Any statisticians around? Warning: Math!

[ronfromfresno]

But even though he played better than others in the league in 2000 he produced less for his team meaning in the real world he cost his team runs, hits, and WINS compared to 2005.

What I would love to see is a break down of all players with at least 7 years in the league for a single position and see how many non-HOF players shoot up the list because they had 5 great season and 15 terrible. I'm willing to bet there are a couple out there and that casts serious doubt on using the 5 year model for greatness and points to a need to take longevity and greatness over time into account.

Using my five year model I have found a total of 2 retired players outside the Hall of Fame who during that span averaged a WAR of 8.0 or higher.

Shoeless Joe Jackson. Who is inelligible for Hall of Fame induction as a result from his banishment from organized, professional play. However, most believe that he was a great player, and had he not been implicated in the 1919 World Series 'fix' he would have been enshrined in Cooperstown.

Ron Santo. Whom many believe is the Hall of Fame's greatest snub, and many argue is the strongest candidate on the current veteran's committee ballot.

As you can see, even using a sample size of five years, which you consider to be "small," only two retired players not in the Hall of Fame have been indentfied. Strong HOF arguments could be made for both players.

Only two out of the thousands of players to play the game, or only two out of the players you have choosen? What about guys like Larry Doyle who lead the league in many categories and lead all 2B for most of his career. And what about players that are border line with slightly lower WAR. Good players can have 5 great years, Will Clark did, Frank Thomas did many pitchers have Mike Hampton comes to mind, but they didn't have the longevity of greatness to become HOFer's yet. I'm still willing to bet if you used this model for all players, all time you'd come up with more than a few guys who shouldn't be considered for the Hall but match the criteria you selected for your adjustments on WAR to determine greatness in your eyes. Now my question is if a non-HOF worhty player pops up on your search because of the 5 year sample size does this make the player great even though the rest of his long career was at or below average? Wouldn't that show your sample size is too small, if and when the method is published these are the types of criticisms that will be investigated.

 

[Guest]

I did not create WAR. Unfortunately, since the owner of Baseball Projection that was the leading proponent of WAR, is now an employee of an undisclosed MLB franchise ... he no longer interacts with us hobbyists and can no longer explain/defend WAR.

To me WAR is a fine measurement of a player's value during a specific season. I've seen arguments made about rankings given to Pujols' respective seasons compared to league values for those seasons. However, I've yet to see anyone point out a specific WAR season that is miscategorized.

On the first post of this thread I list 52 (post-1901) players that averaged a WAR of 8.0 over a span of five seasons. Is there anyone on that list who you could argue statistics show was not a great player?

 

[Guest]

But even though he played better than others in the league in 2000 he produced less for his team meaning in the real world he cost his team runs, hits, and WINS compared to 2005.

What I would love to see is a break down of all players with at least 7 years in the league for a single position and see how many non-HOF players shoot up the list because they had 5 great season and 15 terrible. I'm willing to bet there are a couple out there and that casts serious doubt on using the 5 year model for greatness and points to a need to take longevity and greatness over time into account.

Using my five year model I have found a total of 2 retired players outside the Hall of Fame who during that span averaged a WAR of 8.0 or higher.

Shoeless Joe Jackson. Who is inelligible for Hall of Fame induction as a result from his banishment from organized, professional play. However, most believe that he was a great player, and had he not been implicated in the 1919 World Series 'fix' he would have been enshrined in Cooperstown.

Ron Santo. Whom many believe is the Hall of Fame's greatest snub, and many argue is the strongest candidate on the current veteran's committee ballot.

As you can see, even using a sample size of five years, which you consider to be "small," only two retired players not in the Hall of Fame have been indentfied. Strong HOF arguments could be made for both players.

Only two out of the thousands of players to play the game, or only two out of the players you have choosen? What about guys like Larry Doyle who lead the league in many categories and lead all 2B for most of his career. And what about players that are border line with slightly lower WAR. Good players can have 5 great years, Will Clark did, Frank Thomas did many pitchers have Mike Hampton comes to mind, but they didn't have the longevity of greatness to become HOFer's yet. I'm still willing to bet if you used this model for all players, all time you'd come up with more than a few guys who shouldn't be considered for the Hall but match the criteria you selected for your adjustments on WAR to determine greatness in your eyes. Now my question is if a non-HOF worhty player pops up on your search because of the 5 year sample size does this make the player great even though the rest of his long career was at or below average? Wouldn't that show your sample size is too small, if and when the method is published these are the types of criticisms that will be investigated.

