Jeff's Sports Q & A

2010-06-29T22:03:00.001-07:00

Last Sunday, Evan Longoria and B.J Upton got into an argument over B.J. taking his time chasing down a ball hit to the wall. This may seem like a small spat between teammates, but may have an unseen racial impact with people that saw it.

http://mlb.mlb.com/news/article.jsp?ymd=20100627&content_id=11667090&vkey=news_mlb&fext=.jsp&c_id=mlb

First we will need to take a step back and look at how the brain works. In the book “Blink” by Malcolm Gladwell, he states:

“First of all, we have our conscious attitudes. This is what we choose to believe. These are our stated values, which we use to direct our behavior deliberately.....our second level of attitude, our racial attitude on an unconscious level, the immediate, automatic associations that tumble out before we have time to think. We don't deliberately choose our unconscious attitudes”

Observers of the confrontation between Longoria and Upton probably didn't even notice the race of the players, but our subconscious mind is being told something else. It is see the black player (Upton) as lazy and the white player (Longoria) as the leader and putting the black player in his place.

The racial under tone is not limited to just the Longoria-Upton argument. I know there are plenty of examples that contradict the this example, but are those players making the headlines that we read everyday. Milton Bradley is a trouble maker. David Eckstein is a gritty veterans. Chris Brown is a malingerer. Pete Rose is Charlie Hustle

In most of our normal daily activities, the racist subconscious is a non factor, but as Malcolm Gladwell states:

“It's a powerful predictor of how we act in certain kinds of spontaneous situations.”

The people that have a bunch of spontaneous situations during a game that witness the behavior are the players, managers and umpires. The umpire could have seen the flare up in the Rays dugout on Sunday and he has to make a call on a close play and his mind has to make a decision between a white catcher and a black base runner.

I am not sure if there is a workable solution to this problem. As Galdwell states:

It [un-biased associations] requires that you change your life so that you are exposed to minorities on a regular basis and become comfortable with them and familiar with the best of their culture, so that when you want to meet, hire, date, or talk with a member of a minority, you aren't betrayed by your hesitation and discomfort

The media will not change what they report on, so what we see, hear and read will always be out of our control. I think the best way to fight this issue as fans is to understand the what is going on our subconscious mind and look in our personal lives for chances to train our brain differently.

Status of Blog

2009-07-01T08:24:00.000-07:00

I have not been able to keep up linking my posts and have taken on writing at Driveline Mechanics also. To see my current work, go either:

beyondtheboxscore.com

Drivelinemechanics.com

Update of Posts - 5/22

2009-05-22T08:46:00.000-07:00

I looked at players ont he KC A's and Royals at:

Link

Average and replacement level HOFers using WAR:

Link

Is Farnsworth's nibbling leading to better results:

Link

Outfield UZR Aging Patterns

Link

Recent entries to the 300 win Club

Link

Couple of recent articles at BtB since 5/18

2009-05-18T08:10:00.000-07:00

Members of the 500 Home Run Club Not In The Hall Of Fame

Post-Season Comparison - Is it Tougher to Win the Super Bowl or Make it to the Final Four?

Joined Beyond the Boxscore

2009-05-12T07:59:00.000-07:00

I have joined the group over at Beyond the Boxscore and plan on posting most of my new work there. If any article gets published, I will go ahead and link them here. For now I have two new pieces

A Look a Joakim Soria is the same Pitcher this compared to last.

The All Jeff Team

Analysis of the FSU college baseball team

Pitching Mound Construction Leading to Higher Scoring at Texas?

2009-04-27T08:22:00.000-07:00

Gil Meche, before his last start in Kansas City, was asked about his last start at Texas and here is his response:

"The reason I was stiff," Meche said, "is the mound (in Texas) was so flat. From the first pitch in warm-ups, I knew I was going to have to battle.

"With a flat mound, I don’t know if it’s the way my hips move or the way I stride out, but it affects something. That’s why I like steep mounds. I feel like I can generate more power instead of having my leg land too early and me having to generate power with my arm. That’s the only problem I had. That’s why I threw so many off-speed pitches."

This got me thinking about some work I did previously on individual park factors where I noted Rangers Ballpark in Arlington's official park factor was ~1.05, but it was predicted to be ~1.01. It has been stated that the stadium funnels the air to flow to center field was causing this discrepancy, but there seem to be another factor.

I decided to see if there was any truth to what Gil was talking about. For a initial look to see if there is anything to the accusation, I examined the average and maximum speeds of the opposing starting pitcher's four seem fastball that have pitched so far at Rangers Ballpark in Arlington in 2009. I limited the comparison games to those that happened as close possible to the start at Arlington, in case there was an injury to the pitcher. Also, I wanted to take the Texas pitcher's out of the study because they will be used to the supposedly smaller mound (I will look at Texas's pitchers as I get the time to see how they actually compare). There have been 9 pitchers so far that meet this criteria and here are their numbers:

			Game before Texas		Game at Texas		Game after Texas		Start before Texas		Start after Texas
Date	Team	Starting Pitcher	Average Speed	Max Speed	Average Speed	Max Speed	Average Speed	Max Speed	Difference in Average Speed	Difference in Max Speed	Difference in Average Speed	Difference in Max Speed
Mon, Apr 6	Cleveland	Lee	--	--	89.74	91.6	90.65	92.5	--	--	0.91	0.9
Wed, Apr 8	Cleveland	F Carmona	--	--	92.01	94.1	91.35	93.9	--	--	-0.66	-0.2
Thu, Apr 9	Cleveland	C Pavano	--	--	87.73	89	90.34	93.2	--	--	2.61	4.2
Mon, Apr 13	Baltimore	K Uehara	87.83	89.4	85.78	87.8	87.88	89.4	2.05	1.6	2.1	1.6
Tue, Apr 14	Baltimore	A Simon	92.97	95.3	89.93	91.8	--	--	3.04	3.5	--	--
Wed, Apr 15	Baltimore	M Hendrickson	87.16	88.7	86.24	88.7	86.54	88.1	0.92	0	0.3	-0.6
Fri, Apr 17	Kansas City	Meche	91.32	94.6	89.79	92.4	91.96	94	1.53	2.2	2.17	1.6
Sat, Apr 18	Kansas City	Greinke	92.93	95.6	92	94.7	94.17	97	0.93	0.9	2.17	2.3
Sun, Apr 19	Kansas City	Davies	92.12	94.2	90.62	92.3	90.9	92.9	1.5	1.9	0.28	0.6
		Average Values =	90.72	92.97	89.26	91.35	90.45	92.64	1.66	1.68	1.28	1.36

Just using this small sample size, it can be seen that there seem to be a little truth to ability of pitchers to throw their fastest stuff at Texas. Fausto Carmona is the only pitcher that has thrown faster in a start before or after going Texas compared to the start at Texas. In the other 14 cases the pitcher through slower in Texas by ~1.5 mph.

These numbers have peaked my interest enough that I will continue to update the preceding data as the season continues, expanding it to the Texas pitchers and look back at 2007 and 2008 data with pitch F/X to see if the smaller mound is linked to higher than expected scoring at Rangers Ballpark in Arlington.

Ron Mahay - Analysis Before and After his Trip to the DL

2009-04-27T08:10:00.000-07:00

In 2008 left hander Ron Mahay, just off signing a 2 year, 8 million dollar deal with the Royals, seemed to cruising through the season and helping the Royals solidify their bullpen. On August 15, he had an ERA of 2.91 and help make the pen at Kansas City one of the leagues strongest by joining Ramon Ramirez, Leo Nunez and Joakim Soria. On August 22, Ron was retroactively placed on the 15 DL and since that point, his results have not been the same. Ron was place on the disabled list for “plantar fasciitis in his left foot”. Plantar fasciitis is condition in the foot caused by excessive to the supports to the arches of your feet. Pain is felt in the heel of his plant foot.

After looking at Ron's numbers it seems like injury happened ~5 starts earlier. Here are Ron's stats for the 5 appearances before he went on the DL, the 11 appearances before those 5 and the 11 appearances he has had since coming off the DL.

