.
   18. Carl Goetz Posted: January 14, 2018 at 02:54 PM (#5606273)
Nevermind. The black box you programmed years ago and can't currently explain how it works clearly spits out nothing but truth. We don't need to do any follow up analysis or even double check the results. "RPA" clearly equals "Fact". Science has prevailed thanks to your caps lock and repeated use of the word "Yikes!". Clearly every statistical analyst, manager, GM, scout, player, coach, ball boy, and fan besides you for the last 150 years was wrong to think SS was a valuable defensive position and LF was less so. Thank you for showing us all the light.
   19. caiman Posted: January 14, 2018 at 08:12 PM (#5606450)
Nice answer. Filled with so much factual material. And what 'black box' are you referring to. I published the formulas, with full explanations. I answered all questions given to me here on this site. On what basis, in statistical analysis, do your 'facts' show us that you are right?

Yeah, Managers, scouts, coaches, fans, GM's were WRONG (for 150 years!) when I started advising MLB teams decades ago. They thought I was 'outside the game' with my strange numbers, but they have been proven WRONG.

However, we are still in the 'learning stage' of statistical analysis of player performance.

Until YOU can advance, with research, your findings, you have no credibility with ME.
   20. DL from MN Posted: January 14, 2018 at 09:08 PM (#5606480)
Almost as many balls are hit into each of the three OF positions as are hit into 2B and SS.


The absolute number of chances doesn't matter. The difficult plays matter. SS and 2B have more difficult plays just due to reaction time. No, not every ball hit to the outfield turns into a double if it isn't caught. A pop-up missed usually turns into just a single.
   21. caiman Posted: January 15, 2018 at 01:17 AM (#5606505)
NO, the difficult plays DO NOT MATTER!

The difficult plays would only matter if they were difficult for only one team.

However, all shortstops, on both home and visiting teams, face the same difficulty.

What separates them, as with the outfielders, is how many they turn into outs vs. the other player at the same position. There are lots of easy plays, as well at SS and 2B!!

While the SS and 2B have to have quicker reaction, they only cover a smaller territory as a result. The OF needs both quickness and speed in order to cover the vast spaces in the OF and a high percentage of the balls that do fall in go for extra bases. How you can deny that is beyond me.

As I have said, the use of the positional adjustment is nonsense. The DH is equal, ON DEFENSE, to the median starting SS and Starting 2Bman. There are 30 MLB teams. Only 15 starters, at any position can be a plus on defense. The median starter therefore MUST be rated at ZERO on defense. That's the same rating, on defense, for a DH. ZERO. That means that 15 of the 30 starting SS and 2Bmen MUST have a negative defensive rating.

By the way, the addition of Stanton to the Yankees is an enormous plus for them, but do not count out Cleveland just yet. The 2017 Cleveland Indians were the most dominant AL team since the 1969 Baltimore Orioles! And the most dominant MLB team since the 1976 Big Red machine!! Unfortunately for the Indians, they'll need their pitchers to repeat their fabulous 2017 season because, even before the acquisition of Stanton, the Yankees were not far behind the Indians, on offense, in 2017. Pitching, in the WS, only dominates if the opponent is average on offense. The best hitting teams dominate in the WS (and, I assume, in the playoffs, although I only did a study of the WS teams).
   22. caiman Posted: January 15, 2018 at 01:33 AM (#5606511)
One addendum to my last note:

The Astros were the third best AL team, after Cleveland and NY, but they were the best offensive team in MLB in 2017, once again showing the dominance of offense in the post season. The Astros were only slightly, overall, behind the Yankees, when taking pitching into account, but it was the Astros offense that had carried them all season, not their pitching.
   23. Carl Goetz Posted: January 15, 2018 at 09:30 AM (#5606545)
"However, all shortstops, on both home and visiting teams, face the same difficulty."
Right, but if all the best fielders are congregated at one position, there's going to be less variance at that position. LF is populated by the Brett Gardners and Alex Gordons of the world, but also by the lunks who are only in for their bats. Though I agree that there are less lunks in LF nowadays than there used to be. This creates more variance, not more value. When you are taking a sample of SS, you're sample is skewed towards good fielders because nobody hides a lunk at SS.
"As I have said, the use of the positional adjustment is nonsense. The DH is equal, ON DEFENSE, to the median starting SS and Starting 2Bman."
And we've identified the problem with comparing players to average instead of replacement. When comparing to average, it stands to reason that a DH would rate the same on defense as an average player at any other position. All DH's are average fielding DH's after all. However, when such a model is compared to realty, its beyond idiotic to say that a league average fielding SS (or any other position between the lines) has the same defensive value as someone who doesn't play any defense. Such a model would have to also conclude that any pitcher who doesn't bat in a season is a league average hitter. Point is, players shouldn't be rated as average just because they don't do something. My defensive RPA is zero. Does that make me a league average MLB defender?
"That means that 15 of the 30 starting SS and 2Bmen MUST have a negative defensive rating."
But, as you've pointed out, 15 teams need to field a below average defensive SS. That very fact means that below average defenders have value to major league teams. And what does "Below Average" actually mean? (See below)

Just a general statistics comment: Using average to compare major league players to each other is flawed. The distribution is skewed. MLB talent is not a normal distribution. Perhaps baseball talent on a worldwide population basis is on a normal distribution (I see no way to test this), but lets say it is. Major league talent is on the far far right tail of that normal distribution. On top of that, even the distribution of playing time amongst major league SS (for example) is skewed towards the better half of that sample. If we assume there are 60 rostered SS, we have to assume that roughly the "Above Average" SS are the 30 that are starting. Of course, that's above average when we combine both offense and defense so that would get hairy when we try to determine an average offensive SS and an average defensive SS. Side question, if all positions are compared to each other and given the same base value, why not on offense? Why are SS not compared to SS when hitting?
   24. DL from MN Posted: January 15, 2018 at 09:36 AM (#5606548)
I still suspect RPA's accounting for double plays is fouled up. This is another reason SS has a lot of value defensively. A SS who commits an error on a ground ball could potentially be throwing away two outs. An outfielder who drops a pop fly probably just gave up a single and the baserunner didn't advance because they thought the ball would be caught.
   25. Carl Goetz Posted: January 15, 2018 at 11:22 AM (#5606593)
"I still suspect RPA's accounting for double plays is fouled up."
I agree, though I suspect its not the biggest problem here. No matter how DPs are counted, an average SS rating the same on defense as a DH is a much bigger problem.
   26. Carl Goetz Posted: January 15, 2018 at 12:30 PM (#5606632)
"NO, the difficult plays DO NOT MATTER!

The difficult plays would only matter if they were difficult for only one team.

However, all shortstops, on both home and visiting teams, face the same difficulty."

Just reread this one and hadn't thought about it until now. You're claiming that all SS face the exact same difficulty of plays. Random chance wouldn't allow for that, let alone differences in pitchers on the mound (Lefty vs Righty, ground ball vs fly ball) and hitters at the plate. You are implicitly assuming every SS is facing average offenses (in terms of quality and hitting location tendencies). And it seems you are assuming every ground out to SS and hit that goes past the SS are all of average difficulty and that the difference between hit and out is purely the skill of the SS. I suspect a SS who moves several steps and makes a diving stop up the middle throwing out the runner is being rated the same for that play as he is for the routine grounder right at him. But Carlos Correa plays for an excellent offensive team, that means the offenses he faces while in the field are slightly below average (league average offense with the Astros removed). If we analyzed deeper, some offenses (regardless of quality) will have greater tendencies for hitting balls to SS and some will have lesser. I suspect not controlling for the quality of team is leading to the park factor issues as well. ie a high offense team makes their home park look like a bigger hitters park than it is.
   27. Mike Emeigh Posted: January 15, 2018 at 01:02 PM (#5606653)
However, all shortstops, on both home and visiting teams, face the same difficulty.


Actually, they don't (although I admit that I don't know how much variance there is).

Charlie Saeger and I did some research on this a few years ago, and I doubt it's changed much in the era of shifting. The locations of balls in play (and presumably the degree of difficulty, as fewer plays are made in the hole than on balls hit directly to SS, with balls up the middle being in-between) vary depending on GB-FB mix and left-right hitter mix. GB pitchers tend to allow a higher percentage of easy GB for the SS, and right-handed hitters tend to hit a higher percentage of easy GB to SS. Note that we didn't have velocity numbers, which would impact this - only at zone-level ball distributions.

-- MWE
   28. Carl Goetz Posted: January 15, 2018 at 02:13 PM (#5606710)
"On what basis do you make the claim that Simmons was a great defensive SS in 1997?"
I didn't; he was 8 years old in 1997.

In 2017 however, he led all SS in DRA, UZR and DRS rating. A couple CFs (Cain and Inciarte) did save more runs though, so I'll walk my statement about him being more valuable than all OFs back slightly.
   29. caiman Posted: January 15, 2018 at 07:28 PM (#5606944)
Carl: is that your answer on Simmons: a typo (1997 instead of 2017) wins your argument. Childish!

The defensive rating for any player is not against the median or against the average defensive player at that position.

It is against the composite of all players who he competed against in those games, home and away, in that season.

Andrelton Simmons played for the Angels, in one of the easy parks to play defense in.