As of this post I have processed all members of the HOF, and every player who recieved at least 5% of the votes in a given year of Hall of Fame balloting. I am willing to enter any players by request into this formula. I will post below. I will enter the players you suggested.

 

[Guest]

Players currently eligible (retired 5+ years). Players who failed to rate a pWAR of 5.0+ were not tested beyond the initial sample.

Pitchers
Wood, Wilbur CHW 7.3 6.0 4.5
Cicotte, Eddie CHW 6.6 5.5 4.6
Rucker, Nap BRO 6.6 5.7 4.2
Blue, Vida OAK 6.3 5.5 4.5
Tiant, Luis BOS 6.3 5.8 5.1
Hershiser, Orel LAD 6.1 5.3 4.5
Lolich, Mickey DET 5.9 5.3 4.5
Adams, Babe PIT 5.9 5.2 4.6
Gooden, Dwight NYM 5.8 5.1 4.4
Guidry, Ron NYY 5.6 5.1 4.3
Harder, Mel CLE 5.6 5.1 4.1
Walters, Bucky CIN 5.5 4.7 3.8
Kaat, Jim MIN 5.5 4.8 4.0
Newsom, Bobo DET 5.4 4.8 4.3
Warneke, Lon CHC 5.3 4.8 3.9
Luque, Dolf CIN 5.3 4.6 3.7
Valenzuela, F. LAD 5.3 4.7 3.8
Bridges, Tommy DET 5.2 4.9 4.5
John, Tommy CHW 5.1 4.7 4.1
Scott, Mike 4.8
Maglie, Sal 4.8
Wood, Smokey Joe 4.8
Sain, Johnny 4.7
Morris, Jack 4.7
Newcombe, Don 4.6
Rowe, Schoolboy 4.5
Richard, J.R. 4.5
Derringer, Paul 4.4
Schumacher, Hal 4.2
Reynolds, Allie 4.0
Burdette, Lew 3.9
Hutchinson, Fred 3.8
Parker, Dave 3.7
Vander Meer, J. 3.7
Raschi, Vic 3.5
Fitzsimmons, F. 3.3
Altrock, Nick 2.1
Larsen, Don 2.0
Kerr, Dickey 1.6

The only pitcher with a pWAR of 5.0+ who maintained it for 10 seasons was Luis Tiant.

Catchers
Torre, Joe ATL 5.8 5.3 4.5
Munson, Thurman NYY 5.6 5.0 4.3
Simmons, Ted STL 5.6 5.1 4.6
Howard, Elston 4.9
Parrish Lance 4.1
Cooper, Walker 3.8
Kling, Johnny 3.7
Boone, Bob 3.0
Ruel, Muddy 2.6
Gowdy, Hank 2.1
Wilson, Jimmie 1.5
Criger, Lou 1.4

First Basemen
Bagwell, Jeff HOU 7.9 7.2 6.6
Allen, Dick PHI 7.6 6.9 5.9
McGwire, Mark OAK 6.3 6.1 5.3
Hernandez, Keith STL 6.2 5.8 5.3
Clark, Will SFG 6.1 5.3 4.6
Palmeiro, Rafael TEX 6.0 5.5 5.0
Mattingly, Don NYY 5.8 5.0 4.1
McGriff, Fred TOR 5.5 5.0 4.3
Hodges, Gil LAD 5.3 4.8 4.3
Kluszewski, Ted CIN 5.1 4.1 3.1
Garvey, Steve 4.5
Vaughn, Mo 4.4
Cavarretta, Phil 4.3
Vernon, Mickey 4.2
McInnis, Stuffy 4.1
York, Rudy 4.1
Judge, Joe 3.7
Chase, Hal 2.5
Grimm, Charlie 2.2

The only first basemen with a pWAR of 5.0+ who maintained it for 10 seasons were Jeff Bagwell, Dick Allen, Mark McGwire, Keith Hernandez, and Rafael Palmeiro.