Games Before DL Stint	ERA	FIP	WHIP	KK/9
6 to 16 before	0.73	2.07	0.97	9.44
5 before	31.3	21.03	4.35	7.83
11 after	12.68	6.86	2.96	11.41

Ron was rolling right along until his foot began to bother him an then his numbers just started going through the roof. After the surgery his numbers came down a little, but they still haven't been very good. The interesting stat is that he has been able to strikeout more batters since he came off the DL.

I wanted to see how the injury effects his pitch speed and movement. Looking at the same three time periods using Pitch F/X data, some some interesting results materialize on his 3 main pitches: Change up (CH), Slider (SL) and Fastball (FF)

Type of Pitch	Games Before DL Stint	Start Speed	End Speed	Break Angle	Break Length
CH	6 to 16 before	83.97	77.58	-29.21	6.64
CH	5 before	82.52	76.52	-28.16	7.02
CH	11 after	83.7	77.42	-27.33	7.04

SL	6 to 16 before	82.83	76.81	-3.74	7.47
SL	5 before	81.33	75.66	-1.4	8.04
SL	11 after	81.94	76.31	9.81	4.2

FF	6 to 16 before	90.58	83.1	-39.23	3.93
FF	5 before	90.62	83.27	-36.01	3.82
FF	11 after	90.07	82.57	-32.95	4.07

Lets look at the changes to each pitch type:

Change-up – Right before going on the DL, his change-up dropped by ~1.5 mph and had more of break (mainly do to more time to break before crossing the plate). Since the injury, he has not gotten all his speed back on his change to pre injury level

Slider – Like the change-up, his slider lost significant velocity before going on the DL and he has gained some, but not all of it back. More significantly, since going on the DL, his slider has just half the break while having a different break angle.

Fastball (4 finger) – It lost some velocity since he came off the DL, but it hasn't changed too much otherwise.

As a whole each pitch is just not as sharp as it was before he was hurt. Let's see what the results were for when each type of pitch was hit into play.

	Change ups			Fastballs			Sliders
Games Before DL Stint	Out	Safe	% outs	Out	Safe	% outs	Out	Safe	% outs
6 to 16 before	4	3	0.57	15	5	0.75	6	0	1
5 before	1	1	0.5	4	6	0.4	3	0	1
11 after	3	3	0.5	8	11	0.42	1	3	0.25

The loss in overall speed (and slider break distance) has led to more balls hit into play to be hits, especially for his fastball and slider. Though this is a small sample size, but he just not being very effective. In Ron's last two appearance, Trey Hillman him in the eight inning and Ron has allowed the tying or go ahead run to score. Ron just not the same pitcher he was in the first 2/3s of 2008, but Trey keeps using him as if he was the same pitcher.

Using Pitch F/X to Determine an Umpire's Strike Zone

2009-04-23T19:03:00.000-07:00

With Pitch F/X there has been quite a bit of work on looking at different pitchers. I decided to extend the work I did on umpire scoring and link that information to pitch F/X. I used the standard strike zone of -1, 1 for the width of home plate and then adjusted the batter's height to the vertical standard of 1.5 to 3.5. Here is a spreadsheet with the percentage of pitches that each umpire called strikes and ball and and how they compare to the standard strike zone (-1,1 to 1.5,3.5). I drew the cutoff line to include umpires with more than 1456 called balls and strikes (The bottom 11 umpires called 918 or less strikes).

Then, I ask for some help in determine what percent of pitches in a zone should constitute the area that balls and strikes are normally called by the umpire. I decided to take their advice and use the value of 90%. I increased or decreased the standard strike zone in equal increments until 90% of the pitches were strikes or balls. Here is the dimensions of the strike zone (along with the percentages of correct calls with the normal zone). I added these two differences together to get the amount these areas overlap (inconsistency of the umpire).

Average value for:	Value
% of strikes called strikes	78.8%
% of balls called balls	85.4%
% correct	83.2%
Adjustment for Strikes (90% zone)	0.15
Adjustment of Called Balls (90% zone)	0.09
Combined Adjustment Distance	0.24

Here is an image of the normal strike zone (black) along with the actual zone where strikes are called strikes (inside the red square) and the zone when balls are called balls(area outside the yellow box) 90% of the time.

After that, I went ahead and determined the zone for each individual umpire (on previously link spreadsheet).

Finally for a few of umpires, I have looked to find find how their strike zone compared to a centered zone. Basically, do they call the a little lower or outside. I having a tough time find a good to determine this zone, but am making progress, I will update these values as I figure them out and will let everyone know when the values are complete.

I went ahead and linked the previous data I had on umpire effects to the Pitch F/X, but I am wanting to do the R/G, K/G, etc for the same seasons I did the pitch F/X analysis. I want to include park factors and finally get an umpire factor for each umpire that can be applied to each pitcher to see how much the home plate umpire effected their season. As always thanks for your time and and I am open to comments and suggestions.

UZR Component Comparision

2009-04-17T09:31:00.000-07:00

Initial Position	DPR	RngR	ARM	ErrR	UZR	Final Position	Number of Matched Season

Outfield only
LF	0.00	-1.12	-1.39	-0.06	-2.61	RF	53.66
CF	0.00	7.47	2.39	0.26	10.20	LF	38.05
CF	0.00	8.57	0.40	-0.02	9.01	RF	37.07

Infield only
SS	0.48	2.07	0.00	1.88	4.33	2B	30.50
3B	-0.18	-1.64	0.00	1.61	-0.33	2B	28.42
3B	-0.35	-1.22	0.00	-0.01	-1.60	SS	21.30
1B	-0.57	-0.62	0.00	-2.47	-3.70	3B	19.88
1B	0.29	-2.16	0.00	-0.15	-2.18	2B	6.12
1B	-1.70	-8.50	0.00	-1.32	-11.23	SS	3.26

Infield to Outfield
1B	0.02	-5.00	-0.25	-0.29	-5.55	LF	15.73
1B	0.16	-7.82	-2.44	0.62	-9.47	RF	13.72
3B	-0.22	-3.95	3.61	1.91	1.50	LF	9.27
2B	0.51	3.65	4.07	0.87	9.09	LF	6.87
3B	-0.65	0.76	0.47	2.53	3.00	RF	6.15
2B	1.76	-1.80	2.09	0.54	2.65	CF	5.62
3B	-0.29	-9.80	4.61	3.57	-2.06	CF	4.39
2B	1.30	3.61	-0.63	0.21	4.28	RF	4.13
SS	-0.53	8.01	6.50	6.98	20.71	LF	3.15
1B	-0.29	-8.21	0.72	-0.28	-7.96	CF	2.30
SS	-0.15	-0.25	2.59	8.81	11.26	CF	2.24
SS	1.01	8.76	-2.05	3.18	10.33	RF	1.14

DPR (double play runs)
RngR (range runs)
ARM (outfield arm runs)
ErrR (error runs)
UZR (ultimate zone rating)

State of Players

2009-04-10T21:55:00.000-07:00

With both of the top players in the Royal's farm system coming from warm weather states (Eric Hosmer – Florida and Mike Moustakas – California), I wanted to see if was do to the state's extra days of sunshine or just the state's large population. To make the comparison, I looked at the data using the following guidelines:

I compared the baseball data to the 2000 Census – I could have used some combination of multiple census's (censi?), but figured the 2000 Census should be a reasonable example.
Without spending way more time than I have, I used the state that the player was born. I know that they could have moved, but the number should be similar.
Players born after 1950 were used because expanding the search before that point biased the results towards states that existed. Also the baseball numbers would more closely resemble the 2000 Census.
I compared the state's population to the number of players that appeared in any major league game (quantity),the Win Shares of these players (quality) and the combination of the two values.