Andrelton Simmons played behind a a very good starting rotation that included 5 righties and one lefty (Tyler Skaggs) who only was there for a half season. The five righties, in order of quality performance in 2017 were: 1. Parker Bridwell 2. JC Ramirez 3. Matt Shoemanker 4. Jesse Chavez 5. Ricky Nolasco. Tyler Skaggs was the 6th best and lone lefty.

As such, with a righty dominant and good starting pitching staff, it would be lefty hitters that would put the ball in play the hardest. That would mean that the 2B and not the SS would get the most difficult plays.

Here is the actual performance of Simmons, in 2017, versus the SS on his opposing teams, at home and away:

At Home: Simmmons RPA was .074, the opponent's was .070

At the away parks: Simmons was .082 RPA and the opponents composite was .083.

The opponents were better at Simmons park, but Simmons was better at their parks. Another example of opponents having an advantage at easy to defend parks.

Please note the difference between the Angels home park and the composite away parks. The RPA's are lower for the Angels park! That's the proof it is an easier park to play defense at.

I agree with the criticism that my RPA has not, up till now, added an adjustment for the type of pitcher, that the defensive players perform behind. That is something that could improve my defensive rating and I totally agree. It is not a new idea for me, but it would take an enormous effort on my part to make that adjustment, after first testing to see what adjustment was valid. If you want to improve the RPA, you'd need to do the appropriate tests and results that can I can implement.

However, one must start somewhere, and the only way to start is with the actual results. That is what RPA does. Can it be refined further - or even replaced? Of course! All models are, by definition INCORRECT - but we use them because of their usefulness.

Nope. the #16-30 starting shortstops ALL remove runs from their team, by DEFINITION! They hurt your team!

The OVERALL RPA positional ratings are based upon that individual qualified starting player vs. the 30 starting player performances. The qualified starters are my highest 30 rated players at a position, split into two parts, the top 15, who are the 'premier players' at their position and the bottom 15. Those outside the top 30, are the bench performers, who reduce the possibility for their team even more than any of the bottom 15 qualified starters. Only the top 15 can add runs.

Here is the results for the MLB SS position for the upcoming 2018 season:

The median (not the average) is the dividing line between the top 15 and bottom 15. That median is .128 RPA. The overall RPA rating is based upon hitting,defense and a very small age adjustment over the last two seasons, 60% for the last season and 40% for the prior season.

1. Carlos Correa .170 RPA. Every plate appearance for Correa predictive production is +.042 runs per RPA (over the median starting player)
2. Trea Turner .164 RPA = +.036 per PA
3. Zack Cozart .159 RPA = +.031
4. Corey Seager .153 RPA = +.025
5. Paul DeJong .145 RPA = +.017
6. Francisco Lindor .142 RPA = +.014
6. Jean Segura .142 RPA = +.014
8. Elvis Andrus .137 RPA = +.009
9. Didi Gregorius .134 RPA = +.006

As one goes past number 15, you start to subtract runs, from your team.s production, for every plate appearance.

The DH example was purposefully ridiculous: to show how ridiculous is the positional adjustment! No effort is made by RPA to include defense in the DH rating, OBVIOUSLY. It is just another silly criticism to try to divert my criticism, into comparing it to the pitcher not hitting. Silly! All I am sayng is that the DH, by definition, cannot hurt you on defense, while 15 of the qualified starters will do just that!







   30. DL from MN Posted: January 15, 2018 at 08:06 PM (#5606971)
Those outside the top 30, are the bench performers, who reduce the possibility for their team even more than any of the bottom 15 qualified starters.


This might be true for the truly difficult defensive positions like catcher, SS and CF. However, the backup outfielder is often BETTER than the starting LF on defense. Teams routinely carry a 4th outfielder who can play CF and replace the poor defending slugger in late innings.
   31. caiman Posted: January 15, 2018 at 09:19 PM (#5607001)
DL: what does that have to do with what I said?

Each player is compared on his overall production. The top 30 are based on that. What the 4th or 5th OF does has an effect on the team's results, but not on whether the player is in the top 30 in MLB, at that position.

In the case off LF, I could care less about anything other than those top 30. However, in rating the individual player on defense, he is measured against the composite of ALL the players on defense in LF, in the 162 games that team was in.

If I am running a team, I could care less about the 'Olympic gymnastics' scoring of the LF. All I want to know is how much better or worse is that player vs. all the LF's we have to face, in those 162 games, in regards to his COMPLETE GAME PRODUCTION.

Last season, despite Alex Gordon getting a projected +11 rating in RPA, on defense, his offense was so putrid that I do not have Gordon listed in the top 30 in LF for the 2018 season. His performance, over the last two seasons, does not make him qualified to be a starting LF in 2018. He'd have to show that his offense has improved in order to work his way back into the top 30, des[pite still being a wonderful defensive LF.
   32. caiman Posted: January 15, 2018 at 09:37 PM (#5607006)
DL: If the backup OF or backup IF or Backup C, provide positive defense, it will all be placed their RPA rating - and - in some cases, if the player is fabulous on defense, his overall rating may be higher than the starter and place him in the top 30, despite being a non starter.

My RPA ratings specialty is in finding those non starters WHO SHOULD BE GIVEN A STARTING JOB! In fact, every year, I identify non-starters who should be starters.

The best example from 2017 is Tommy Pham. Three years ago I contacted Milwaukee sports radio broadcaster Steve "Homer" True, that there was a player, Tommy Pham, whom the St. Louis Cardinals did not realize just how good he was, that the Brewers should pursue. in my overall LF RPA ratings for the upcoming season, I have Pham rated the #2 LF in all of MLB!!!

Matt Szczur is not a starter, but I have him as a qualified starter, as the 12th best LF in MLB for the upcoming season. That puts him in the 'premier group' and the thing that got him there was his +12 on defense, along with a +2 age adjustment and a decent .125 offensive RPA for a total RPA of .139. The median RPA for LF was .138, so although Szczur is just a +.001, it still means that not only should he be given a shot at starting, he figures to be a plus performer, within the top 15, even if only barely.
   33. Carl Goetz Posted: January 16, 2018 at 08:13 AM (#5607094)
"In the case off LF, I could care less about anything other than those top 30. However, in rating the individual player on defense, he is measured against the composite of ALL the players on defense in LF, in the 162 games that team was in."
But we don't have balanced schedules. What if a specific SS happened to play against above average SSs? It would make him look less dominant by comparison.

"Carl: is that your answer on Simmons: a typo (1997 instead of 2017) wins your argument. Childish!"

No. this part wins my argument. "In 2017 however, he led all SS in DRA, UZR and DRS rating. " Every viable defensive measure in existence says he's easily the best. And for 5 years running (though other got closer in a couple years). And again we need to establish that RPA works before quoting RPA results changes my mind about a player. I've cited many issues with RPAs handling of defense, particularly middle infielders. Not trying to be combative, but I didn't just trust WAR (or WS or Kiko's Player Wins) out of the gate either.

"The DH example was purposefully ridiculous: to show how ridiculous is the positional adjustment! No effort is made by RPA to include defense in the DH rating, OBVIOUSLY. It is just another silly criticism to try to divert my criticism, into comparing it to the pitcher not hitting. Silly! All I am sayng is that the DH, by definition, cannot hurt you on defense, while 15 of the qualified starters will do just that!"

The DH example is purposefully ridiculous, to show how important a positional adjustment is. Its true that the DH can't hurt you on defense, because he doesn't do it! I can't hurt you at Brain Surgery because I don't do it. That doesn't make me of equal value at brain surgery as the average brain surgeon. When you ignore the value of an average SS (set it to zero; the equal of a DH), that means you are ignoring defense for that player. This is actually worse than just ignoring defense, you are ignoring defense only for select players and punishing actual major league players for playing defense. But at least you are rewarding the best LFs & Rfs on defense.

"Matt Szczur is not a starter, but I have him as a qualified starter, as the 12th best LF in MLB for the upcoming season. That puts him in the 'premier group' and the thing that got him there was his +12 on defense, along with a +2 age adjustment and a decent .125 offensive RPA for a total RPA of .139. The median RPA for LF was .138, so although Szczur is just a +.001, it still means that not only should he be given a shot at starting, he figures to be a plus performer, within the top 15, even if only barely."
What is the age adjustment for? Are you rating players based on the season that happened? Or predicting the upcoming season? Is 2017 RPA rating players on 2017? Or projecting for 2018?
   34. DL from MN Posted: January 16, 2018 at 10:03 AM (#5607149)
However, in rating the individual player on defense, he is measured against the composite of ALL the players on defense in LF, in the 162 games that team was in.


That is not how I understood the method you had described above. You mentioned everyone below the 15th best player as negative. Now you are saying they are measured against the composite of all players. Your explanations are confusing.

You seem to be much more interested in "winning an argument" than explaining your system.

