Any statisticians around? Warning: Math!

[hofautos]

it would be nice to have the author or any member of the WAR team here so we can get some straight answers.

Also interesting is that Fangraphs and Baseball reference publish different WAR values. That makes me think that there is some "tweaking" going on, or has already gone on. Of course the team that doesn't publish disclaimers about margins of errors or inaccuracies may tweak things without version control.

Anyone else here know WAR? Wheres that tribe guy :?:

Please go here to read all about WAR before you talk about it anymore.

Mr apr 2010 9 poster, please tell me who you are?
Will going there tell me why fangraphs and and baseball reference publlish different WAR values? If so, i don't see it...please enlighten me.

What does it matter who I am, when I joined, and how many posts I have? As long as I'm giving you correct information that should suffice.

For the offensive part of the WAR calculation Fangraphs uses wOBA (weighted on-base average) and BR uses Rbat (number of runs better or worse than average the player was as a hitter).

For pitching, Fangraphs uses FIP (fielding independent pitching) to determine RAR and BR uses simple Runs Allowed and adjusts for the quality of the opponent and the team's Total Zone Rating.

Enlightened?

You have been thanked! Pretty good 10 posts, 2 thanks.
I wanted to know who you were, hoping to get some details from you on your experience with WAR......would still like to know.
Yes, thanks, you enlightened me that WAR doesnt always equal WAR...you would think they would have a different name for different values...i guess it is too new.

So are the WAR values intended that you can compare pitchers and hitters using one scale to determine value?
E.g is a pitcher with WAR 10 equal to same value as hitter with WAR 10?

Anyway, thanks, and I apologize if I sounded rude...just a little frustrated at the time.

 

[pigskincardboard]

it would be nice to have the author or any member of the WAR team here so we can get some straight answers.

Also interesting is that Fangraphs and Baseball reference publish different WAR values. That makes me think that there is some "tweaking" going on, or has already gone on. Of course the team that doesn't publish disclaimers about margins of errors or inaccuracies may tweak things without version control.

Anyone else here know WAR? Wheres that tribe guy :?:

Please go here to read all about WAR before you talk about it anymore.

Mr apr 2010 9 poster, please tell me who you are?
Will going there tell me why fangraphs and and baseball reference publlish different WAR values? If so, i don't see it...please enlighten me.

What does it matter who I am, when I joined, and how many posts I have? As long as I'm giving you correct information that should suffice.

For the offensive part of the WAR calculation Fangraphs uses wOBA (weighted on-base average) and BR uses Rbat (number of runs better or worse than average the player was as a hitter).

For pitching, Fangraphs uses FIP (fielding independent pitching) to determine RAR and BR uses simple Runs Allowed and adjusts for the quality of the opponent and the team's Total Zone Rating.

Enlightened?

You have been thanked! Pretty good 10 posts, 2 thanks.
I wanted to know who you were, hoping to get some details from you on your experience with WAR......would still like to know.
Yes, thanks, you enlightened me that WAR doesnt always equal WAR...you would think they would have a different name for different values...i guess it is too new.

So are the WAR values intended that you can compare pitchers and hitters using one scale to determine value?
E.g is a pitcher with WAR 10 equal to same value as hitter with WAR 10?

Anyway, thanks, and I apologize if I sounded rude...just a little frustrated at the time.

Just go by 'The Book.' and then read the various websites. The community is very open about their math and the only time problems occur is when teams snipe people away. At that point, they're not supposed to enlighten you but they'll still post a solid link for you if you happen to ask them a question.

 

[hofautos]

^^^ too much work... I would rather solicit people here...there's a pretty good group here, but if i am unsuccessful, and my desire enthralls me, i may do that

 

[elmalo]

Best way would be to rate by position, not on one scale, but if you feel the need to have one scale, don't use just a 5 year. I commended him, and believe that is a good basis, but added longevity career numbers should be figured in.

The bolded I completely agree with.

I can understand your view on the top list, because it does seem absurd that 15 players from one position could be better than the best player at any given position.

especially when you figure pitchers play only 1/4(on avg) the amount of games as a catcher.

[quote="Anthony K.":3lmnydkz]
I have read how you would do it, I guess I should have asked IF you would do it. Then, we could see your top 50 or top 100 list and could open that up for debate :mrgreen:

I told him i would LOVE to create a formula or two, and asked several times if he would give me his pWAR values...but he won't, and I am too lazy to grab them for the amount of time it would consume.

but a quick and easy method that would be a VAST improvement over his would be to do this:
First seperate pitchers from position players.