With these constraints, all the results are available on Google Docs, but here are the states that produce the most and least players per state capita (the Overall Rank is based on a combination of the Win Shares and total player's rankings):

Overall Rank	State	State Population (Millions – 2000 Census)	WS/person	People per MLB Baseball Player
1	California	33.9	844	4881
2	South Dakota	0.8	998	6621
3	Mississippi	2.8	1223	6822
4	Wyoming	0.5	1120	7964
5	Louisiana	4.5	1421	7813
6	Oklahoma	3.5	1584	7807
7	Delaware	0.8	1602	7836
8	Alabama	4.4	1521	8519
9	Illinois	12.4	1614	8489
10	Ohio	11.4	1663	8733

42	Wisconsin	5.4	3269	18689
43	Maine	1.3	4099	20237
44	Idaho	1.3	3635	23106
45	Colorado	4.3	5522	22286
46	New Mexico	1.8	664	32483
47	Nevada	2.0	6364	27373
48	Utah	2.2	7919	29002
49	New Jersey	8.4	16762	13885
50	Vermont	0.6	7805	30441
51	Montana	0.9	16707	25061

The numbers seem to indicate that most players come from Sunbelt states and Midwest states. I wonder if the high ranking of the Midwest states if do to the long culture of baseball young kids are encourage to be big leagues. These Midwest states, though, produce about 2/3 the number of MLB ballplayers than California per capita. The states of Delaware and Wyoming don't seem to fit in that group especially since they have a lot more in common with the bottom ten states that are colder mountainous states. I have never lived in New Jersey and can't explain why it is ranked so low and maybe someone that lives or lived that can inform everyone. As always. I am open to comments and suggestions.

Home Plate Umpire's Effect of on Game Events

2009-03-26T21:08:00.000-07:00

With the start of the baseball season around the corner, it time to get the stats predictions for the most cherished people on the field, the umpires. More specifically, I wanted to examine the men behind the plate and how they effect the scoring of the games they call. The main portion of the data I used was from the website: The Logical Approach. It contains all the data on umpires since 2000, the year they were combined from being separate AL and NL groups to one MLB group. The following data was used: K/BB, BB/9, Base runners/IP, ERA and Total Runs/Game.

Procedure: To get to the final numbers, I calculated the total of all game scores for each umpire, added these up for all the seasons and averaged them to get the all time average values. Then I used this average and the yearly averages to create yearly “umpire factors” for each year to normalize the stats. I took the umpires that umped in all the nine years I had data for and determined the point when they had regressed to 50% (r-value) of there final value. Using these r-values, I predicted what the umpire's 2009 numbers. I have made the sheet available for all at Google Docs that includes all the umpires assigned to crews in 2009 (including the crew in which they are assigned) and also any umpire that umped at all in the majors in 2008. http://spreadsheets.google.com/ccc?key=pYilmL4DSNmkgSYzN_o80ZA.

Here is a chart of the averages off all the years data, the high and low values of the predicted values and the number of games where the umpire regresses to 50% of their actual value:

	K/BB	BB/9	Base Runner/inning	ERA	Total RPG
Average Values for All Umpires	2.01	3.16	1.40	4.59	9.55
Highest Predicted Value	2.85	3.53	1.47	4.91	10.33
Lowest Predicted Value	1.67	2.55	1.29	4.11	8.70
Number of games to get to 50% regression	35	97	133	97	194

To get a general idea which umpires are pro-pitcher or pro-hitter, I ranked each umpire in the 5 categories and then added these 5 figure together (not the most scientific method, but will work for general analysis). Here are the top 10 umpires for being pitcher and hitter friendly:

	Hitter Friendly
Rank	Name	Crew
1	Derryl Cousins	O
2	Paul Schrieber	H
3	Jerry Crawford	M
4	Mike Reilly	Q
5	Gerry Davis	L
6	Jerry Layne	K
7	CB Bucknor	L
8	Gary Cederstrom	D
9	Greg Gibson	N
10	Tim McClelland	N

	Pitcher Friendly
Rank	Name	Crew
1	Doug Eddings	J
2	John Hirschbeck	A
3	Brian Gorman	L
4	Phil Cuzzi	M
5	Bob Davidson	C
6	Bill Miller	O
7	Brian O'Nora	D
8	Ron Kulpa	B
9	Jim Wolf	D
10	Dan Iassogna	G

Finally, I used these numbers to see how they would effect a single pitcher. Using the ERA numbers, I looked at the Royal's Gil Meche in 2008. I added how much the each home plate umpire's ERA differed from the average ERA. He had an extra 2 runs scored against him because of umpire differences over the season. Taking these differences into account, his ERA drops from 3.98 to 3.90.

Improvements – I wanted to get this data able available before the start of the season, but feel it is far from complete. I plan on adding these two improvements for future analysis:

Incorporate Pitch F/X. I currently don't have all the data downloaded, but hope to have it soon. I plan on implementing something like Jonathan Hale did in this article for the Hardball Times. I would like to see how well the data correlates from umpires actual strike zone to the game stats.
Use Retrosheet data and Park Factors. With Retrosheet I would be able to go back further in time and be able to get any stat need (please suggest any additions that would be helpful to the research). Some umpires only work certain areas of the country, so their numbers might be skewed do to the parks they work in.

Conclusion: With a range of +/- 1.6 R/G for all the umpires, teams and pitchers will need to adjust for the umpire behind the plate. Hopefully I give a people a better understand of which umpires have what tendencies. Finally, I plan on expanding the research to have more and better data, including the data from Pitch F/X. As also I open to comments and suggestions.

Royal WAR

2009-03-25T15:03:00.000-07:00

Chone Smith went back recently and calculated the WAR (df's description in the comments) for all the batters back to 1955. I have finally gotten around to getting the single season and overall totals (link to the complete lists - let the comparisons begin) for the top 100 Royals in lifetime at bats. This worked out to 563 at bats and I will add to the list as time permits. So without further ado, here are your top (and bottom) 10 Royal's seasons and careers when measured with WAR.

Career Leaders and Laggards:

Name	Total WAR
Brett, George	89.3
Otis, Amos	42.7
Wilson, Willie	37.8
McRae, Hal	25.8
Beltran, Carlos	25.1
White, Frank	25
Sweeney, Mike	22.7
Mayberry, John	21.2
Porter, Darrell	18.1
Patek, Freddie	17.5

Gload, Ross	-1.5
Motley, Darryl	-1.5
Perez, Neifi	-1.7
Nelson, Roger	-1.8
Shumpert, Terry	-1.9
Thurman, Gary	-2
Vitiello, Joe	-2
Hedlund, Mike	-2.2
Hernandez, Jackie	-2.2
Tabler, Pat	-2.4

Individual Season Leaders and Laggards:

Name	Year	WAR
Brett, George	1980	9.8
Brett, George	1979	9.1
Porter, Darrell	1979	8.7
Brett, George	1985	8.3
Brett, George	1976	8.3
Beltran, Carlos	2003	7.7
Wilson, Willie	1980	7.6
Brett, George	1977	7.5
Otis, Amos	1978	7.5
Mayberry, John	1975	7.1


Guiel, Aaron	2002	-1.3
LaCock, Pete	1980	-1.4
Balboni, Steve	1987	-1.4
Gload, Ross	2008	-1.6
Perez, Neifi	2002	-1.7
Motley, Darryl	1986	-1.7
Berroa, Angel	2006	-1.9
Pena, Tony	2008	-2
Tabler, Pat	1989	-2.3
Piniella, Lou	1973	-3

Pitching WAR Aging Patterns

2009-03-13T12:52:00.000-07:00

I decided, along with Sky Kalkman's Bad Contracts discussion, to look into finding a players aging multipliers for WAR and apply them to project a few contracts by selecting a pitcher and seeing what his numbers project to do one year to the next.

Note/Disclaimer/Venting Rant: I originally started this endeavor with wanting to look at pitchers (starters and relievers) considered to be regulars and what they did the next year. I decided on 40 innings as being my point of reference for a regular pitcher. I went to Fangraphs and download all the data of pitchers with 40 innings or more. As you can tell, when I went to match the pairs later, I wouldn't catch the pitchers that had less than 40 innings in the second year, but greater than 0. I didn't notice the problem until I was almost done and was looking at my beloved Royals and couldn't find the 2008 numbers on Jimmy Gobble. It was then I relized my error. I decided to go forward with the semi-biased data and might go back later and redo it when I am not so mad at myself.