WAR has a positional difficulty adjustment. This is built into the system because you can take darn near any MLB SS and put him in LF and get average LF defense but you can't do that swapping a LF into SS. This makes a player with those athletic skills more valuable on a 25 man roster.
   35. bjhanke Posted: January 16, 2018 at 10:11 AM (#5607155)
The header doesn't have WAR for Felix Hernandez. The number is 3.7, just for everyone's use. As for Greg Luzinski, I just assumed that the reason he was playing left field was that he was blocked up at first base by John Kruk. Bad as Luzinski was on defense, the Phillies apparently thought that Kruk would be worse. If you've ever seen John Kruk play defense, you might not think that the Phillies were idiots there.
   36. Carl Goetz Posted: January 16, 2018 at 10:44 AM (#5607172)
Luzinski was a little before Kruk, though you're point remains the unchanged. Luzinski was "blocked" at 1B by the likes of Willie Montanez, Dick Allen, Richie Hebner, and Pete Rose; all of whom where atrocious defenders (at least during the period they were starting 1B for the Phillies; I haven't looked at their whole careers). These guys may have even made Kruk look adequate in the field.
Thanks for the King Felix update btw.
   37. bjhanke Posted: January 16, 2018 at 11:51 AM (#5607249)
Carl - Thanks. I started to worry about my memory and looked Luzinski's career up, and you have correctly identified the people who played first while Greg was in the outfield, about 3 years out of each of them. Greg's last year was 1984, while Kruk debuted in 1986. All of them, as you noted, were older than Luzinski, although, except for Rose, they were generally only 3 years older or so. I have no idea why Luzinski played LF in preference to any of them (except Rose), and especially to Montanez and Allen, who could at least run and had arms. Luzinski did get started in the majors very young - only 19 - and maybe they thought he'd learn to play left as he got older.
   38. Carl Goetz Posted: January 16, 2018 at 12:14 PM (#5607276)
They probably thought that HRs were in short supply at the time and Garry Maddox makes up for a fair amount of poor leftfielding.
   39. caiman Posted: January 17, 2018 at 11:29 AM (#5607987)
The age adjustment only has to do with the player's age. Nothing more. All players at the same age will have the same age adjustment. It is just a measure of the average improvement or decline for a player of his age.
   40. Carl Goetz Posted: January 17, 2018 at 11:38 AM (#5607992)
But why would an age adjustment be necessary to evaluate what happened in a given past season? An age adjustment is only necessary when predicting the coming season. Its irrelevant in evaluating in 2018 how well a player played in 2011.
   41. caiman Posted: January 17, 2018 at 11:44 AM (#5607998)
Carl: I cannot believe that a simple concept is so difficult to understand!

If brain surgery were a part of the sport of baseball, and there were 30 brain surgeons, one per team, would we give the brain surgeon greater value than the SS? Of course not, no matter how skillful or rare the skill.

A team would want to have one of the 15 best brain surgeons on their MLB team, not one of the worst 15. The worst 15 would all be DEFICITS, no matter how skilled or rare their skill.

If the MLB team also was required to have, at the other end of the skill set say, a ditch digger, the team would want to have one of the 15 best ditch diggers, not one of the worst ditch diggers.

The value of the brain surgeon or the ditch digger, to the team, could only be measured against their own skill position, be it brain surgeon or ditch digger. That is all I am saying about the value of the DH. The median SS, the median Brain surgeon, the median ditch digger value would be zero BY DEFINITION, whether it be for offense or for defense.

You are the one including a value for the DH on defense, NOT ME!

In order to rate the positions differently, on defense, you are MISTAKENLY giving a value to the DH on defense!

How did you arrive at that value, for each position and how did you test that value? In so doing, you are destroying both logic and math.
   42. caiman Posted: January 17, 2018 at 11:55 AM (#5608020)
DL: Huh? I repeatedly stated that the player is measured against the composite of all players, at that position, at the same park.

The top 30 refers to the final ratings FOR ALL players at that position, not just the top 30. It is the rating for the player after I have evaluated his offense, his defense and his age to get a final overall RPA rating.

There are way more than 30 SS in MLB, but only 30 are qualified to be starters, simply because there are 30 teams. Of those 30 qualified starting SS, only 15 can help your team, simply because they are better than the worst 15 qualified starters.

The measure of value is not something ethereal. Value is measured only against something tangible. You win or lose games against those other teams. You measure your SS vs. their SS.

In MLB, you need to measure a player's value against the players who start at that same position. Are they in the top 30? If so they are qualified to be a starting player at that position.

Then you must ask "where in the top 30 are they located"? If in the top 15: "Yippee". If in the bottom 15: "How do we repair this problem?"
   43. caiman Posted: January 17, 2018 at 12:12 PM (#5608050)
DL: You say that you can take a SS and place them in LF. So what? Why is that important when evaluating a players current value to a team? If they are not playing the position, there is no reason to evaluate a player at that other position that they do not perform at.

In any case, there is a reason why the average SS CANNOT PLAY LF: They can't hit a lick!

Why don't MLB managers simply bench their slugglish LF? Because that sluggish LF can hit! They have made the calculation, when taking into account offense and defense, that it would be suicide to play the SS in LF.

So lets play a terrific defensive SS in LF, like J.J. Hardy, who is a +16 on defense at SS, but can't hit a lick. That would be a disaster, You would LOSE the defense advantage that Hardy gives you at SS, and I do not believe that he'd likely be a big plus on defense in LF, if any plus at all in LF, while making his offensive woes even more glaring in LF, because of the better hitters at that position. It would be a lose-lose situation. You'd lose on both the offensive and the defensive value for Hardy!

In other words your very example speaks loudly against your argument. You cannot swap a SS into LF any more easily than you can swap an LF into a SS. In fact, I already gave you an example of that swap in the Kevin Mitchell example from 1986. The fact that some swaps can happen both ways has nothing to do with the position value. It has to do with the individual player value.
   44. caiman Posted: January 17, 2018 at 12:16 PM (#5608061)
Carl:

The age adjustment is NOT for evaluating past performance. As stated, my RPA ratings, each year, for players at each position, are for the coming season, after evaluating the player's past two seasons. As such I add or subtract points for the age of the player.
   45. MrC. Posted: January 17, 2018 at 12:33 PM (#5608092)
2011 Preliminary Ballot

Hitters: Use a WAR framework with value added runs as the basis for RAA. Adjust for position, non SB base running, DRS defense. I only use 60% of the replacement runs used by Fangraphs and Baseball Reference. Convert this to wins (a stat I call Wins above reduced replacement: WARR). Post season adjustment based on WPA for the post season (from Baseball Gauge).

Pitchers: Again use value added runs for RAA. Use similar adjustments as baseball reference (strength of competition, role, team defense and leverage for relievers). Convert to wins and post season adjustment same as for hitters. Also add in the Pitchers WARR for hitting.

1. Jacob Ellsbury 8.82 WARR A career year. Great offense with above average defense.
2. Ryan Braun 8.27 WARR Top 3 offensive player in NL, above average fielder and a very good post season.
3. Miguel Cabrera 7.81 WARR best offense in AL, below average defense, but massive post season
4 Matt Kemp 7.79 WARR Slightly better offense than Braun, but poorer defense
5. Cliff Lee 7.75 WARR Interesting battle between Lee and Halladay. Lee was the poorer pitcher, but a much better hitter, had the poorer post season, but still eeked out a slight victory over Halladay.
6. Jose Bautista 7.73 WARR A great offensive season
7. Roy Halladay 7.72 WARR
8. Ben Zobrist 7.60 WARR Solid offense, great defense, poor post season
9. Jason Verlander 7.31 WARR best pitcher in AL by a whole bunch
10. Evan Longoria 7.04 WARR A great defense year.
11. Joey Votto 6.91 WARR Was comparable to Braun and Kemp offensively, but was apoor base runner playing a hitter's position.
12. Dustin Pedroia 6.87 WARR A great defensive season for both Pedroia and Gordon. Nothing really to choose between the two.
13. Alex Gordon 6.87 WARR
14. Clayton Kershaw 6.54 WARR Similar raw stats as Lee and Halladay but pitched behind a better defense in a more pitcher friendly park.
15. Jered Weaver 6.42 WARR

Rest of the top 20

16. Adrian Gonzalez
17. Ian Kinsler
18. Robinson Cano
19. Mike Napoli
20. Ricky Romero



   46. Carl Goetz Posted: January 17, 2018 at 12:37 PM (#5608098)
"A team would want to have one of the 15 best brain surgeons on their MLB team, not one of the worst 15. The worst 15 would all be DEFICITS, no matter how skilled or rare their skill."

I'm guessing the 30 best brain surgeons are all really good, just like the 30 best SS are all really good. 30 SS are starting for MLB teams; none are a deficit to their team. You are assuming a player has negative value if he is below average. Clearly MLB teams disagree since they are paid well above the minimum. You are correct that a team would want to have an above average player at every position, but that is really heard to do, so even the best teams field below average players. They still have value since until you reach replacement level, their replacement would be worse. A 1 WAR player is certainly below average, but still has value because if he were replaced with a minor leaguer at his position, his team would lose 1 Win (again this is an estimate).

"You are the one including a value for the DH on defense, NOT ME!"
Actually we are both including the correct value on defense for DH, Zero. The problem is that you are rating an average defensive SS (who has defensive value) as equal to zero as well. Clearly that average SS has more value on defense than a DH, because he plays defense.

"In order to rate the positions differently, on defense, you are MISTAKENLY giving a value to the DH on defense!"
Again, zero is a value and it is the value we are both placing on the DH's defense.

"How did you arrive at that value, for each position and how did you test that value? In so doing, you are destroying both logic and math."
I didn't. I use the work of others on this front and I use my understanding of statistics to evaluate their methods. I do place trust in them that they are faithfully and accurately implementing the process that they've described. You claim to value logic and math, but you don't seem to understand that:
a)zero is a number.
b)that a negative number is less than zero.
c)a&b together mean that you are giving half the players in baseball that play defense a lower defensive rating than DHs.

I cannot make this more clear. So this will be my last attempt to try.