Second Basemen
Doyle, Larry 4.8

Third Basemen
Santo, Ron CHC 8.0 7.3 6.3
Boyer, Ken STL 6.3 6.0 5.4
Nettles, Graig NYY 5.9 5.5 5.0
Evans, Darrell ATL 5.8 4.2 4.6
Williams, Matt 4.8
Madlock, Bill 4.6
*****, Jimmy 2.8

The only third basemen with a pWAR of 5.0+ who maintained it for 10 seasons were Ron Santo, Ken Boyer, and Graig Nettles.

Shortstops
Trammell, Alan DET 6.7 6.3 5.6
Larkin, Barry CIN 6.2 6.0 5.5
Dark, Al 4.7
Groat, Dick 4.4
Wills, Maury 4.3
Wright, Glenn 4.1
Concepcion, Dave 4.1
Marion, Marty 3.8
Crosetti, Frankie 3.1

The only shortstops with a pWAR of 5.0+ who maintained it for 10 seasons were Alan Trammell and Barry Larkin.

Left Fielders
Jackson, Joe CLE 8.2 7.6 6.2
Rose, Pete CIN 6.8 6.4 5.9
Minoso, Minnie CHW 6.3 5.8 5.1
Keller, Charlie NYY 6.3 5.5 4.3
Raines, Tim Mon 6.3 5.8 5.1
Belle, Albert CLE 5.9 4.9 3.8
Foster, George CIN 5.6 5.1 4.1
O'Neill, Tip STL 5.1 4.1 3.1
O'Doul, Lefty 4.7
Kuenn, Harvey 3.6
Carter, Joe 3.5
Meusel, Bob 3.3
Lewis, Duffy 3.0
Jackson, Bo 1.7

The only left fielders with a pWAR of 5.0+ who maintained it for 10 seasons were Shoeless Joe Jackson, Pete Rose, Minnie Minoso, and Tim Raines.

Center Fielders
Murphy, Dale ATL 6.7 5.9 4.6
Pinson, Vada CIN 5.9 5.3 4.6
Lynn, Fred BOS 5.7 5.2 4.3
Flood, Curt 4.5
Van Haltren, G. 4.2
McGee, Willie 4.2
DiMaggio, Dom 4.2
Davis, Chili 4.0
Oliver, Al 4.0
Martin, Pepper 3.4
Williams, Cy 3.3
Moore, Terry 3.0
Cramer, Doc 1.6

Right Fielders
Walker, Larry COL 6.6 6.0 5.3
Parker, Dave PIT 6.2 5.3 4.0
Oliva, Tony MIN 6.1 5.5 4.3
Bonds, Bobby SFG 6.0 5.7 5.2
Evans, Dwight BOS 5.5 5.2 4.6
Staub, Rusty Mon 5.4 4.8 4.0
Maris, Roger NYY 5.4 4.8 3.8
Clark, Jack SFG 5.1 4.7 4.3
Canseco, Jose OAK 5.0 4.3 3.7
Henrich, Tommy 4.8
Herman, Babe 4.7
Gonzalez, Juan 4.6
Mondesi, Raul 4.0
Bauer, Hank 3.7
Southworth, Billy 3.0
Stengel, Casey 2.7

The only right fielders with a pWAR of 5.0+ who maintained it for 10 seasons were Larry Walker and Bobby Bonds.

Designated Hitters
Martinez, Edgar SEA 6.4 6.2 5.8
Baines, Harold 3.3

The only designated hitter with a pWAR of 5.0+ who maintained it for 10 seasons was Edgar Martinez.

 

[ronfromfresno]

Players currently eligible (retired 5+ years). Players who failed to rate a pWAR of 5.0+ were not tested beyond the initial sample.


Second Basemen
Doyle, Larry 4.8

So explain the theory behind Larry Doyle he was the MVP in 1912 and third in voting in 1911 so in the eyes of his peers he has at least two MVP type seasons. He also had the best offensive WAR in 1915 when he lead the league in BA, 2B and hits. Add into that second in total bases three times and top 5 in HR for 5 seasons. Top four in getting on base in 6 different seasons and top 4 in extra base hits in 6 different seasons. He wasn't amazing on defense, he made a ton of errors, but he was in the top 4 in assists, put outs, and double plays for most of his career. He played 14 season, by no means a compiler and his .408 career slugging average was the best by an NL second baseman when he retired, as were his career totals in hits 1887, doubles 299, triples 123, total bases 2654 and extra base hits 496. So he was amoung the best 2B of his era in several categories but you have him listed as less than an All-Star compared to his peers and history. How can that be when he was the best or amoung the best for at least 5 years during his career if not more?