I looked at all the pitchers between 2002 and 2008 that pitched 40 innings in consecutive years. Then I divided the first year by the second year for the Starting, Relief and Total War and also for Starting, Relief and Total innings pitched to get the multiplier factor for each of these factors. Here are the results:

Adjustment (going to age):	Staring War	Starting IP	Reilver War	Relief IP	Total WAR	Total Innings	Number of Paired Seasons
22	1.38	1.41	1.75	0.90	1.42	1.34	11
23	1.53	1.26	1.10	1.05	1.51	1.23	34
24	1.18	1.16	0.86	1.08	1.12	1.15	72
25	1.17	1.08	1.23	1.01	1.18	1.07	94
26	0.99	1.03	1.25	1.17	1.02	1.05	118
27	1.02	0.98	0.74	1.05	0.98	0.99	132
28	0.99	1.03	1.25	1.17	1.02	1.05	134
29	0.80	0.84	1.05	1.08	0.85	0.89	132
30	0.92	0.99	1.02	1.04	0.94	1.00	141
31	1.03	1.02	0.63	0.94	0.90	1.00	133
32	0.85	0.84	1.14	1.03	0.92	0.90	104
33	0.68	0.82	0.69	1.07	1.04	0.89	86
34	0.84	0.88	0.74	1.05	0.81	0.94	58
35	0.80	0.85	0.73	0.99	0.77	0.90	39
36	0.80	0.91	0.89	1.06	0.81	0.96	19
37	0.80	0.85	0.93	0.90	0.87	0.87	20
38	0.60	0.87	0.74	1.09	0.68	0.98	10

There is for sure a lack of data and the extremes, but significant number for reliable numbers from 24 to 34. Here is the graphs of starting and relief WAR for each year:

A few notes I have observed:

Pitchers add to their WAR to ~28 years old and stay fairly level until they begin to decline at age 32.
I averaged the Total WAR values of pitchers that did not reach 40 innings the next year their WAR averaged 0.46. Replacement level pitchers get replaced.
70% of all pitchers did not reach the 40 innings from year to year.

With this information, I decided to look at the contracts signed this off season by C.C. Sabathia and A.J. Burnett by the Yankees. I just used there 2008 WAR numbers to base the next year on. What exact number is used is debatable, but I used 2008 for ease of discussion/calculation. I used the dollar value of the wins from Fangraphs of 4.5M to be paid for each WAR in 2008. I needed to calculate the inflation value for the future. The percent increase in the past was for baseball salaries was ~10%. I decided to go a little more conservative consider the economy and went with 7%. Here are the two pitchers:

Name	Amount of Contract (Millions)	Years of Contract	Amount per Year
C.C. Sabathia	$161.00	7	$23.00		Multiplier Used
					1.07
Age	Innings	Innings Multiplier	WAR	Starter Multiplier	Projected Value of 1WAR	Projected Amount of Value (millions)
28	353		7.7		4.5
29	296.37	0.84	6.19	0.8	4.82	$29.83
30	293.63	0.99	5.68	0.92	5.15	$29.24
31	299.3	1.02	5.85	1.03	5.51	$32.25
32	251.94	0.84	4.95	0.85	5.9	$29.18
33	205.83	0.82	3.38	0.68	6.31	$21.31
34	180.27	0.88	2.85	0.84	6.75	$19.26
35	152.53	0.85	2.27	0.8	7.23	$16.43
		Total War =	31.17		Estimate Value of contract =	$177.50
		Value with constant value for WAR (4.5M) =	$140.26

Name	Amount of Contract (Millions)	Contract Length (Years)	Amount per year
A.J. Burnett	$82.50	5	$16.50		Multiplier Used
					1.07
Age	Innings	Innings Multiplier	WAR	Starter Multiplier	Projected Value of 1WAR	Projected Amount of Value
32	220		5.7		4.5
33	196.51	0.89	3.89	0.68	4.82	$18.73
34	184.26	0.94	3.29	0.84	5.15	$16.93
35	166.68	0.9	2.62	0.8	5.51	$14.45
36	160.51	0.96	2.09	0.8	5.9	$12.34
37	139.83	0.87	1.68	0.8	6.31	$10.59
		Total War =	13.57		Estimate Value of contract =	$73.04
		Value with constant value for WAR (4.5M) =	$61.05

As you can see C.C.'s contract doesn't look like a good deal with today's prices ($140M vice $161M) , but when you take inflation into consideration, it doesn't look as bad ($177M vice $161M). A.J.'s on the other hand doesn't look like a good deal for the Yankees either way ($61M/$73M vice $82.5M).

Let me know if you have any questions/commnets and remember to double check your data.

Ranking of Royal Shortstops According to individual season WAR

2009-03-11T12:54:00.000-07:00

Chone Smith has recently released Wins Above Replacement Level (WAR) for all positional player 1955 to 2008. I went and look at the recent KC black hole, the shortstop position. Ever since the pair or Freddy Patek and U.L. Washington (with both of them splitting duties in 1979 for the transition), the Royals have not be able to fill the position until hopefully Mike Aviles steps up.

Here is the list of top and bottom ten seasons for Royal Shortstops according to there WAR. I included Win Shares for comparison.

WAR Rank	Year	Name	WAR	Win Shares	Win Shares Rank
1	1997	Jay Bell	5.5	21	2
2	2008	Mike Aviles	3.9	17	5
3	1971	Freddie Patek	3.8	24	1
4	2003	Angel Berroa	3.8	16	6
5	1999	Rey Sanchez	3.6	12	17
6	1993	Greg Gagne	3.2	18	4
7	1976	Freddie Patek	3.1	14	13
8	2001	Rey Sanchez	3.1	10	26
9	1982	U L Washington	2.9	18	3
10	1973	Freddie Patek	2.6	16	7

44	1986	Buddy Biancalana	-0.5	4	40
45	1987	Angel Salazar	-0.5	5	33
46	1982	Onix Concepcion	-0.9	2	49
47	1970	Jackie Hernandez	-1.0	3	44
48	1985	Onix Concepcion	-1.0	5	34
49	1979	Freddie Patek	-1.1	3	45
50	1969	Jackie Hernandez	-1.2	7	31
51	2002	Neifi Perez	-1.7	6	32
52	2006	Angel Berroa	-1.9	1	52
53	2008	Tony Pena	-2.0	0	53

A few observations:

Gotta love the bottom three, just gotta.
Berroa has the largest gap (5.7 WAR) between best and worst season, with Patek coming in 2nd (4.9 WAR)
Why didn't we keep Jay bell longer
The Dye for Perez trade looks even worse considering what Sanchez was giving us.

Here is the list of all the Shortstop(s) over the seasons with the their stats combined:

Year	Name(s)	War	WinShares
2008	Pena Aviles	1.9	17
2007	Tony Pena	0.8	10
2006	Angel Berroa	-1.9	1
2005	Angel Berroa	0.5	11
2004	Angel Berroa	0.5	10
2003	Angel Berroa	3.8	16
2002	Neifi Perez	-1.7	6
2001	Sanchez/Perez	3.1	13
2000	Rey Sanchez	2.3	9
1999	Rey Sanchez	3.6	12
1998	Lopez/Halter	0.3	8
1997	Jay Bell	5.5	21
1996	David Howard	0.4	9
1995	Greg Gagne	1.9	10
1994	Greg Gagne	1.5	11
1993	Greg Gagne	3.2	18
1992	Rossy/Wilkerson	0.9	12
1991	Stillwell/Howard	0.5	13
1990	Kurt Stillwell	1.1	12
1989	Kurt Stillwell	1.9	15
1988	Kurt Stillwell	2.6	16
1987	Angel Salazar	-0.5	5
1986	Salazar/Biancalana	0.3	8
1985	Biancalana/Concepcion	-1.1	8
1984	Washingtion/Concepcion	1.7	10
1983	U L Washington	0.6	13
1982	Washingtion/Concepcion	2	20
1981	U L Washington	1	8
1980	U L Washington	2.1	16
1979	Washingtion/Patek	-0.4	8
1978	Freddie Patek	1.5	11
1977	Freddie Patek	2.3	15
1976	Freddie Patek	3.1	14
1975	Freddie Patek	0.8	11
1974	Freddie Patek	2.1	15
1973	Freddie Patek	2.6	16
1972	Freddie Patek	2.4	13
1971	Freddie Patek	3.8	24
1970	Matchick/Hernandez/Severson	0.1	10
1969	Jackie Hernandez	-1.2	7

How do I determines a player's real/predicted UZR including age adjustments?

2009-02-27T11:15:00.000-08:00

Question: How do I determines a player's real/predicted UZR including age adjustments?