   47. Carl Goetz Posted: January 17, 2018 at 12:42 PM (#5608103)
"The age adjustment is NOT for evaluating past performance. As stated, my RPA ratings, each year, for players at each position, are for the coming season, after evaluating the player's past two seasons. As such I add or subtract points for the age of the player."
So you're voting on players for 2011 MMP based on your projections for the 2011 season? We know what happened in 2011 already. We don't need projections.
This would explain the Dan Uggla being rated better than Robinson Cano issue from 2010. He's rating 2010 based on 2008 & 2009 plus an age adjustment. Based on a combo of 2008 & 2009, I would rate Uggla better than Cano as well (even without using his park factors and minimized 2B defense). Problem is that we were voting on what they did in 2010.

Problems with defense and positional adjustments aside, surely MMP voters are required to vote on what the players did in the year for which we are voting!
   48. caiman Posted: January 17, 2018 at 12:52 PM (#5608113)
Another example would be trying to transfer Troy Tulowitzki to LF.

JJ Hardy did not even make my top 30, but Tulo did, but in the bottom 15.

Tulo's offense + defense + Age accounted for his RPA rating. His offense was evaluated at .106. His efense was a +6 and the age adjustment a +1 for a total RPA of .113.

The median SS, among the top 30, for the coming season is .128. That would make Tulo a -15 at SS. What would that be in LF, even if, doubtfully, he was able to maintain the +6 in LF?

Tulo would be a -25 because the median of the top 30 in LF is .138. In order to maintain the -15 rating from SS to LF, Tulo would have to go from a +6 to a +16 at a position where even Alex Gordon only gets a +12. In other words, it is more likely that Alex Gordon could go from LF to SS, than Tulo could go from SS to LF. In either case, it would be a bad move, but it would likely be worse for Tulo than Gordon.
   49. Carl Goetz Posted: January 17, 2018 at 01:03 PM (#5608122)
"The median SS, among the top 30, for the coming season is .128. That would make Tulo a -15 at SS."
"Tulo would be a -25 because the median of the top 30 in LF is .138."

You've maintained you weren't doing any positional adjustments (in fact, you've been railing against them pretty regularly) through a couple weeks worth of debate. Now you are saying that you are?! The method described here is a 10 point positional adjustment between SS and LF for the given year (or 2 prior years, anyway).
   50. caiman Posted: January 17, 2018 at 01:15 PM (#5608134)
Carl: Stop twisting my words!

I do not know what your problem is. Past performance is past performance. HOW DARE YOU IMPLY that i said anything else! 2011 is based on 2011. Period. Stop the nonsense!

As far as 'accepting others work', why not mine? I DID THE WORK. You did not!

I have critiqued their work - NOT YOURS.

What testing did they do? I have tried, to my best to explain how I arrived at my conclusions.

I have looked at their formulas and found them factually incorrect. It is bad statistical analysis.

I have a basis for my criticism, simply because I did the testing and explained my methodology.

I did have an advantage of using dialectical materialism in order to develop my method. Without it, there would be no RPA. That's not the 'bean counting' methods that I see being commonly used. Unfortunately, bad statistical analysis is what is taught in our 'educational' system, based upon rote learning rather than logic and dialectics. That is true of much of what is taught. That is why I have put out my videos opposing the current accepted nonsense in theoretical physics. Sorry for the digression, but its a problem that is much wider than baseball. Statistics are misused and abused, when I watch the media reporting, because so few understand how to use statistics, even those using them! The media quotes a ridiculously low unemployment rate, so widely variant from reality, that I have had to publish, monthly, the REAL unemployment rate figures, based on the US government's BLS monthly published data. Watching the corporate news readers on TV, repeating the nonsense from the BLS, about the unemployment rate, is infuriating.
   51. Carl Goetz Posted: January 17, 2018 at 01:37 PM (#5608153)
"As far as 'accepting others work', why not mine? I DID THE WORK. You did not!

I have critiqued their work - NOT YOURS. "
I don't accept it blindly. They explain their methods and I evaluate them. You've been vague and contradictory

"Carl: Stop twisting my words!"
"The age adjustment is NOT for evaluating past performance. As stated, my RPA ratings, each year, for players at each position, are for the coming season, after evaluating the player's past two seasons. As such I add or subtract points for the age of the player."

You're words are vague and keep changing; don't hand me a pretzel and accuse me of twisting it.

"I did have an advantage of using dialectical materialism in order to develop my method."
As in Karl Marx? I'm not going to touch that one. I use baseball and statistical analysis of it to get away from politics.
   52. DL from MN Posted: January 17, 2018 at 01:47 PM (#5608166)
Moving Tulowitzki from SS to LF probably means he is at least a median LF. Moving Alex Gordon to SS probably means he goes from +12 to being the worst SS in baseball (-15? -20?). Moving a -10 LF to SS could mean a -60 defensive rating at SS or worse. Of course this wouldn't ever happen because the team would quickly get tired of losing. You can see this in practice. When a team's SS gets hurt they replace him with a different player who can play SS acceptably. They don't move the DH and his "average" glove to SS and call up another bat.

"There are way more than 30 SS in MLB, but only 30 are qualified to be starters, simply because there are 30 teams. Of those 30 qualified starting SS, only 15 can help your team, simply because they are better than the worst 15 qualified starters."

That 15th player is not your median "0" performer. The median has to include ALL the innings played at SS by all players. MLB starters DO NOT play every inning at SS. There is value for a team of having the 18th best SS, especially if he is durable enough to make sure you don't need the 56th best guy to play in his place for an extended stretch of time. Given player scarcity an MLB team with the 18th best SS probably is not looking to replace him because there are no freely available replacements for him. Average players get multi-year contracts all the time.
   53. DL from MN Posted: January 17, 2018 at 01:49 PM (#5608169)
As stated, my RPA ratings, each year, for players at each position, are for the coming season, after evaluating the player's past two seasons. As such I add or subtract points for the age of the player."


This sounds much more like a player projection system than a player value system. I will repost the rules, bolding the relevant section.

Candidate Eligibility: Any North American professional baseball player is eligible for the Most Meritorious Player (MMP) award including players on independent teams. Voters should consider the player’s on-field contribution to Major League Baseball (MLB) team(s) in that season only. If part of the season was spent outside MLB, that value may be considered as well. However, the player’s on-field contribution should be judged in relation to the highest level major league, not relative to a minor league. A season may include playoff or World Series games but does not include spring training or exhibition games. No credit will be given for games not played due to injury, wartime service or contract holdouts.

Ballot Length: For 2017, each voter should rank 15 players.

Voter eligibility: All voters who did not vote in the previous year’s election must post a preliminary ballot in the ballot discussion thread at least 2 days before voting ends. All voters must fill out a complete ballot. Voters must briefly explain their ballot choices. One person, one vote; anyone determined to have voted with multiple accounts will be banned and their votes will be disallowed. The MMP ballot committee has authority to exclude any ballot that does not meet these requirements.
Scoring: Points will be given in descending order with the highest-ranked player receiving 15 points, the second highest 14 points, and so on until the last player on the ballot receives 1 point. The player with the highest point total will be named the Most Meritorious Player. In case of a tie, the tiebreaker will be number of 1st place votes. If the first tiebreaker does not determine a winner the players will share the title of Most Meritorious Player.

   54. Kiko Sakata Posted: January 17, 2018 at 02:22 PM (#5608206)
Tulo's offense + defense + Age accounted for his RPA rating. His offense was evaluated at .106. His efense was a +6 and the age adjustment a +1 for a total RPA of .113.

The median SS, among the top 30, for the coming season is .128. That would make Tulo a -15 at SS. What would that be in LF, even if, doubtfully, he was able to maintain the +6 in LF?

Tulo would be a -25 because the median of the top 30 in LF is .138.


I've only been half paying attention to this conversation, because I haven't actually voted in an MMP election in a while (maybe posting this will prompt me to change that). So, apologies if this has been explained before.

Hypothetically, suppose you were considering where to place Troy Tulowitzki on an MMP ballot (obviously, the current version wouldn't make it, seeing as he's below average - and fair enough, he was below average in 2017 in my system too). You say he has an RPA of .113 which is -15 at SS. Now, suppose there's also a LF version of Tulo who was also .113 which, you say, is -25 at LF.

In your MMP rankings, would these two Tulo's be tied at .113 or would SS Tulo rank ahead of LF Tulo -15 to -25?

Thanks! And again, sorry if I'm asking a question that you've already answered.
   55. 6 - 4 - 3 Posted: January 17, 2018 at 07:08 PM (#5608428)
I'm still not understanding the rationale for the age adjustment component in the context of a MMP discussion. If two players put up 8 win seasons (or substitute whatever metric you choose--point is that they're equal), why care if one is 25 and one is 35? Equal on-field value is equal on-field value.
   56. Petunia inquires about ponies Posted: January 17, 2018 at 07:22 PM (#5608434)
You guys are tilting at the weirdest windmill here. A system that rates a full-time DH as more valuable than an identical hitter playing the 16th-best shortstop defense in the league is prima facie ridiculous and merits no attempt at rational engagement.
   57. Carl Goetz Posted: January 17, 2018 at 08:09 PM (#5608459)
"I'm still not understanding the rationale for the age adjustment component in the context of a MMP discussion. If two players put up 8 win seasons (or substitute whatever metric you choose--point is that they're equal), why care if one is 25 and one is 35? Equal on-field value is equal on-field value."
Pretty sure he's using a projection system rating the previous 2 years with an age adjustment. This explains to me how he rated Uggla ahead of Cano for 2010. It makes no sense otherwise. It still makes no sense to vote on a season that already happened using a projection system......