Then take pWAR + derivative of career WAR (such that its weight is approx half the weight of pWAR), and then you would have 2 lists
One for hitters and one for pitchers, and those "top lists" would be a dramatic improvement over pWAR.

I could offer much better formulas, but above method would only take 10 minutes once I had the data.
Other formulas (using pWAR 7, and pWAR10, along with career WAR and better compensation for catchers) but that would take more time, and I would be happy to do it, but again, Chevy doesnt want to give me his pWAR values...whatever.[/quote:3lmnydkz]
And pitchers have more of an impact. An ace can be the difference between a team making the playoffs or not, and winning the world series or not. Look at Koufax, look at Gibson, the Dodgers and the Cardinals dont win World Series without them. Johnny Bench was on a teams full of Hall of Famers.

 

[Guest]

This is in response to the people who argue that "five years" is too small a sample size. Here's a comparitive Top 10 (with pitchers and position payers seperated) over 1 season, 5 seasons, 10 seasons, and career.

Only 1901-present considered.

PITCHERS

Single Season WAR (only best season per player considered)
1. Walter Johnson 12.4 (1913)
2. Steve Carlton 12.2 (1972)
3. Bob Gibson 11.9 (1968)
4. Dwight Gooden 11.7 (1985)
5. Cy Young 11.2 (1901)
6. Sandy Koufax 10.8 (1963/'66)
7. Wilbur Wood 10.7 (1971)
T-8. Joe McGinnity 10.5 (1903)
T-8. Gaylord Perry (1972)
T-10. Roger Clemens (1997)
T-10. Dick Ellsworth (1963)

pWAR (5 best WAR seasons, averaged)
1. Walter Johnson 10.6
2. Grover Alexander 9.2
3. Bob Gibson 9.1
T-4. Roger Clemens 8.8
T-4. Lefty Grove 8.8
T-4. Christy Mathewson 8.8
7. Ed Walsh 8.7
T-8. Sandy Koufax 8.6
T-8. Cy Young 8.6
10. Gaylord Perry 8.4

10 best WAR seasons, averaged
1. Walter Johnson 9.2
2. Roger Clemens 8.1
3. Christy Mathewson 7.9
4. Lefty Grove 7.8
5. Pete Alexander 7.7
6. Bob Gibson 7.4
7. Randy Johnson 7.2
T-8. Gaylord Perry 7.1
T-8. Tom Seaver 7.1
10. Phil Niekro 6.9

Career WAR
1. Roger Clemens 128.4
2. Walter Johnson 127.7
3. Tom Seaver 105.3
4. Pete Alexander 104.9
5. Lefty Grove 98.3
T-6. Greg Maddux 96.8
T-6. Phil Niekro 96.8
8. Gaylord Perry 96.3
9. Warren Spahn 93.4
10. Randy Johnson 91.8

POSITION PLAYERS

Single Season WAR (only best season per player considered)
1. Babe Ruth 14.7 (1923)
2. Rogers Hornsby 13.0 (1924)
3. Mickey Mantle 12.9 (1956)
4. Barry Bonds 12.5 (2001)
5. Carl Yastrzemski 12.2 (1967)
T-6. Lou Gehrig 12.0 (1927)
T-6. Joe Morgan 12.0 (1975)
9. Ted Williams 11.8 (1946)
10. Honus Wagner 11.6 (1908)

pWAR (5 best WAR seasons, averaged)
1. Babe Ruth 13.3
2. Barry Bonds 11.7
3. Rogers Hornsby 11.5
4. Mickey Mantle 11.3
5. Ty Cobb 11.2
6. Ted Williams 10.9
T-7. Lou Gehrig 10.5
T-7. Willie Mays 10.5
9. Honus Wagner 10.3
T-10. Eddie Collins 10.2
T-10. Joe Morgan 10.2

10 best WAR seasons, averaged
1. Babe Ruth 12.2
2. Barry Bonds 10.7
3. Rogers Hornsby 10.1
4. Willie Mays 10.0
5. Ty Cobb 9.7
6. Lou Gehrig 9.6
T-7. Honus Wagner 9.3
T-7. Ted Williams 9.3
9. Mickey Mantle 9.1
10. Eddie Collins 8.8

Career WAR
1. Babe Ruth 172.0
2. Barry Bonds 171.8
3. Ty Cobb 159.4
4. Willie Mays 154.7
5. Hank Aaron 141.6
6. Honus Wagner 134.5
7. Tris Speaker 133
T-8. Rogers Hornsby 127.8
T-8. Stan Musial 127.8
10. Eddie Collins 126.7

 

[hofautos]

^^^ It really just depends on what one's objective is.
Randy Johnson was a workhorse. It's nice to see he made the 10year top 10...curiosity asks if he makes a 7year top 10 as well.