Why I asked the question: I needed a way for my Manger's Scorecard to determine a player's defensive ability and I couldn't find a good method. I was told that an average player's UZR should decrease by ~1 UZR per year , but my work on CF UZR/150 showed different results.

Analysis:

Positional adjustments

I collected the UZR for all players since 2002 from Fangraphs that had over 100 innings at each position. I decided not use true UZR because a few bad plays in a small sample size can lead to a player with a fairly large positive or negative UZR. I needed to regress the data to a value that is more indicative of the player's actual ability. I used the value of 125 games (as recommended by Mitchel Lichtman) to be the point when 50% of player's UZR can determined from their current stats and the rest will be the league average. The value of 125 games is a little more conservative value than the 100 value mention in this article at Fangraphs (http://www.fangraphs.com/blogs/index.php/fielding-update-arms-and-double-plays), but I will use it for my analysis.

I assumed that a player regresses to a value of 0, so for each player I got their yearly regressed UZR (rUZR) with the following equation:

rUZR = (1-((125)/(number of games + 125))) * Yearly UZR)

I ran this formula on all players (~3200 players), except pitchers and catchers (who have no UZR data on them) born after 1968 (the data from Fangraphs doesn't have playerID, so I was getting too many duplicate names). I combined all the positions and got the following results:

It is tough to make too many conclusions from the data, but the following can be observed:

Players are significantly better when they are younger than 30 than when they are older than 30. Both of these age groups seem to have values grouped together.
Even though the general trend in downward, there seem to be a learning curve for a few years, a peak and then a slow decline.

After that, I summed up the data for each position and got the following average numbers for the yearly rUZR for 23 to 34 year olds (above and below these ages, the amount of data for certain position, shortstop or example, was severely lacking):

	Yearly regressed numbers
Position	Change in UZR per year	R-squared
1B	-0.042	0.155
2B	0.075	0.296
SS	0.054	0.161
3B	0.006	0.001
RF	-0.124	0.458
CF	-0.127	0.274
LF	-0.127	0.732
All Positions	-0.032	0.431

The r-squared shows the level of correlation in a direct change over time with the left field numbers actually being statically significant and the third base numbers not having any correlation.

Here is a chart of the actual rUZR values by year and position:

Age	1B	2B	SS	3B	RF	CF	LF
23	0.30	-1.04	-0.15	1.64	0.76	-0.17	0.50
24	0.44	-0.56	-0.11	-0.81	0.14	0.86	0.70
25	-0.23	0.15	-0.17	0.33	1.21	1.38	0.49
26	-0.46	0.00	0.39	-0.53	0.29	0.84	-0.07
27	0.01	0.55	0.74	-0.41	-0.03	-0.17	0.63
28	0.67	0.43	0.44	0.19	-0.36	1.27	0.41
29	0.02	0.63	-0.32	0.83	0.40	0.98	-0.44
30	0.36	0.27	0.27	0.71	0.46	0.54	-0.33
31	0.06	0.22	0.11	-0.06	0.01	-0.78	0.01
32	0.07	0.20	-0.25	0.63	-0.56	-0.60	-0.54
33	-0.40	-0.23	0.20	0.53	0.11	-0.60	-0.76
34	-0.56	0.57	1.36	-0.07	-1.38	0.08	-0.66

The preceding data seems to indicate that there are 3 categories for rUZR:

Outfielders – All three groups show about the same decrease of ~0.125 rUZR per year of age.
Corner Infielders – These both are pretty steady over the years with values close to 0, but 1B seems to decline a little more than 3B and its decline rate is closest of any position to the overall decline rate.
Middle Infielders – These two positions were a little confusing in that they had a trend of an increase in regressed UZR over the years. I needed to look into them a little more. I selected the players that played 4 straight seasons at SS and looked to see how much of change there was from year to year and the results were the same. The players seem to having an improving rUZR to a peak around age 27 and then decline after that. Here are the graphs for these 2 positions:

Once you look at the positional rUZR, the overall rUZR values makes a little more sense when the following 3 trends are combined:

General overall decline(outfielders)
Increase at the beginning (middle infielders)
General overall leveling (corner infielders)

Creating a prediction formula

After collecting the information on individual positions, I wanted to look at using the age adjustment to help predict the players actual rUZR scores. Through the discussion with Mitchel Litchman, he recommended the following formula which weights the recent years more than the previous years:

((5* # of games last year * UZR last year)+

(4 * # of games 2 years ago * UZR 2 years ago)+

(3 * # of games 2 years ago * UZR 3 years ago)+

(2 * # of games 2 years ago * UZR 4 years ago))/

((5 * # of games last year) +

((4 * # of games 2 years ago) +

((3 * # of games 3 years ago) +

((2 * # of games 4 years ago))

I ran this formula against all the players with over 100 games at a position for 5 straight years and using this system, it was a very good predictor. There was a slight decimal difference for perfect prediction, but it was a very good method. I also found out that using a weight of 2,1 (2 for last year, 1 for 2 years ago and 0's for the rest) was equally as good.

The one fact I did find was that -.03 was not a good indicator of the loss of defensive ability from these players. The main reason I saw a larger increase in the negative direction is that I was predicting using 4 years of data to predicted the 5^th year, so the youngest players could not be predicted until they are ~25 years old. Here are the age adjustments and standard deviations on being able to predict a player's UZR using the 5,4,3,2 method and the 2,1 method:

	2,1 Age Adjustment	Standard Deviation for 2,1	5,4,3,2 Age Adjustment	Standard Deviation for 5,4,3,2
Year's Regressed UZR	-0.35	1.61	-0.70	1.87
Year's Actual UZR	-1.00	4.44	-1.35	4.45

r-squared of rUZR vs predicted UZR = 0.65 and r-squared of UZR vs predicted UZR = .66

I might go back sometime and figure out the overall and positional age adjustment using the 2,1 formula. For now I feel the data is good enough for the Manager's Scorecard, but it does raises some questions on exact positional aging patterns.

I have put a spreadsheet on the web (with a web page coming soon) that a person can download and enter a player's Games/Chances, UZR, regression adjustment and age adjustment and the player's rUZR will be exported.

There is not great understanding of defensive metrics, but my hope is to help remove some of the old clutter (and possibly created new clutter) and help people interpret the defensive metrics like UZR better. As always, I open to comments and suggestions.

How do to measure how well a manager manages?

2009-02-24T11:39:00.000-08:00

Question: How do to measure how well a manager manages?

Why I asked the question: There has always been a debate on how much of an impact a manger has on a team's performance and a game's and season's outcome. I wanted a thorough, yet relatively easy way to measure a manager usefulness. My main focus now with the initial release is to try to find any huge glaring errors so it won't have too many mid-season changes.

Analysis:

General overview

Finding a way to measure the many different aspects of a manager's job is close to impossible, but this my best shot. I am calling it, the Manager Scorecard and am hoping that others would join me (if not, oh well) We could start getting some measurable impact of mangers over an entire season, especially with the number of games the team wins.

The Scorecard will actually look at all of the team's managers. If the pitching coach leaves a pitcher in too long or a 3^rd base coach sends someone home and they get thrown out, the entire staff will take the hit. I am not looking at any decisions that would be made by the general manager (trades, player call ups). The General Manager Scorecard is something on my to do list and should be ready by the beginning of this upcoming off season.

Finally, some of the data I plan on collecting can easily be mined from old game logs to determine how a manager performs. At the end of the season, I plan on comparing how a quick sample of data (IBB, CS/SB, Sacrifice Buts, pitcher abuse, etc) to the results from the Scorecard and see how the numbers correlate.

I was able to find some way to measure the manger's effect in following aspects of a game:

Theoretical runs gained or lost

Lineup utilization - This area is probably the most over analyzed while having the least impact on the outcome on the game.
IBB - Looking to see if managers intentionally walking someone is a good or bad decision.
Fielding – UZR/150 is going to be used to see if a player is obviously out of position.
CS/SB – Does the advantage of the extra base out weigh the disadvantage of causing an out?
Effectiveness of hit and run – How much this call helps or hurt in getting the batter safe and the runner safely advanced to another base?
Sacrifice bunts – Does the giving the out up, justify moving the running up another base?
Pinch hitter/pinch runner - Are they effectively utilized later in a game to generate more run scoring opportunities?