"You guys are tilting at the weirdest windmill here. A system that rates a full-time DH as more valuable than an identical hitter playing the 16th-best shortstop defense in the league is prima facie ridiculous and merits no attempt at rational engagement."
Agreed, but up until now, his vote has been counting for MMP elections and at 10%, that's a major skew to the results.
   58. toratoratora Posted: January 17, 2018 at 09:39 PM (#5608520)
2011 Prelim.
The usual blend of WAR systems.
No adjustments. No Post-season.
Comments will come with final ballot

1-Jose Bautista
2-Matt Kemp
3-Jacoby Ellsbury
4-Justin Verlander
5-Roy Halladay
6-Cliff Lee
7-Ryan Braun
8-Clayton Kershaw
9-Miguel Cabrera
10-Ben Zobrist
11-Dustin Pedroia
12-Joey Votto
13-Adrian Gonzalez
14-Jered Weaver
15-Prince Fielder
15 (Tied)-Brian McCann

The best of the rest, in order:
Ian Kinsler, Robinson Cano, Alex Gordon, Justin Upton, Curtis Granderson, Evan Longoria, Lance Berkman


   59. Kiko Sakata Posted: January 18, 2018 at 12:44 AM (#5608567)
Okay, having been drawn to this thread (and the 2010 Results thread) by the somewhat weird conversation with caiman re: his RPA (caiman: could you please answer my #54 if you have a chance? Thanks!), I'm going to go ahead and do one of these.

I have my own statistic: Player won-lost records

I calculate them two ways: pWins tie to team wins, eWins neutralize the context. I compare against positional average, replacement level (which is one standard deviation below positional average), and "star" level (which is one standard deviation above positional average).

I set up a page that allows one to calculate an "uber-stat" based on various weights. For regular Hall-of-Merit voting, I do a (2/3), (1/3) split between eWins and pWins. For a season, I thought maybe a .5/.5 split made more sense to recognize the context of what actually happened in 2011. I also usually weight all of wins, WOPA, WORL, and WO*, but I decided to drop raw wins here. Differences in total wins aren't necessarily due to "skill" (or even playing time) but may be due to position. I may do something similar for the 2019 Hall of Merit ballot; we'll see.

Anyway, this is what I came up with.

Here's the top 31 (I'll explain why 31 below):

1. Joey Bautista
2. Clayton Kershaw
3. Ryan Braun
4. Roy Halladay
5. Justin Verlander
6. Cliff Lee
7. Evan Longoria
8. Jered Weaver
9. Mike Napoli
10. Matt Kemp
11. Miguel Cabrera
12. Dan Haren
13. Curtis Granderson
14. Alex Avila
15. Josh Hamilton
16. Jacoby Ellsbury
17. Lance Berkman
18. Adrian Beltre
19. Matt Cain
20. Cole Hamels
21. Prince Fielder
22. Doug Fister
23. Dustin Pedroia
24. Ian Kennedy
25. Ian Kinsler
26. Albert Pujols
27. Robinson Cano
28. James Shields
29. C.C. Sabathia
30. Pablo Sandoval
31. Troy Tulowitzki
...
41. Ben Zobrist

Two things strike me as comment-worthy. First, positional balance. You have to get down to #23 before you find a middle infielder and #31 before you get to a shortstop. I don't think there's anything necessarily wrong with that: there are 3 second basemen between #23 and #27; it's just that none of them necessarily separated himself from the others.

The other noteworthy thing is Ben Zobrist at #41 as opposed to his 0.5-win lead in bWAR. As I'm sure you can all guess, the issue here is fielding. Baseball-Reference shows Zobrist with an oWAR of 5.7 - that's what his WAR would be if he was an average fielder, which would place in the 25-30 range in bWAR. So, how good a fielder was Ben Zobrist in 2011?

Baseball-Reference's DRS - which they use in calculating bWAR - says he was +29 - i.e., 3 wins.
But Baseball-Reference also shows an alternate fielding statistic, Total Zone (Rtot). According to that, Zobrist was -4.

Fangraphs uses UZR, which shows Zobrist as +10.1 runs. Interestingly, DRS and UZR basically agree on Zobrist's career value, +60, but DRS says he earned his +60 in three years (2009-11) and has been dead average over the rest of his career, while UZR gets to +60 by basically showing Zobrist +10 for a 6-year stretch from 2009-14.

Baseball-Gauge shows numbers using DRA, Michael Humphreys' stat. From that, Zobrist appears to be +5.6 runs if I'm reading their tables correctly (I want "single-year" not "multi-year", right?).

And what do I say? My Player won-lost records say that Zobrist had a fielding record of 5.4 - 5.1, so, about 0.3 net wins. My numbers show that Zobrist's fielding rate (winning percentage) was almost precisely his career rate at both 2B and RF (slightly above average in both cases).

If EVERYBODY ELSE thought that Ben Zobrist was worth 2 wins in the field, I might be inclined to give him a judgmental boost. But my numbers are basically right in line with DRA, broadly consistent with UZR, and MORE favorable to Zobrist than Total Zone. So, I'm comfortable with Zobrist ending up off-ballot.
   60. caiman Posted: January 18, 2018 at 01:48 AM (#5608571)
6-4-3: I do not use the age adjustment in MMP. MMP is for past performance in that season ONLY. I am amazed that there is confusion about what I said. The two-year averaging and the age adjustment is ONLY for projecting the likely performance for the player for the upcoming season.
   61. caiman Posted: January 18, 2018 at 02:12 AM (#5608572)
Kiko:

Thanks for asking the question!

Tulo's .113 was part of my projection for the upcoming season, based on two years of performance, not his performance last season alone.

Tulo's offense in 2016 was .115 and last season was horrible at .093. Simply terrible.

I do not use the RPA projections for MMP voting. I use the actual run value of that production in my MMP voting.

In Tulo's case, he cost the BlueJays 8.21 runs on offense while saving them 2.60 runs on defense for a total MMP value of -5.61 runs.

Contrary to what others said, I do not believe anyone can claim to know what a player's value would be at another position, other than it's not likely he would be as good as what he produced at his 'natural' position.

In other words, the only 'guarantee' is that Tulo's .106 offensive RPA would remain the same in LF and his age adjustment would remain +1, for a total RPA of .107.

Tulo wold need to be a +31 to even be rated at zero in LF, something almost impossible to do and he's not even likely to be a +6, as he was at SS. He would likely end up being even lower than .113 in LF, which would leave him at least a -25 in LF.

As for Gordon playing SS, I doubt very much that he'd be a terrible SS or even a bad SS. In order to be a very good OF, and he's one of the best, you have to be a very good athlete with very good reaction time and quickness, all of which would play well at SS. That's the most I can say about Gordon, without seeing him play SS. He could be bad, he could be good. My guess is that he is more likely to be good than bad. The problem for Gordon, these days, as for Tulo, is that his bat has disappeared!
   62. Carl Goetz Posted: January 18, 2018 at 10:32 AM (#5608652)
Responding generally to Kiko's #59 Zobrist comments.
Typically DRS is my defensive stat of choice in the years its available, though I'm still considering regressing it a bit based on our earlier conversation. That said, its hard to ignore DRA, TZ, UZR, and your own Player Wins all saying he was worth between 2 & 3.5 wins less than DRS. That's a major discrepancy. In fact the entire reason I haven't posted a preliminary ballot is that I'm still figuring out what to do with Zobrist's defense.

"Two things strike me as comment-worthy. First, positional balance. You have to get down to #23 before you find a middle infielder and #31 before you get to a shortstop. I don't think there's anything necessarily wrong with that: there are 3 second basemen between #23 and #27; it's just that none of them necessarily separated himself from the others."

Not sure if this was directed at me and my criticism of caiman's ballot, but I wanted to comment to clarify my position even if that wasn't your intent. While I do feel there are 2B deserving of a top#15 ranking (or at least consideration), I agree with your central point (I think your central point) that not every position needs to be on a ballot every year. Some years, certain positions just aren't top heavy and SS in 2011 certainly qualifies. My central point is that caiman's ballots look like this every year. The only times I found a 2B-3B-SS in a top-15 slot on his ballot were in some of A-Rod & Chipper's monster offensive seasons (I believe Donaldson made it once too) and even then, they were way down in the back half of the top 15. This is due to his "alternative" defensive spectrum.

   63. Kiko Sakata Posted: January 18, 2018 at 10:50 AM (#5608671)
Not sure if this was directed at me and my criticism of caiman's ballot, but I wanted to comment to clarify my position even if that wasn't your intent.