And as usual...thanks for the research and data you are doing. It will be nice to get all the pwar values when you are done.

 

[hofautos]

A quick google search will show MANY people have a real problem with WAR for catchers.
My guess is that within a year we will see some different war values on both "BaseballReferrence" and Fangraphs for catchers...
interestingly enough, both sites calculate WAR differently....what we have here is WAR in it's infancy stages.

It will be interesting also to see who's WAR is more widely accepted since both "Baseball Reference" and Fangraphs calculate differently.

It's good stuff though...and I like it.

http://www.google.com/search?q=WAR+catc ... nG=Find+It

 

[Guest]

To those of you who cite longevity as an important measure of a player's 'greatness' I did some research and found out that not many Hall of Famers have actually had the types of careers that match public perception.

"10 years" is what I have heard, not just in this thread, but in the pages of Sports Illustrated, on ESPN and MLB TV, and on sports talk radio. I think that there is a growing consesus that a player "needs" 10 good seasons to be a Hall of Famer. If you define a 'good season' as All-Star Quality (WAR 5.0)+ then only the a handful of current Hall of Famers meet this standard. Please note that this is NOT a 10-year average war. This is simply looking at each year of a player's career and counting the number of WAR 5.0+ seasons. Only members of the HOF were eligible for this list.

P
Alexander, Pete
Blyleven, Bert
Grove, Lefty
Johnson, Walter
Mathewson, Christy
Nichols, Kid
Niekro, Phil
Perry, Gaylord
Seaver, Tom
Spahn, Warren
Young, Cy

1B
Gehrig, Lou

2B
Collins, Eddie
Hornsby, Rogers
Lajoie, Nap
Morgan, Joe

3B
Mathews, Eddie
Schmidt, Mike

SS
Wagner, Honus

LF
Henderson, Rickey
Musial, Stan
Williams, Ted

CF
Cobb, Ty
DiMaggio, Joe
Mantle, Mickey
Mays, Willie
Speaker, Tris

RF
Aaron, Hank
Ott, Mel
Robinson, Frank
Ruth, Babe

I'm sure this does not meet with the public's general perception about the quality players have had over time.

 

[ronfromfresno]

To those of you who cite longevity as an important measure of a player's 'greatness' I did some research and found out that not many Hall of Famers have actually had the types of careers that match public perception.

"10 years" is what I have heard, not just in this thread, but in the pages of Sports Illustrated, on ESPN and MLB TV, and on sports talk radio. I think that there is a growing consesus that a player "needs" 10 good seasons to be a Hall of Famer. If you define a 'good season' as All-Star Quality (WAR 5.0)+ then only the a handful of current Hall of Famers meet this standard. Please note that this is NOT a 10-year average war. This is simply looking at each year of a player's career and counting the number of WAR 5.0+ seasons. Only members of the HOF were eligible for this list.

Now in this age of better conditioning, better travel methods and generally better health, yes ten years seems to be the standard. For the history of the game it would be different, but that's the beauty of the voting, the voters can see past the numbers and understand what a player did in his time.

 

[Guest]

To those of you who cite longevity as an important measure of a player's 'greatness' I did some research and found out that not many Hall of Famers have actually had the types of careers that match public perception.

"10 years" is what I have heard, not just in this thread, but in the pages of Sports Illustrated, on ESPN and MLB TV, and on sports talk radio. I think that there is a growing consesus that a player "needs" 10 good seasons to be a Hall of Famer. If you define a 'good season' as All-Star Quality (WAR 5.0)+ then only the a handful of current Hall of Famers meet this standard. Please note that this is NOT a 10-year average war. This is simply looking at each year of a player's career and counting the number of WAR 5.0+ seasons. Only members of the HOF were eligible for this list.

Now in this age of better conditioning, better travel methods and generally better health, yes ten years seems to be the standard. For the history of the game it would be different, but that's the beauty of the voting, the voters can see past the numbers and understand what a player did in his time.