Bullpen management – Are pitchers used correctly as the game goes on? How long is the starting pitcher used compared to when they usually begin to break down? Also, is the bullpen used so that the pitchers face the batters when they have a distinct advantage?

Bullpen usage – Are the best relief pitchers being used in situations when the game is at the most critical moments?

Bullpen Fatigue - Looks to see if certain relievers get over used and their performance declines because of this over usage.

Starting Pitcher Abuse - Looking to see if some managers over use their starters and the effects on them later in their career. The up to date stats can easily be found at Baseball Prospectus website . Also the number of innings pitched will be tracked for the Verducci effect. The rule states that young pitchers that increase significantly the number of pitches thrown from one year, take a major step back the next year.

Finally, there are other aspects that a manager does that are not easy to measure such as team chemistry, player development and scouting. These might be added later as I can find a way to measure them.

Source of statistics:

I was wanting to limit the number of locations to visit to get player projections and the season's stats. I have it so the user only has to go to three web sites:

BaseballProspectus.com - stater abuse points
Baseball-Reference.com - starting pitching breakdown points
Fangraphs.com - everything else

It will take a little daily updating, but once the spreadsheet is filled out for the first game, it will be fairly easy to maintain. Also, game data should be drawn from any source possible, articles, TV, radio, MLB games, game box scores and public questioning in order to get all the facts from the game.

I am not for sure exactly how much of an effort it will take to keep a scorecard for an entire season, so I am only going to following two managers this season, Trey Hillman's of the Royals and Ron Gradenhire's of the Twins. I feel this first season, there will be several changes to the system due to unforeseen game situations and some aspects being incorrectly weighted, but I will need to begin the evaluation.

The initial scorecard is available for download from here: Link I will update the scorecard with more detailed instructions closer to the beginning of the season. Also, a website will be set up to display the various final scores of Hillman and Gardenhire for comparisons and other individuals, that want to, can post the individual game scores of the managers they track.

Basic Instructions for the Manager Scorecard

Important: Even tough I have have written some rules/guidelines, the user of the scorecard need to use good judgment when making filling it out. It might be ideal for a player to be in a lineup everyday for maximized run generation, but players need breaks too. An example would be if Trey Hillman plays Mike Jacobs at 1B for more than 10 games during the season. I can see a few games to give others a break, but his defense is so bad, he should only be used as a DH. The decisions are totally up to the users.

The entire Manager Scorecard is a spreadsheet with 5 main tabs (Seasons_Total Score, Starting_Pitcher_Abuse, Bullpen_Fatigue_and_Abuse,Roster_Projections and Additional_Comments) and two tabs for each game (Gameday_Sheet_XX_XX_XX and Gameday_Sheet_XX_XX_XX_copy) . Here are the instructions on what each sheet is looking for and how to set each one up.

Note – I use the “Calc” program from OpenOffice when creating the spreadsheet and save it out as an .xls. If there are any problems with it working in Excel or other spreadsheet programs please let me know and I will try to resolve the issue.

Tab 1: Seasons_Total_Score

This tab contains totals from the rest of the tabs for easy access. I am hoping to have each section weighed somewhat evenly (I can see some multiples being added as the season goes on). The only work that needs to be done on this sheet is to copy over the daily Bullpen Management and Runs Gained/Lost from the daily copy of the Gameday sheet (more on why a copy of each Gameday sheet is used in its section). The rest of the data should update automatically.

Tab 2: Starting_Pitcher_Abuse

The top half of the page checks for the Verducci effect. The number of innings pitched in the previous year in the major and minor leagues and the total number of innings pitched during the current season will be inserted into this page. For every five innings over the threshold, add one point to the Points of Abuse column. The threshold for the number of innings the pitcher is not allowed to go over is set to 30.

On the bottom half of the page, the numbers are inserted from the Baseball Prospectus web page on pitcher abuse. The data can be entered into the section by coping and pasting into the correct place or entering data by hand. The final column labeled Stress will be divided by 10 and they the totals will then be added up for the team.

Tab 3: Bullpen_Fatigue_and_Abuse

This has been the toughest to figure out and I have found very little information on it on the web. I basically have the numbers set up so that when a pitcher pitches for 3 days in a row, the manager will get “awarded” with an abuse point. For each day I am taking .7 times the previous days value and adding the pitches thrown that day to come up with that day's value. If the number exceeds 90, the pitcher gets an abuse point. If anyone knows a better method please let me know. All the relievers will need to be entered by hand and the number of pitchers per game will need to be copied over from each Gameday Sheet.

Tab 4: Roster_Projections

Initially filling out this tab will be the most time consuming event and it will need to be updated daily also. There are 3 distinct areas, Hitters, Bullpen and Starting Pitchers.

Hitters – A weighted average of the 4 prediction (25% Marcels 25% Chone 25% Bill James 25% Oliver) from Fangraphs.com is used. Feel free to use what ever projection system or combination you feel appropriate. For determining the players ability, the yearly predicted numbers will be added to the year-to-date numbers to get the player's current ability/stats.

Also, I am attributing the lifetime UZR/150 from FanGraphs.com for each position the player has ever played (I have recently come upon another method, besides life time numbers and need to test it first. If it is easy to use and understand. If it is easy, I will enter it in the final version).

Bullpen – using The weighted averages of the different available projection systems from FanGraphs.com is used to determine the pitchers FIP. I am using FIP as it is the best measure of pitching performance available from projection data.

Starting Pitchers – In this area, the lifetime splits from baseball-reference.com are entered to see if there is point during the game when the pitcher really breaks down. The three areas that can be measured are:

Inning
Total pitches thrown
Time through the lineup

These breaks aren't that precise, but as of right now it is the best available information I can find.

Finally, I am really torn here on including all available player information, while still making it easy to do a assessment of each game in under 5 minutes. As of right now, I am including just the player's projected line and calculating their OBP, SLG and FIP.

Tab 5. Additional_Comments

This tab contains the factors that aren't measurable, but might have a major effect on the team's performance (e.g. kicked out of game, called out player/team in public, etc). If there are enough of these comments/factors not checked in other categories, they could be added together in a new area or they could give a point total for each event.

Tab 6: Gameday_Sheet_XX_XX_XX (where XX_XX_XX is the date)

Areas where run expectancy can change

Lineup – All batters will need to be imported and put in their actual lineup spot. Then using a lineup analysis tool, such as this one at Baseball Musing , the run expectancy of the lineup is generate. The OBA and SLG will need to be adjusted depending on if they are facing RHP or LHP. The user needs to determine if the lineup could be improved in any way (e.g. different order of players or different players starting). If there is a better way to construct the lineup, put the run differential in to this category.

Defensive Placement – Use UZR/150 to determine if the correct people are positioned in the field. If there should be an adjustment, take the difference in UZR/150 and divide by 150 to get the runs lost for incorrect positioning. Make sure if you add a batter for his offensive ability, adjust for his defensive ability also.

Pinch hitter usage - To determine if a pinch runner should be used, subtract .074 from OBP and .042 from SLG along with the correct left or right hand splits. Compare the pinch hitter to players already in the lineup and see if he should insert as a batter late in the game. If the bench player is better than one in the game, insert the player in to the on-line lineup run predictor calculator to get the difference in runs after they are in the lineup. Take the difference and divide it by 38 PA/G (~ average PA per game over the last 3 years) to get the amount of runs lost.

Other - Any situations that arise in the season where any of the preceding situations don't apply.

Manager Game Time Decisions

The various situations that manager has control over that can be measured with a run value are:

Intentional Base on Balls
Sacrifice bunts
Stolen bases
Base running – The user might need to watch/ listen to the game see if it was the manager or the player's decision to take an extra base. If there wasn't a play at the base, it should not count towards or against the manager
Hit and run

When the situation happens, take the run expectancy before and after the event from the Fraphgraphs game log, and insert the difference. Late in the game the Win Expectancy might increase even though the Runs Expectancy didn't. These events should not be counted against the manager.