 

[Guest]

Players currently eligible (retired 5+ years). Players who failed to rate a pWAR of 5.0+ were not tested beyond the initial sample.


Second Basemen
Doyle, Larry 4.8

So explain the theory behind Larry Doyle he was the MVP in 1912 and third in voting in 1911 so in the eyes of his peers he has at least two MVP type seasons. He also had the best offensive WAR in 1915 when he lead the league in BA, 2B and hits. Add into that second in total bases three times and top 5 in HR for 5 seasons. Top four in getting on base in 6 different seasons and top 4 in extra base hits in 6 different seasons. He wasn't amazing on defense, he made a ton of errors, but he was in the top 4 in assists, put outs, and double plays for most of his career. He played 14 season, by no means a compiler and his .408 career slugging average was the best by an NL second baseman when he retired, as were his career totals in hits 1887, doubles 299, triples 123, total bases 2654 and extra base hits 496. So he was amoung the best 2B of his era in several categories but you have him listed as less than an All-Star compared to his peers and history. How can that be when he was the best or amoung the best for at least 5 years during his career if not more?

A large focus of my collection is the New York Giants. Larry Doyle was clearly an important part of the Giants teams of that era, but he was playing against two of the greatest players in the history of the game at his position: Eddie Collins and Nap Lajoie. Johnny Evers was also putting up competing numbers on the Cubs, the Giants great rival of the era.

Doyle was a good player. Not a great player. And WAR reflects this.

As for MVP voting? Any voting of any kind (MVP, All-Star, People's Choice awards, political elections, etc.) is flawed. The 1912 NL MVP should have been Honus Wagner.

And I think this is where the fundamental difference is. Your arguments are with statistics themselves, not just my statistic. I would be perfectly content to strip voting of all MLB awards, and simply let a computer award them and select all-star teams.

 

[ronfromfresno]

[quote="Chris Levy":12vbz5hu]Players currently eligible (retired 5+ years). Players who failed to rate a pWAR of 5.0+ were not tested beyond the initial sample.


Second Basemen
Doyle, Larry 4.8

So explain the theory behind Larry Doyle he was the MVP in 1912 and third in voting in 1911 so in the eyes of his peers he has at least two MVP type seasons. He also had the best offensive WAR in 1915 when he lead the league in BA, 2B and hits. Add into that second in total bases three times and top 5 in HR for 5 seasons. Top four in getting on base in 6 different seasons and top 4 in extra base hits in 6 different seasons. He wasn't amazing on defense, he made a ton of errors, but he was in the top 4 in assists, put outs, and double plays for most of his career. He played 14 season, by no means a compiler and his .408 career slugging average was the best by an NL second baseman when he retired, as were his career totals in hits 1887, doubles 299, triples 123, total bases 2654 and extra base hits 496. So he was amoung the best 2B of his era in several categories but you have him listed as less than an All-Star compared to his peers and history. How can that be when he was the best or amoung the best for at least 5 years during his career if not more?

A large focus of my collection is the New York Giants. Larry Doyle was clearly an important part of the Giants teams of that era, but he was playing against two of the greatest players in the history of the game at his position: Eddie Collins and Nap Lajoie. Johnny Evers was also putting up competing numbers on the Cubs, the Giants great rival of the era.

Doyle was a good player. Not a great player. And WAR reflects this.

As for MVP voting? Any voting of any kind (MVP, All-Star, People's Choice awards, political elections, etc.) is flawed. The 1912 NL MVP should have been Honus Wagner.

And I think this is where the fundamental difference is. Your arguments are with statistics themselves, not just my statistic. I would be perfectly content to strip voting of all MLB awards, and simply let a computer award them and select all-star teams.[/quote:12vbz5hu]

I have no problems with the statistics, I agree Doyle wasn't a superstar over the course of his career, but he might have been the biggest part of the Giants offense from 1911-1915, a five year span where he was also amoung the league leaders and up to that point all-time 2B leaders in production. At the very least an all-star caliber player, if not an MVP type player, but pWAR has him as slightly less than an All-Star and more like an everyday guy, which simply isn't the case.