Well, unfortunately for me I cannot simply turn over my bank account to the whims of a bunch of sports writers and retired players with agendas who have shaped (or mis-shaped) the Hall of Fame over the past 71 years.

I need to find a lithmus test of sorts. One thing I can look at to count/discount players.

This may sound 'scarey' to some, but I for one would feel like a giant monkey was off my back if I could find a statistical scandal that generated a list of 'great' players that I could then automatically correct without the worry of bias.

I said an average of 5 MVP Quality seasons (with an average 10 All-Star Quality seasons) was a fine basis for a list. However, I was taken back by the level of venem, and resistance people have had in my attempt to quantify these players. I believe if I had say "These players were 5x MVP, 10x All-Stars" people would have been a lot more receptive, but because I injected WAR it brought out the anti-WAR people.

So I ask you, ron. Do you have faith in WAR?

If so, are there a certain combination of averaged MVP and All-Star seasons you would consider make a player 'great'?

If you dislike averages, are there a certain number of counted MVP and All-Stars you would consider make a player 'great'?

Here are my problems. I genuinely dislike career statistics because they discount players who had shortened careers for reasons beyond their control. And I dislike actual MVP voting/All-Star voting/HOF voting because I find voting of any kind problematic.

 

[elmalo]

[quote="Chris Levy":5j9g9llt]To those of you who cite longevity as an important measure of a player's 'greatness' I did some research and found out that not many Hall of Famers have actually had the types of careers that match public perception.

"10 years" is what I have heard, not just in this thread, but in the pages of Sports Illustrated, on ESPN and MLB TV, and on sports talk radio. I think that there is a growing consesus that a player "needs" 10 good seasons to be a Hall of Famer. If you define a 'good season' as All-Star Quality (WAR 5.0)+ then only the a handful of current Hall of Famers meet this standard. Please note that this is NOT a 10-year average war. This is simply looking at each year of a player's career and counting the number of WAR 5.0+ seasons. Only members of the HOF were eligible for this list.

Now in this age of better conditioning, better travel methods and generally better health, yes ten years seems to be the standard. For the history of the game it would be different, but that's the beauty of the voting, the voters can see past the numbers and understand what a player did in his time.

Well, unfortunately for me I cannot simply turn over my bank account to the whims of a bunch of sports writers and retired players with agendas who have shaped (or mis-shaped) the Hall of Fame over the past 71 years.

I need to find a lithmus test of sorts. One thing I can look at to count/discount players.

This may sound 'scarey' to some, but I for one would feel like a giant monkey was off my back if I could find a statistical scandal that generated a list of 'great' players that I could then automatically correct without the worry of bias.

I said an average of 5 MVP Quality seasons (with an average 10 All-Star Quality seasons) was a fine basis for a list. However, I was taken back by the level of venem, and resistance people have had in my attempt to quantify these players. I believe if I had say "These players were 5x MVP, 10x All-Stars" people would have been a lot more receptive, but because I injected WAR it brought out the anti-WAR people.

So I ask you, ron. Do you have faith in WAR?

If so, are there a certain combination of averaged MVP and All-Star seasons you would consider make a player 'great'?

If you dislike averages, are there a certain number of counted MVP and All-Stars you would consider make a player 'great'?

Here are my problems. I genuinely dislike career statistics because they discount players who had shortened careers for reasons beyond their control. And I dislike actual MVP voting/All-Star voting/HOF voting because I find voting of any kind problematic.[/quote:5j9g9llt]
What do you think of Alberte Belle?

 

[Guest]

What do you think of Alberte Belle?

Me, personally? I don't think anything about Albert Belle. Just like I don't think much of any player. Organized professional baseball has been played for 140 years. When you sit and think about the number of players who have come and gone over the years it's simply overwhelming.

Perhaps you can look over a player and quickly discern if he was 'good' or 'great,' but I can not. I know this is going to open me to personal attacks, but there are literally nights when I lay awake, restless worrying about how to figure out who was good enough to be in my collection.