Pitching Staff Usage

Bullpen – This area is where to rank which relief pitchers to use verses right and left hand batters. The league average splits for FIPS are located in this area. There is also a chart from Between the Numbers that shows the situations when the team's ace reliever show be used. Besides the chart on the Ace reliever, the usage of which reliever to use when is not very precise. Previous usage, game score, match ups, etc will come into effect to determine who should. What the user should be looking for here is see if the obvious case when a pitcher should not be in the game. To measure incorrect usage, the number of batters faced incorrectly with be counted..

Also in this area is where the number of pitches are tracked to help measure the each reliever's fatigue. I used the “pitch one inning for 2 days, one day off, repeat and you won't be getting fatigued” mantra as my measure stick. Also, a pitcher that pitches 3 days in a row, but to only 1 batter each time won't also get fatigued. I assumed it takes the average 24 (12 full speed and 36 not at full speed – divided 36 by 12) pitches to warm up from reading various articles. If it is determined by observation that each individual pitcher uses a different number of pitches, feel free to use the actual number the pitcher uses. Once the number of pitches are calculated, they will be link to Tab 3 (Bullpen_Fatigue_and_Abuse).

Starting Pitcher – This area shows when the starting pitcher begins to fatigue. These values are taken from the Roster_Projections tab and are the time when a pitcher breaks down. Once 2 of the 3 situations occur, each batter the pitcher pitches beyond that point is added up and added to count of batter not faced in optimal conditions. A batter can only be incorrectly counted once, not once for each the bullpen and the starter.

Tab 7: Gameday_Sheet_XX_XX_XX_copy

Make a copy of the each Gameday_Sheet before starting a new and paste it into this sheet using Paste Special (in OpenOffice Calc). The regular Gameday_Sheet has values that are linked to other parts of the scorecard and these values on the Gameday_Sheet will change due to other parts of the Scorecard changing. This copy is created to preserve the data exactly as it was on the day it happened.

Additional comments

Here are some Possible other additions to the scorecard. Currently, I believe they take up too much time for not extra information, but that could always change:

Look at LH and RH splits greater than or less than the league average
Other splits - power/finesse/average, fly ball vs ground ball,
Rotation management – rest between starts for starting pitching

As always, I will gladly accept suggestions (too much data, too little data, too hard to implement,etc) on ways to add to or improve the Manager's Scorecard.

Post Season Equivalencies for various sports

2009-02-15T22:04:00.000-08:00

Here is a look at post season equivalents between various sports. I came up with 9 levels, with 9 being the most difficult to reach (least percentage chance).

Notes: Would love to include golf be not sure what is best way since it doesn't really have a post season. Winner and everyone else. Pick a tournament like the Masters and rank from there. Feedback on this would be great.

Sport - Total Teams - Teams in post season - % making post season

MLB – 30 – 8 - 26.7%

NFL - 32 - 12 - 37.5%

NHL - 30 - 16 - 53.3%

NBA - 30 - 16 - 53.3%

College Football – 120 – 68 - 56.7%

College Basketball - 347 - 97 - 18.7%

College Baseball - 196 - 64 – 32.7%

NASCAR – 42 -12 – 28.6%

World Cup – 205 – 32 – 15.6%

Level 1 (53.3% to 56.7% of all teams)

NHL - Conference Quarterfinals

NBA - First Round

NCAA Football – any bowl game

Level 2 (26.7% to 37.5% of all teams)

MLB - First Round

NFL - Wild card Round

NCAA Baseball - Regional Round

NASCAR – Chase for the Sprint Cup

Level 3 (25.0% to 28.6% of all teams)

NFL - Divisional conference round

NHL - Conference Semifinals

NBA - 2^nd Round

NCAA Basketball – any post season tournaments (NIT or NCAA tournament)

Level 4 (13.3% to 20.0% of all teams)

MLB - NLCS or ALCS

NFL - AFC or NFC conference championship

NHL - Conference Finals

NCAA Basketball – NCAA Tournament

NCAA Football – Major New Year's Day Bowl

World Cup – Group Stage

Level 5 (6.3% to 9.2% of all teams)

MLB – World Series

NFL – Super Bowl

NHL - Stanley Cup Finals

NBA – Finals

NCAA Football – BCS Bowl game

NCAA Basketball – Round of 32

NCAA Baseball - Super Regional

World Cup – Knockout State – 16 teams

Level 6 (3.9% to 4.6% of all teams)

NCAA Basketball – Sweet Sixteen

NCAA Baseball – College World Series

World Cup – Knockout State – 8 teams

Level 7 (2.0% to 3.3% of all teams)

MLB – World Series Champion

NFL – Super Bowl Champion

NHF – Stanley Cup Winner

NBA – Champion

NCAA Basketball – Elite 8

Sprint Cup Champion

World Cup – Knockout State – 4 teams

Level 8 (1.0% to 1.7% of all teams)

NCAA Football – BCS National Championship Game

NCAA Basketball – Final 4

NCAA Baseball – Final best of 3 series

World Cup – Knockout State – Championship Game

Level 9 (0.3% to 0.8% of all teams)

NCAA Football – BCS National Champion

NCAA Basketball – National Champion

NCAA Baseball – College World Series Champion

World Cup – Champion

How many games does it take to accurately determine a player's UZR/150 defensive rating?

2009-02-12T15:30:00.000-08:00

Question: How many games does it take to accurately determine a player's UZR/150 defensive rating? (Link to explanation of of UZR and UZR/150)?

Why I asked the question: In The Hardball Times 2009 Baseball Annual, Coco Crisp was given a defensive grade of F-. It seems that he was not the worst possible center fielder and his ranking was unfair. I need to find out how close a single season's value corresponds to a player's lifetime value.

Analysis: For the study, I needed to find at what levels are seasonal UZR/150 reflective of lifetime skills. I chose to use the data for the 43 Center Fielders who had a combined 1000 innings played in center field since 2004 (year that UZR/150 goes back to). This data was obtained from wwwl.FanGraphs.com.

Note on Data: I wished to increase the sample size, but being able only being able to copy and paste each player's stats, so I limited the scope on this initial look. Also the data only goes back to 2004 for UZR and it is only divided into single season data. I understand that there is some definite issue about lack and quality of the data, but I feel I used the best that is available.

For these 43 players, I wanted to compare the player's available lifetime UZR/150 to the UZR/150 the player had for each year. I used standard deviation to compare the difference of the two values. I also divided the data for the number of games played in a season. From the research I got the following results:

Minimum Games Played	Minimum Innings Played	Standard Deviation
0	0	19.98
25	225	6.56
50	450	5.95
75	675	5.81
100	900	5.24
125	1125	5.36
150	1350	6.45

So generally, a 66% (1 standard deviation measure that value that 66% of all available values fall in) of all players will have have their UZR/150 vary from 0 to 6 runs while 95% of all players will have their UZR/150 vary from ~ 0 to 12 runs.

Also I looked at how a player's lifetime UZR1/150 compares to the UZR/150 at each time in the player's career. Again I used standard deviation to measure the variation:

Minimum Games Played	Minimum Innings Played	Standard Deviation
0	0	22.92
50	450	3.45
100	900	3.07
150	1350	2.65
200	1800	2.78
250	2250	2.74
300	2700	2.14
350	3150	2.09
400	3600	1.88
450	4050	1.67
500	4500	1.39
550	4950	1.18
600	5400	1.05
650	5850	0.80
700	6300	0.60
750	6750	0.76

As you can see, over a player's career they get within 2 points of their final UZR/150 within 400 games or ~3 years of time. The reason these numbers have less variation vice the yearly data is that with so few of years worth of data, and with some players only having a couple years in majors, the final value me be closer to the current value.

To make sure the previous data isn't too far off, I took the 24 players that had 5 or more years in center field and compared their UZR average from the first 3 years to the their final. They had a standard deviation of 2.62 runs over 1028 innings which compares close to the preceding graph.

Next, I need to find how UZR/150 changed as the players aged. I grouped the players depending on their age and got the following information:

Age	UZR/150
22	0.02
23	-0.06
24	0.48
25	0.97
26	0.92
27	0.01
28	1.08
29	0.64
30	0.54
31	-0.06
32	0.00
33	-0.62
34	-0.22
35+	-0.88

It seems that defensive the center fielders peak between the ages of 25 and 28 (not sure what they are doing in their 27^th year) and seem to decline to the league average at the age 31. Not great information here, there does seem to be an ~2 point swing in UZR/150 over a player's career.