 

[Guest]

I have no problems with the statistics, I agree Doyle wasn't a superstar over the course of his career, but he might have been the biggest part of the Giants offense from 1911-1915, a five year span where he was also amoung the league leaders and up to that point all-time 2B leaders in production. At the very least an all-star caliber player, if not an MVP type player, but pWAR has him as slightly less than an All-Star and more like an everyday guy, which simply isn't the case.

In terms of WAR, Art Fletcher and George Burns were found to be more valuable in terms of WAR for the pre-1920 Giants teams. Doyle had two seasons of WAR 5.0+ (All-Star), so I feel pWAR accurately rated him at 4.8.

I did not develop WAR. pWAR is just a way of interpreting WAR using a sample size other than one season or an entire career.

I think Doyle's rating is an accurate performance over the best five years of his career. Five years which you feel were great, but WAR has determined to be a mix of both good and above average.

 

[ronfromfresno]

I have no problems with the statistics, I agree Doyle wasn't a superstar over the course of his career, but he might have been the biggest part of the Giants offense from 1911-1915, a five year span where he was also amoung the league leaders and up to that point all-time 2B leaders in production. At the very least an all-star caliber player, if not an MVP type player, but pWAR has him as slightly less than an All-Star and more like an everyday guy, which simply isn't the case.

In terms of WAR, Art Fletcher and George Burns were found to be more valuable in terms of WAR for the pre-1920 Giants teams. Doyle had two seasons of WAR 5.0+ (All-Star), so I feel pWAR accurately rated him at 4.8.

I did not develop WAR. pWAR is just a way of interpreting WAR using a sample size other than one season or an entire career.

I think Doyle's rating is an accurate performance over the best five years of his career. Five years which you feel were great, but WAR has determined to be a mix of both good and above average.

To each their own, he was named the MVP and top 5 in several offensive categories over that time span, not top five at his position, top five in all of baseball but because others at his position were all-time greats he is devalued. Fletcher never lead the league in any category except HBP and never finished higher than 12th in MVP voting. Burns was a great run scorer and SB king and I'd say he was very comparable to Doyle but they played different positions. How would Doyle fair compared to SS of the era.

 

[Guest]

I have no problems with the statistics, I agree Doyle wasn't a superstar over the course of his career, but he might have been the biggest part of the Giants offense from 1911-1915, a five year span where he was also amoung the league leaders and up to that point all-time 2B leaders in production. At the very least an all-star caliber player, if not an MVP type player, but pWAR has him as slightly less than an All-Star and more like an everyday guy, which simply isn't the case.

In terms of WAR, Art Fletcher and George Burns were found to be more valuable in terms of WAR for the pre-1920 Giants teams. Doyle had two seasons of WAR 5.0+ (All-Star), so I feel pWAR accurately rated him at 4.8.

I did not develop WAR. pWAR is just a way of interpreting WAR using a sample size other than one season or an entire career.

I think Doyle's rating is an accurate performance over the best five years of his career. Five years which you feel were great, but WAR has determined to be a mix of both good and above average.

To each their own, he was named the MVP and top 5 in several offensive categories over that time span, not top five at his position, top five in all of baseball but because others at his position were all-time greats he is devalued. Fletcher never lead the league in any category except HBP and never finished higher than 12th in MVP voting. Burns was a great run scorer and SB king and I'd say he was very comparable to Doyle but they played different positions. How would Doyle fair compared to SS of the era.

Him being named MVP is irrelevant. He played in the media capitol of the country, with the largest grouping of sports reporters. He was reportedly a very well-liked fellow. It's easy to see how "one of the boys" could simply be given a MVP trophy as a jesture.

1910-'12, '15, and '19 were the WAR seasons used to generate the pWAR of 4.8. Two of those seasons were determined by WAR to be All-Star level, while the remaining three were determined to be Starter level.

Here is what killed Doyle. If you take out defense, Doyle's p(oWAR) is 5.2 (All-Star). The terrible defensive statistics he posted during that same span lower his WAR below the 5 threshold.

 

[hofautos]

Although Gehrig was great on the field, Much of Gehrig's "greatness" is measured by his character

These are the types of statements that drive me absolutely bonkers. Gehrig is an all time great because of what he did on the field. Lou gehrig was a man of great character but no one would know who he was if it wasnt for the fact he excelled at the game of baseball and has the numbers to prove it. If he had the iron man record but was a .250 hitter who hit 15 homers a year we would not talk about it in the same regard.


If "much" of a mas greatness was his character, I can think of a few players no one would ever talk about.