I admire the player collectors, team collectors, and set collectors. Their ability to stick to goals without getting distracted is worthy of my admiration. I tried building a set, but they all started to look the same to me after a while, and once I had the 'stars' it was tough finding the energy to chase down the 'commons'. I tried player collecting Mel Ott, but after buying a very clean cut, a tough patch card, several low-end relics, and a vintage example I felt as if I had too many cards of the same player and quickly lost steam. I tried collecting the 1950s Dodgers, the team I grew up listening to stories of my grandmother, mother, and uncle talk about, but I quickly added most of the team to my collection only to find out as I began following statistics that the players (other than the obvious Robinson, Snider, Reese, and Campanella) weren't nearly as good as family memories.

I tried collecting Hall of Famers next, but there are almost 300 of them, and it seemed like an unenviable task. To motivate myself I read stories about the players, tried to pick out onces that seemed interesting. I tried to start with ones who were local at the time (the NY teams) but found out that a good number of them (Ross Youngs, George "High Pockets" Kelly, Dave Bancroft, Freddie Lindstrom, and Travis Jackson) were not elected for any great on the field performance. They were elected in large part due to their teammate Frankie Frisch, who used the Veteran's Committee as a tool for getting friends in. Once I looked behind the curtain and saw how the process actually worked, I realized that the HOF was 'flawed' and that I'd never be happy collecting those players. So I needed to find a way to determine an arbitrary way of deciding, which brings us to this thread.

Because after all this is a baseball card forum, and essentially I developed pWAR first and foremost as a way of giving me an arbitrary way of letting a statistic dictate who would (and would not) be in my collection using the push of a button. No more thinking about. No more worrying about it. No more trying to weigh home runs, hits, RBI, SB, etc. I could simply look at one stat and in less than seconds know with certainly whether or not that player would end up in my collection.

I was remarkably at piece with that, and immediately went out and spent a considerable sum based on the results. However, I am still shocked at the venom and resistance this has received. I never simply thought that by averaging a player's 5 best WAR seasons, and sorting them in order from highest to lowest would make someone's head explode, but it seems as if it has. Personal attacks. Name calling. Trolling. Questioning my formal education. Questioning the merits of the formula, when essentially it's just an average of X years. I suppose most of these people just don't want what I want, a simple "I win" button I can push that immediately settles the debate as to a player's worth. Baseball has always had 'fun' arguments about who was the best since the days of Cap Anson and King Kelly. The problem is century-long arguments don't help me add cards to my collection stress-free.

Back to Albert Belle. I trust in WAR. And WAR tells me that Albert Belle had 3 All-Star Quality seasons. So now, whenever I think of Albert Belle I think "3 time All-Star." You may say "that's too simple. the stat is flawed. he was more than that." Maybe. Maybe not. But if I sit here thinking and worrying about Albert Belle, then I have to sit here thinking and worrying about everyone ... and that way madnless lies. Since this thread is about pWAR, I will also tell you that his pWAR is 5.9. Which basically means that his 3 AS seasons were high enough that when averaged with his two best seasons as a Starter, he averaged AS performance over the five year span.

 

[hofautos]

^^ I appreciate what you shared here, as I have had the exact same dilemma, trying to define my collection. I have changed my collection so many times. And if I were to start all over again, I would do "very similar" to you. I would only want the best card of the greatest players, within my budget (of which you profess to have none), and to me the best card is the rookie card. Rookies are not a necessity for you.

If you can't sleep deciding who to collect, How do you decide which card you buy?

Also my magic button would not be pwar, it would be a modified pwar, as I only want the greatest, and the greatest cannot be calculated by 5 years only.

You will always have critics when you try to sell the concept of your magic button, when better magic buttons can be had.
enjoy your "magic button" and let others enjoy their own.

Again, i cant say enough about my appreciation for pWAR, and i thank you for it..after all my button would be a modification of it...

I suspect our collections would be very similar, except for that i would have players with the likes of johnny bench and pete rose in place of your gaylord perry and robin roberts, because your button is based on only 5 years and has no respect for the catcher position or MLB records, where mine will.

 

[hofautos]

^^^ I also appreciate you taking the time to explain WAR to me a little better. After thinking about what you told me, without even looking up his name, I guessed Mariano Rivera would have very little value based on my better understanding...and sure enough, i was right.

According your definition of great, Mariano has not had even one great season....not even one good season, and he has never really had any good trade value either.

I read an article where they talked about giving Mariano a "LI" levaraging index....to take his 2.4 to a 4.8...and that was by some staff at fangraphs.


I like WAR, but only if used in certain ways.

 

[hofautos]

Here is a decent scale that includes extra weight for peak performance

http://www.beyondtheboxscore.com/2010/1 ... eplacement