Finally, did Coco deserve better than his F- ranking. To begin with her is a look at Coco's available defensive stats over the years he has played center field:

Season	Team	Position	Innings	Lifetime Games	UZR	UZR/150	Lifetime total UZR	Lifetime UZR/150
2002	Indians	CF	269	30	-0.2	-1	-0.2	-1
2003	Indians	CF	462	81	-2.2	-6.43	-2.4	-4.43
2004	Indians	CF	807	171	6.6	11.04	4.2	3.69
2005	Indians	CF	79	180	-1.7	-28.98	2.5	2.09
2006	Red Sox	CF	900	280	0	0	2.5	1.34
2007	Red Sox	CF	1216	415	22	24.42	24.5	8.86
2008	Red Sox	CF	886	513	-8.9	-13.56	15.6	4.56
Totals			4620		15.6

Since the data in the annual was base on just the 2008 season, his score of -13.56 corresponds to their grade of F- . The problem is that 1 year's worth of data is not enough to determine how well a player plays defense over his career. Coco's defensive numbers have been all over the place over his career (much more then the average center fielder), but his lifetime value of 4.56 for UZR/150 pretty close to his ability (average center fielder – Grade: C). In 2008 he was about 18 runs off his lifetime average, which is about 3 standard deviations (18/3) off. Only 5% of all players will vary is much in a season, but it is not totally out of the norm. In my opinion a respectable publication like the Hardball Times should included, besides the seasonal defensive grade, is a lifetime or 3-year grade.

How does the fielding of 2nd and 3rd baseman compare to when they play the opposite position?

2009-01-28T14:20:00.001-08:00

Question: How does the fielding of 2^nd and 3^rd baseman compare to when they play the opposite position?

Why I asked the question: Mark Teahen of the Royals is being asked to practice at 2^nd base for the up coming season. Using other players as comparison, how should Mark project at 2^nd.

Analysis: I used UZR as the defensive metric which is available at Fangraphs.com. UZR is the best freely available defensive metric I can find. As with any defensive metric, it is far from perfect, but for now it is the best to use. I looked at all players from the last 3 season that played at least 100 innings at both 2^nd and 3^rd base. I came up with 43 players. To equally weight the weight the at bats I determined the Harmonic mean in at bats between innings and UZR at 2^nd and 3^rd

Harmonic mean =2/((1/innings at 2nd)+(1/innings at 3rd))

Note: I had a discussion with Tom Tango on his blog on the best way to measure the differences.

I then added the UZR and innings for all players at 2^nd and 3^rd and got the following results:

Average UZR/150 games at 2B: -1.90
Average UZR/150 games at 3B: -0.85

Then I group the the players that played the majority of time at 2B and got the following results

Average UZR/150 games at 2B: -1.73
Average UZR/150 games at 3B: -0.49

Finally I group the the players that played the majority of time at 3B and got the following results

Average UZR/150 games at 2B: -2.21
Average UZR/150 games at 3B: -1.49

There are the general numbers showing how it seems that 2^nd base is harder to play than 3^rd base for this group of players over the last 3 years.

Lets take this data and extend it further to see how it effects the Royals and the decision to move Mark Teahen to 2^nd. From the above data, we can determine that Mark then should see a change of -1 UZR/150 from his lifetime 3^rd base average (-11.6) to -12.6, which would make him the worst defensive 2^nd baseman when comparing his -12.6 UZR/150 to all 2^nd baseman with over 400 innings played last year. So if Mark played 2^nd base for 150 games next season, his defense would cost us 15 runs compared to your average defensive 2^nd basemen.

Callaspo on the other hand rates at 5.9 UZR/150 at 2^nd base. Using a Run Generator calculator to determine the difference, the Royals will score 6 more runs with Teahen in the lineup vice Callaspo if they were each to play 150 games at 2nd. As a whole, the Royals are -12.5 runs worse as a team if they were to play Teahen at 2^nd vice Callaspo, or ~1 win.

How does a players ability and the team he signs withs payroll, performance and home team discount measure how much of a salary he will be offered/acc

2009-01-12T21:06:00.000-08:00

Question: How does a players ability and the team he signs withs payroll, performance and home team discount measure how much of a salary he will be offered/accept?

Why I asked the question: I was wonder the amount that smaller teams have to pay more than better (more payroll and/or higher winning percentage) teams to get good talent to come to their team.

Analysis: I decided to look at the 2007-2008 free agent class when doing this initial study (will expand as time permits and suggestions are taken into account) The variables for most of the data was pretty strait forward.

Free Agents – I used the ESPN list of free agents, their salary and if they switched teams (home team discount). Only players that were given Major league contracts and that had a VORP from the previous season were used.
Total team salaries – CBSSportsline.com. I looked at the final 2008 value.
VORP – Baseballprospectus.com list both VORP for pitchers and hitters. The players VORP from the previous season were only used. Most free agent decisions are based on what the ball player has done recently.

Note: As you can tell the data was from several different sources, so if there happens to be any inconsistencies or transposition errors, please let me know.

After running a multiple regression for a predicted salary against payroll, VORP, wins over .500 and home team discount I got the following values:

Factor	Salary Amount
Base Salary	108,500
Extra paid per win over 500	12,000
Hometown Discount	-108,000
1 point of VORP above league average	209,000
Amount paid per 10,000,000 in total payroll	233,398

r-squared: 0.622

Standard Deviation of: 3,000,000

This wasn't what I expected. It seems that teams that are winning actually pay more for the same level of talent as teams that win less and those with higher payrolls over pay (i.e. Yankees). This is probably do to the bidding wars that teams get into for the top talent. Also, the home discount of ~1 million is much more of factor than the team's winning percentage that signs the player.

For example let's look at Mark Loretta's contract of 2.75 million. He starts with a base salary of 108,500 and almost gives it all up with his hometown discount of -108,000 since he stayed with Houston. His VORP of 10.5 leads to a salary of 2,000,000. While Houston being 8 games below .500 means they paid 95,000 less, but their payroll of 89 million means he should of another 2.1 million more. His estimated salary would be about 3.2 million or 450,000 less than he is currently making.

As I am a huge fan of charts, here are some of more of the results:

Top ten Overpaid Players:

Name	Team Signed by	Amount over paid
Andruw Jones	LA Dodgers	13,933,330
Torii Hunter	LA Angels	6,542,104
Mariano Rivera	NY Yankees	6,132,745
Francisco Cordero	Cincinnati	5,884,827
Jose Guillen	Kansas City	4,969,959
Eric Gagne	Milwaukee	4,956,344
Jason Jennings	Texas	4,132,468
Alex Rodriguez	NY Yankees	3,760,672
Andy Pettitte	NY Yankees	3,755,589
Kip Wells	Colorado	3,552,833

Top Ten Underpaid Players:

Name	Team Signed by	Amount under paid
Jorge Posada	NY Yankees	5,919,700
Chad Durbin	Philadelphia	4,702,272
LaTroy Hawkins	NY Yankees	4,687,160
Shawn Chacon	Houston	4,367,667
Sean Casey	Boston	4,280,449
David Riske	Milwaukee	3,521,124
Josh Fogg	Cincinnati	3,356,606
Matt Wise	NY Mets	3,310,750
Doug Brocail	Houston	2,923,741
Chad Paronto	Houston	2,868,081

Three players that projected to an actual negative salary (they should have paid the team that signed them to play):

Name	Team	Amount owed to team
Emil Brown	Oakland	-740,500
Jason LaRue	St. Louis	-326,622
Jason Jennings	Texas	-132,468

The amount underpaid (+ values) or overpaid (- values) per position

Position	Amount Underpaid/Overpaid
1B	2,673,270
LF	1,063,523
C	617,522
RP	192,965
SP	79,669
3B	-74,550
2B	-256,537
SS	-385,656
RF	-1,898,907
CF	-4,999,808

I looking at expanding the study to include more years. I have the data for 2007-2008 off season, but salaries and Free Agent signings before that are not as readily available. I would also like to incorporate any other factors that might help explain the data. Finally I might remove the games over .500 from the regression analysis since it has very little effect, but I want to wait to see if any other factors are added.