Your comment to caiman probably pushed this forward in my mind, but, as a general rule, I like to see balanced ballots. As you say, some years, for whatever reason, there just aren't any middle infielders worth a vote. I just want to make sure that I'm not doing something systematic - as it appears caiman may be doing - that would lead to unbalanced ballots every year. I want to fiddle around a bit with this. Just as a quick check, using the same weights for 2010, Robinson Cano popped up at #3 with Tulowitzki at #4 (behind Josh Hamilton and Joey Votto). In 2012, Robinson Cano and Buster Posey are in a virtual tie for #1. So, I think what I'm doing is okay, and I'm just going to end up with no middle infielders on my 2011 ballot. But I'll probably run a few more seasons to see if maybe I need to give a bit of a boost to middle infielders (I was actually a little surprised that Cano wasn't #1 in 2010). It could be that the weighting scheme I use for looking at careers in general HOM voting doesn't translate as well to single seasons (or that the weighting scheme that I'm using in general sucks).
   64. Carl Goetz Posted: January 18, 2018 at 11:03 AM (#5608693)
"6-4-3: I do not use the age adjustment in MMP. MMP is for past performance in that season ONLY. I am amazed that there is confusion about what I said. The two-year averaging and the age adjustment is ONLY for projecting the likely performance for the player for the upcoming season."

If anyone can read caiman's series of vagueries and contradictions and not be confused, please let me know. In one post, he says he doesn't use defense prior to 2006 because he doesn't have the information available. In another, he says he's been using defense in his ratings for more than 2 decades. In one post he rates on the 2 prior years plus an age adjustment; here he says he doesn't.
For me, it comes back to Uggla rating 15 runs better than Cano in 2010. With both playing 2B and with 2B being the lowest rung on his defensive spectrum, I doubt defense comes into play much in his comparison between the 2. oWAR has Cano leading 6.4-5.3. It would take a pretty large park factor difference for Uggla to move him from 1 win down offensively to 1.5 wins up.
But, if we take 2008 & 2009 plus an age adjustment, suddenly the ranking (offensively anyway) appears much more reasonable due to Cano's poor 2008 at the plate.
2008-09 oWAR
Uggla: 4.0 + 3.3 = 7.3
Cano: 1.3 + 4.5 = 5.8

About 1.5 wins (15ish Runs) better total. I assume a rational age adjustment would work slightly to Cano's benefit since 2010 was his age 27 season (Age 30 for Uggla), but at least we're in the ballpark.

I certainly can't claim to decipher every contradiction that comes from caiman, but in this case, the evidence would indicate that its more likely that his statement that he uses the 2 prior years (2008 & 2009 in this case) plus an age adjustment to rate the current year (2010) is the true one. If someone has an alternative explanation, I'd love to hear it. But, 2.5 win park adjustment just doesn't make sense to me.
   65. Kiko Sakata Posted: January 18, 2018 at 11:52 AM (#5608771)
For me, it comes back to Uggla rating 15 runs better than Cano in 2010. With both playing 2B and with 2B being the lowest rung on his defensive spectrum, I doubt defense comes into play much in his comparison between the 2. oWAR has Cano leading 6.4-5.3. It would take a pretty large park factor difference for Uggla to move him from 1 win down offensively to 1.5 wins up.


Actually, if I look at context-neutral batting only for 2010, relative to average, I find that Uggla beats Cano 1.6 - 1.3 wins. Now, that's only about 3 runs, not 15 runs. And Cano gains back one of those runs with baserunning and Cano beats Uggla on fielding by twice the batting difference. And if you put their records in context, Cano jumps dramatically ahead (Cano had 4.7 pWORL vs. 3.5 eWORL; Uggla's pWORL/eWORL differed by only 0.3). So, as I said earlier, Cano would have been top 5 on my hypothetical 2010 ballot and Uggla would have been well off it. But if I squint just so, I can see the beginning of an argument that Uggla was better.

One other possible issue. Uggla played in the NL, where, of course, pitchers bat, while Cano played in the DH league. To control for this, my comparisons (in the link above) are to what I call WOPA_b, which is the league average batting performance for non-pitchers - which is .500 by construction in the AL, but will be something over .500 in the NL. If you just compare to everybody in a league, Uggla will be more above average than Cano in part because he's being compared to a lower average - that includes pitchers - than Cano - whose average includes, for example, David Ortiz and Jim Thome.
   66. Carl Goetz Posted: January 18, 2018 at 01:13 PM (#5608851)
"Actually, if I look at context-neutral batting only for 2010, relative to average, I find that Uggla beats Cano 1.6 - 1.3 wins. Now, that's only about 3 runs, not 15 runs. And Cano gains back one of those runs with baserunning and Cano beats Uggla on fielding by twice the batting difference."

Ok, so 2 runs after we net in the baserunning runs. That still leaves 13 runs on offense for differences in park factor to bridge. That's still a big gap for park factors to explain. I don't think using context dependent numbers bridges the gap since Cano wins on WPA 4.07-2.71 which I believe is only offense in context.

"One other possible issue. Uggla played in the NL, where, of course, pitchers bat, while Cano played in the DH league."
He does claim to be comparing to the median 2B so not sure this one works.
   67. DL from MN Posted: January 18, 2018 at 02:16 PM (#5608912)
The more voters we have, the better.
   68. Carl Goetz Posted: January 18, 2018 at 02:39 PM (#5608932)
Fair enough.
   69. caiman Posted: January 18, 2018 at 02:53 PM (#5608942)
Again: I have used a defensive rating for well more than two decades. However, the MMP is different. It requires me to work out the actual run value of that defense, for each and every player in past seasons. As I have stated, it would require me to go back over every single player, in every single past year, in order to determine that value. MMP, therefore, has nothing to do with my projective defensive evaluations of players for upcoming seasons.

Why does this separation exist? It is because I only began my historical ratings within the last decade and the defensive modification to that historical ratings only within the last three years. The historical ratings require me to proceed on an across-the-board MLB study of those seasons. Even though I have rated players defensively in the past, my new retroactive work is being done differently from the projective ratings. Those prior RPA rating were always projective, in order to evaluate upcoming seasons. While I have historical ratings back to 1900, the defensive evaluation only goes back to 2009, simply because it would require me to run the Retrosheet play-by-play data, followed by player-by-player measurements of the play-by-play results, for each year. That a lot of work and I do not have that time available. I have a life. I am involved in many areas of life, from politics, to physics, to unions, to economics, etc...
   70. Kiko Sakata Posted: January 18, 2018 at 02:59 PM (#5608947)
Again: I have used a defensive rating for well more than two decades. However, the MMP is different. It requires me to work out the actual run value of that defense, for each and every player in past seasons. As I have stated, it would require me to go back over every single player, in every single past year, in order to determine that value. MMP, therefore, has nothing to do with my projective defensive evaluations of players for upcoming seasons.


I'm a bit confused. In comment #9, you cite a defensive rating for Jose Bautista (who is also #1 on my ballot), "-2.17 runs on defense".
   71. caiman Posted: January 18, 2018 at 03:14 PM (#5608958)
Kiko: I have done 2011. That's why I have a defensive run value for Bautista. As I said, I have done that historical study, that includes defense, back to 2008. I mis-spoke, above, when I said 2009. I've done it back to 2008 from and including the recent 2017 season.
   72. Kiko Sakata Posted: January 18, 2018 at 03:16 PM (#5608962)
Okay, thanks.
   73. Eskimo3842 Posted: January 19, 2018 at 09:29 AM (#5609371)
I find it odd that dialectical materialism would be used in this manner - Marx used it to attempt to describe social constructs, rather than ideal constructs. I would think that identifying a mid point as the "average" is by nature an ideal construct, as it assumes that all participants in the system behave accordingly. Whether to their benefit or detriment, it is apparent that teams do not behave this way, *shrug* I think I am with many in trying to figure out caiman's system and not seeing it, at least not yet. I do think shouting down other systems and statistics borders on appeal to ignorance, and supposed failures of other systems ought not be used to demonstrate the viability of an alternative, which feels like what some of caiman's defense has been, though there are also interesting points he raises at times.

All of that being said, discussion is always good, and after reading it for quite a while, I figured I'd chime in.

I find this a challenging year to evaluate. I'm tinkering, but for starters, I have this as an order:
1 Cliff Lee
2 Roy Halladay
3 Justin Verlander
4 Jacoby Ellsbury
5 Ben Zobrist
6 Matt Kemp
7 Dustin Pedroia
8 Jose Bautista
9 Ryan Braun
10 Miguel Cabrera
11 Clayton Kershaw
12 CC Sabthia
13 Ian Kinsler
14 Joey Votto
15 Evan Longoria

I tend to start with WAR, though I do favor offensive contribution over defense a bit more. Also credit for postseason. Uncomfortable with this much pitching at the top. Lee's hitting, as others have said, right now giving the edge.
   74. caiman Posted: January 19, 2018 at 10:14 AM (#5609397)
Eskimo3842: Dialectical materialism is a method of analysis of matter in motion, i.e., of change. Marx' analysis of social structure is a product of that method in one area of science, the most important for the question of class struggle and revolution, however it is not limited to that. It applies in all sciences. The two primary works of dialectical materialism are Engels "Anti-Duhring" and Lenin's "Materialism and Empirio-Criticism".

Baseball players are 'matter-in-motion'. Classical science tends to look at things isolated from that motion. Dialectical Materialism looks at things from the context of their motion, in relation to the motion of related matter.

The midpoint is the median, not the average player, when I do the position-by-position projective RPA's.



   75. caiman Posted: January 19, 2018 at 10:34 AM (#5609408)
Just for fun, here's the worst players in 2017:

1. Derek Holland -27.39 runs (Note: 5 runs = 1 win. Therefore Holland cost his team 5.478 wins)
2. Alcides Escobar -24.19 runs
3. Tommy Joseph -23.26 runs
4. Albert Pujols -23.24 runs
5. Austin Romine -21.74 runs
6. Ubaldo Jimenez -21.43 runs
7. Matt Moore -20.95 runs
8. Matt Wieters -20.58 runs
9. Maikel Franco -20.47 runs
10. Kevan Smith -20.36 runs
11. Matt Harvey -20.12 runs
12. Chris Tillman -19.68 runs
13. Mike Pelfry -19.44 runs
14. Rougned Odor -17.50 runs
15. James Shields -17.32 runs
   76. Carl Goetz Posted: January 19, 2018 at 12:39 PM (#5609497)
I've decided to apply a 25% regression to BBRef dWAR based on my previous conversations with Kiko.
1)Cliff Lee(9.1) - Great year but it still takes offense contribution to push him ahead of Halladay.
2)Roy Halladay(8.7) - Best Pitcher of 2011 (only pitching)
3)Matt Kemp(8.5) - Defensive regression helps him. Just a fantastic year. He should have beaten Braun in MVP vote.
4)Justin Verlander(8.4) - Big year for pitchers.
5)Jose Bautista (8.2) - Another guy helped by defensive regression.
6)Jacoby Ellsbury(7.7) - One of his best years.
7)Ryan Braun (7.6) - Big year that was fun to watch in Milwaukee; even if chemically fueled.
8)Miguel Cabrera(7.7) - Big stick and easily the best 1B.
9)Dustin Pedroia (6.8) - Virtually tied with Zobrist, but gave Pedroia the nod due to more of his value being on offense.
10)Ben Zobrist(6.8) - Despite this downgrade, still a great year.
11)CC Sabathia(7.4) - The big man was closing in on the end of his prime, but still put up a big season here.
12)Clayton Kershaw(7.2) - I prefer not to be this pitcher heavy, but seems to be a top-heavy season for pitchers.
13)Evan Longoria (6.0) - Pretty confident he's the best 3B.
14)Adrian Gonzalez(6.1) - Little better than Votto for 2nd best 1B.
15)Curtis Granderson(6.1)

Others considered but not ranked: Jerad Weaver, Cole Hamels, Joey Votto, Ian Kinsler, Robinson Cano, Jose Reyes, Troy Tulowitzki, Pablo Sandoval, Adrian Beltre, Alex Gordon
I have David Ortiz as best DH and David Robertson as best reliever but neither was a serious consideration for top 15 ranking.
   77. Eskimo3842 Posted: January 19, 2018 at 02:10 PM (#5609620)
caiman: I'm pretty well versed in dialectical thinking, including dialectical materialism - I get what you are saying, but whether it is average or median, you are dealing primarily with an ideal construct (something that is a priori)- looks more like Hegelian dialectic from my vantage point. IN any event - I find it an interesting approach. It doesn't gel with what I "see" or "intuit", but that makes it more interesting to probe and test its validity. It seems like the debate (in part) is whether the system being analyzed is limited to the starters at each position, or expanded to include consideration of those who are in the system but do not start (or even play), thus the debate about replacement level player vs. average or median player at a position? In other words, the median, synthetic third proposition is necessarily a function of where you set the poles.

Carl: is that conversation with Kiko something you can point to on the boards? Your updated list looks a lot like what I was trying to do with mine, but probably a bit further along in figuring out how to scale the defensive contributions.
   78. Carl Goetz Posted: January 19, 2018 at 02:21 PM (#5609634)
Eskimo3842: I think its in the 2018 Hall of Merit discussion, but I don't recall for certain.
   79. Kiko Sakata Posted: January 19, 2018 at 02:23 PM (#5609637)
Carl: is that conversation with Kiko something you can point to on the boards?


The conversation starts at comment #352 here (that's the most recent Hall of Merit Discussion thread).
   80. Carl Goetz Posted: January 19, 2018 at 02:50 PM (#5609677)
Thanks Kiko.
   81. caiman Posted: January 19, 2018 at 06:36 PM (#5609865)
Eskimo3842: The results that I gave are not limited to the starters at that position. The results are for every player.

My top 3o include the top 30 of all the players rated. That is why Matt Szczur is rated #12 as a premier starter in LF, even though he's a sub.

The top 30 is in order to determine who are the best qualified to be starting players, simply because there are only 30 MLB teams and there can only be 30 starting positions available.

After determining the top 30, only the top 15 can give a team a plus against the bottom 15 of qualified starters in this starter vs. starter comparison.

This is not a determining factor for PAST PERFORMANCE. It is only used for projection for the coming season, in order for me to evaluate the teams.

Based on that, Last season I projected that the Dodgers and the Indians were the strongest teams in each league.

The process is this: I add up the differences, at each position + the 5 starters and three main relievers, between the actual starter on that team vs. the median rated of the top 30. The total leaves me with a value number for each team.

When I used this method to advise Dan Duquette, decades ago, we were able to pick off many players from other teams that did not realize the value of their own player -and/or - trade away a player that others thought highly of, but my ratings said otherwise.
   82. Eskimo3842 Posted: January 20, 2018 at 11:42 AM (#5610024)
@caiman: right - but the median point being at fifteen is an ideal construct, related to the idea that teams/players as "matter in motion" will behave relative to poles that are the best team and worst team possible to field given the 30 best players at each position? Even if the predictive value is what you propose, does it accurately evaluate the "past" being the actual performance that occurred, (which is what he MMP voting seeks to do, with that median as the zero value as opposed to a lower replacement value? Not being willfully dense or anything, just trying to put my finger on the exact moment of discrepancy between your system and others (notably the varieties of WAR). I know that this ground has been otherwise covered, but the conversations I have followed devolved at various points, and it seems to me this is the crux of disagreement where they devolved - what the zero value player ought to be. Given that dialectical materialism is your stated philosophical underpinning, I am trying to determine whether the zero value is derived from actual behaviors or a chosen "ideal". If it is the latter, there is more room for debate.

@kiko&Carl; - thanks - I'm heading over to read up on that conversation.
   83. caiman Posted: January 20, 2018 at 03:15 PM (#5610118)
The RPA rating is the basis for all my calculations. The RPA takes in all players. RPA rates all players using the same formula that starts with adjusting all values to a 'standard year'. The RPA uses .115 as the standard RPA to which all performances, regardless of the year, are normalized to.

In other words, if the actual RPA for all scoring, in a particular year is, say, .121, the RPA normalizes the value of all performances that season to .115 in order to make all seasons equivalent.

When each stadium is evaluated for its offense, in that year, it's scoring is compared to that year's RPA, for the stadium variant.

Each player or pitcher's RPA value is gotten via the stadium variant, within the adjusted year adjustment to the standard of .115

In terms of historical ratings, the RPA is used to determine how many runs, plus or minus, the player provided. It can also be used to rank players by field position or league-wide, either by RPA value or by run production value. One player can have a much higher RPA than the other player who has produced more runs, simply because of more or less plate appearances or batters faced.

as an example, Mike Trout had the highest RPA of any player in 2017 with a .228 RPA. But he produced fewer runs than Aaron Judge. Trout produced 54.85 runs and Judge produced 59.14 runs, even though Judge's RPA was lower than Trout's at .200. That is because of Trout's time on the DL. Judge, produced at a lower RPA rate, but accounted for a greater value, due to the aggregate value of his performance.

In terms of projecting performance, the RPA uses all the data from all player performances and ratings, and places each player, performing at each position, in numerical RPA order of the results.

There are, 49 MLB third basemen, with at least 200 computed MLB plate appearances, over the last two seasons, listed number one to 49 on my rPA ratings, from best to worst.

Since only 30 can be starting third basemen, because there are only 30 positions that can be filled, I find it most useful in evaluating teams, to compare the top 15 to the 16-30 qualified starters, per the RPA formula.

Am I completely ignoring the 19 bench warmers? I suppose so, and that is a choice I made for simplifying my team analysis.

Frankly, I am more concerned about young prospects who might become the actual starters, but who have little or no MLB experience, but could have a huge affect on team projections. Generally speaking, the bench warmers can likely only hurt the team they are on and, therefore, I feel free to ignore them, unless they suddenly show an upsurge in their abilities.

The use of the top 15 vs. the bottom 15 is simply for utility. All I want to know is how much better or worse is one starter at that position vs. the other starter at that position, in evaluating team prospects for the upcoming season.

The RPA formula itself HAS NOTHING TO DO WITH THIS practical use of the RPA for the projection of performance. The RPA is a formula developed by dialectics. The use of the top 15 is just how I use the results, for practical purposes, AFTER determining a player's RPA value.

   84. DL from MN Posted: January 22, 2018 at 12:28 PM (#5610899)
NLDS Cardinals 3 Phillies 2
Player Name  G  AB  R  H  2B  3B  HR  RBI  BB  SO  BA  OBP  SLG  OPS  SB  CS  E
Albert Pujols 5 20 2 7 3 0 0 1 2 4 .350 .409 .500 .909 1 1 0

Hunter Pence 5 19 3 4 0 0 0 4 2 2 .211 .286 .211 .496 0 1 0
Shane Victorino 5 19 2 6 1 0 0 2 0 0 .316 .316 .368 .684 0 0 1

Pitcher Name  G  GS  ERA  W  L  SV  CG  IP  H  R  ER  BB  SO  WHIP
Roy Halladay 2 2 2.25 1 1 0 0 16.0 9 4 4 2 15 0.688
Cole Hamels 1 1 0.00 1 0 0 0 6.0 5 0 0 3 8 1.333
Cliff Lee 1 1 7.50 0 1 0 0 6.0 12 5 5 2 9 2.333 
   85. DL from MN Posted: January 22, 2018 at 12:49 PM (#5610924)
NLDS Brewers 3 Diamondbacks 2
Player Name  G  AB  R  H  2B  3B  HR  RBI  BB  SO  BA  OBP  SLG  OPS  SB  CS  E
Ryan Braun 5 18 5 9 4 0 1 4 3 3 .500 .571 .889 1.460 1 0 0 
Prince Fielder 5 18 2 5 2 0 1 3 2 5 .278 .381 .556 .937 0 0 0

Miguel Montero 5 20 3 6 2 0 0 2 2 6 .300 .364 .400 .764 0 0 0
Justin Upton 5 20 3 4 0 0 2 3 3 6 .200 .304 .500 .804 0 0 0
Chris Young 5 18 5 7 1 0 3 4 2 5 .389 .450 .944 1.394 2 0 0

Pitcher Name  G  GS  ERA  W  L  SV  CG  IP  H  R  ER  BB  SO  WHIP
Ian Kennedy 2 2 4.26 0 1 0 0 12.2 13 6 6 3 8 1.263 
   86. DL from MN Posted: January 22, 2018 at 12:52 PM (#5610926)
ALDS Tigers 3 Yankees 2
Player Name  G  AB  R  H  2B  3B  HR  RBI  BB  SO  BA  OBP  SLG  OPS  SB  CS  E
Alex Avila 5 16 0 1 0 0 0 1 2 7 .063 .167 .063 .229 0 0 0 
Miguel Cabrera 5 15 2 3 0 0 1 3 5 5 .200 .400 .400 .800 1 0 0
Austin Jackson 5 16 3 2 1 0 0 0 4 8 .125 .300 .188 .488 1 0 0

Robinson Cano 5 22 2 7 2 0 2 9 2 4 .318 .375 .682 1.057 0 0 0
Brett Gardner 5 17 3 7 1 0 0 5 1 4 .412 .444 .471 .915 0 0 0 
Curtis Granderson 5 20 4 5 1 1 1 3 4 7 .250 .375 .550 .925 0 0 0

Pitcher Name  G  GS  ERA  W  L  SV  CG  IP  H  R  ER  BB  SO  WHIP
Doug Fister 2 1 6.52 1 1 0 0 9.2 12 7 7 4 10 1.655 
Justin Verlander 2 2 5.00 1 0 0 0 9.0 6 5 5 5 12 1.222

Mariano Rivera 2 0 0.00 0 0 0 0 1.1 0 0 0 0 1 0.000
David Robertson 2 0 0.00 0 0 0 0 2.0 0 0 0 0 2 0.000
CC Sabathia 3 2 6.23 0 0 0 0 8.2 10 6 6 8 11 2.077
   87. DL from MN Posted: January 22, 2018 at 12:59 PM (#5610938)
ALDS Rangers 3 Rays 1
Player Name  G  AB  R  H  2B  3B  HR  RBI  BB  SO  BA  OBP  SLG  OPS  SB  CS  E
Elvis Andrus 4 14 1 2 0 0 0 0 3 3 .143 .333 .143 .476 0 1 0
Adrian Beltre 4 15 5 4 0 0 3 4 0 1 .267 .313 .867 1.179 0 0 1
Ian Kinsler 4 16 2 4 2 0 1 3 3 3 .250 .368 .563 .931 1 0 0
Mike Napoli 4 14 3 5 0 0 1 4 1 5 .357 .400 .571 .971 1 0 0

Evan Longoria 4 16 2 3 0 0 1 3 2 8 .188 .278 .375 .653 0 0 0
Ben Zobrist 4 17 2 4 0 0 0 0 0 4 .235 .278 .235 .513 0 0 0

Pitcher Name  G  GS  ERA  W  L  SV  CG  IP  H  R  ER  BB  SO  WHIP
Matt Harrison 2 1 4.76 1 0 0 0 5.2 6 3 3 3 9 1.588 
C.J. Wilson 1 1 10.80 0 1 0 0 5.0 7 8 6 1 6 1.600

James Shields 1 1 12.60 0 1 0 0 5.0 8 7 7 0 6 1.600
   88. DL from MN Posted: January 22, 2018 at 01:04 PM (#5610945)
NLCS Cardinals 4 Brewers 2
Player Name  G  AB  R  H  2B  3B  HR  RBI  BB  SO  BA  OBP  SLG  OPS  SB  CS  E
Albert Pujols 6 23 5 11 4 0 2 9 4 4 .478 .556 .913 1.469 0 0 0

Ryan Braun 6 24 2 8 3 0 1 6 1 6 .333 .385 .583 .968 0 0 0
Prince Fielder 6 20 4 4 2 0 2 3 4 4 .200 .360 .600 .960 0 0 0
   89. DL from MN Posted: January 22, 2018 at 01:10 PM (#5610958)
ALCS Rangers 4 Tigers 2
Player Name  G  AB  R  H  2B  3B  HR  RBI  BB  SO  BA  OBP  SLG  OPS  SB  CS  E
Elvis Andrus 6 25 4 6 0 0 0 1 3 2 .240 .321 .240 .561 1 1 1
Adrian Beltre 6 27 4 6 3 0 0 2 2 10 .222 .276 .333 .609 0 0 0
Ian Kinsler 6 24 6 7 2 0 0 6 4 2 .292 .414 .375 .789 1 1 0
Mike Napoli 6 24 6 7 0 0 0 1 3 4 .292 .370 .292 .662 0 0 0

Alex Avila 6 25 1 2 0 0 1 1 0 9 .080 .080 .200 .280 0 0 0
Miguel Cabrera 6 20 5 8 4 0 3 7 7 6 .400 .556 1.050 1.606 1 0 0
Austin Jackson 6 25 3 6 1 0 1 4 3 11 .240 .345 .400 .745 0 1 1

Pitcher Name  G  GS  ERA  W  L  SV  CG  IP  H  R  ER  BB  SO  WHIP
Matt Harrison 1 1 3.60 0 0 0 0 5.0 3 2 2 3 3 1.200
C.J. Wilson 2 2 6.75 0 1 0 0 10.2 14 8 8 7 11 1.969

Doug Fister 1 1 2.45 1 0 0 0 7.1 7 2 2 0 3 0.955
Justin Verlander 2 2 5.56 1 1 0 0 11.1 13 7 7 5 13 1.588 
   90. DL from MN Posted: January 22, 2018 at 01:22 PM (#5610982)
World Series Cardinals 4 Rangers 3
Player Name  G  AB  R  H  2B  3B  HR  RBI  BB  SO  BA  OBP  SLG  OPS  SB  CS  E
Albert Pujols 7 25 8 6 1 0 3 6 6 3 .240 .424 .640 1.064 0 0 2

Elvis Andrus 7 29 5 8 1 0 0 0 1 7 .276 .300 .310 .610 0 0 1
Adrian Beltre 7 30 5 9 2 0 2 3 0 8 .300 .323 .567 .889 0 0 0
Ian Kinsler 7 25 2 9 1 0 0 2 7 4 .360 .500 .400 .900 1 3 2
Mike Napoli 7 20 2 7 1 0 2 10 6 4 .350 .464 .700 1.164 0 0 1

Pitcher Name  G  GS  ERA  W  L  SV  CG  IP  H  R  ER  BB  SO  WHIP
Matt Harrison 2 2 7.04 0 2 0 0 7.2 11 8 6 3 4 1.826 
C.J. Wilson 3 2 2.92 0 1 0 0 12.1 8 5 4 11 9 1.541

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Dynasty League Baseball

Support BBTF

donate

Thanks to
Chicago Joe
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Hot Topics

2019 Hall of Merit Ballot Discussion
(106 - 6:31pm, Apr 21)
Last: Kiko Sakata

Most Meritorious Player: 1942 Ballot
(1 - 11:54am, Apr 20)
Last: DL from MN

Most Meritorious Player: 1942 Discussion
(10 - 9:55am, Apr 19)
Last: DL from MN

Latin American Players and Pitchers Home Page
(68 - 8:01pm, Apr 18)
Last: Dr. Chaleeko

Carlos Moran
(70 - 6:27pm, Apr 11)
Last: Dr. Chaleeko

Oliver Marcelle
(12 - 6:24pm, Apr 11)
Last: Dr. Chaleeko

Connie Marrero
(10 - 1:40pm, Apr 06)
Last: Dr. Chaleeko

Leroy Matlock
(68 - 1:39pm, Apr 06)
Last: Dr. Chaleeko

Max Manning
(4 - 1:38pm, Apr 06)
Last: Dr. Chaleeko

Marvin Williams
(30 - 1:36pm, Apr 06)
Last: Dr. Chaleeko

Most Meritorious Player: 1941 Results
(1 - 4:29pm, Apr 04)
Last: DL from MN

Most Meritorious Player: 1941 Ballot
(12 - 3:16pm, Apr 04)
Last: DL from MN

Most Meritorious Player: 1941 Discussion
(17 - 4:17pm, Mar 31)
Last: TomH

Jim Gilliam
(9 - 12:58pm, Mar 30)
Last: Dr. Chaleeko

George Scales
(71 - 12:55pm, Mar 30)
Last: Dr. Chaleeko

Syndicate

Page rendered in 0.7604 seconds
42 querie(s) executed