Did I say "how much" ones character defines their greatness. I just said it is in the formula...at least in determining HOF'ers...like it or not, it's a fact...bonk on.

 

[hofautos]

Bash Bash BASH!!!

I hope you didnt' include me in the group that you suggested is calling him names and bashing him. I even defended him once.
I appreciate his efforts. I like what he is doing, and I think he shed a lot of light to many. I don't agree 100% with his formula as defining one's greatness, but do feel it is a good step in the right direction. I hope Chevy feels the same way...

 

[hofautos]

I like to collect Record Holders, those that are considered best at their position, and Fan Favorites....besides, they will hold their value better.

But what happens when a record is broken? You sell off your collection of the player and chase the new guy?

.

Well most of the major records I don't feel will be broken, or at least no more frequently then someone's WAR stats are beaten by someone else...so what's the difference?

 

[hofautos]

I like to collect Record Holders, those that are considered best at their position, and Fan Favorites....besides, they will hold their value better.

But what happens when a record is broken? You sell off your collection of the player and chase the new guy?

I look at it like this. The MVP trophy as we have come to know it was absent from the first 60 years of the game. The Cy Young award came even later. Both were also voted on in a subjective manner. The same can be said about All-Star appearances.

So I throw out MVP awards, Cy Young awards, and All-Star appearances as a way of measuring a player's 'greatness.'

Instead, I credit a player with a WAR 8.0+ MVP as a MVP/Cy Young winner, and a WAR of 5.0+ MVP as an All-Star. This is what the developer of the formula intended.

My pWAR formula states X amount of MVP seasons = "great."

The question is, how many MVP seasons does it take for public opinion to believe someone is great? I propose five.

I have no problem with that logic....I have a problem when it is suggested that you can accurately state how great someone is based by their war numbers.

e.g. everyone with 5+ 8.0 WAR numbers are great players, but player with war value 9 is not necessarily greater than player with war value 8.8

 

[hofautos]

Just out of curiosity, is this based on top 10 seasons of all HOFERS or only players with 10 pWAR 5.0+ (All-Star) seasons?
e.g. if someone had 9 pWAR 5.0+ (All-Star) seasons, were they not considered?

Well here's the thing.

If you take a player's five best seasons and he has a pWAR of 5.7, and then you expand the data to seven years, you're adding two lesser seasons, and with ten you're adding five lesser seasons. Therefore the 7 year WAR will drop from the 5 year one, and the 10 year WAR will drop from the 7 year one. Since you're only adding lesser seasons, the number cannot increase.

The player's on that list averaged a WAR of 7.0 (or higher) over their best ten seasons.

I meant, on your 10 year list, if a player has these 10 pwar values would he be considered:

10.0
10.0
10.0
10.0

4.8
4.8
4.8
4.8
4.8
4.8

I mean did a player HAVE to have x amount of PWAR+8 or x amount of PWAR+5 to be considered on the 10 year list you made?

 

[George_Calfas]

I like to collect Record Holders, those that are considered best at their position, and Fan Favorites....besides, they will hold their value better.

But what happens when a record is broken? You sell off your collection of the player and chase the new guy?

I look at it like this. The MVP trophy as we have come to know it was absent from the first 60 years of the game. The Cy Young award came even later. Both were also voted on in a subjective manner. The same can be said about All-Star appearances.

So I throw out MVP awards, Cy Young awards, and All-Star appearances as a way of measuring a player's 'greatness.'

Instead, I credit a player with a WAR 8.0+ MVP as a MVP/Cy Young winner, and a WAR of 5.0+ MVP as an All-Star. This is what the developer of the formula intended.

My pWAR formula states X amount of MVP seasons = "great."

The question is, how many MVP seasons does it take for public opinion to believe someone is great? I propose five.

I have no problem with that logic....I have a problem when it is suggested that you can accurately state how great someone is based by their war numbers.

e.g. everyone with 5+ 8.0 WAR numbers are great players, but player with war value 9 is not necessarily greater than player with war value 8.8

This is one of the major problems with a mathematical antilogarithm such as this. The variables are of an arbitrary selection. I have showed this formula to several of the Stats professors here at the University of Illinois and the main comment was that the selective data sets used (and not full career numbers) fail to gain statistical confidence.

 

[hofautos]

Thanks for explaining, it does clear some things up for me...but i understood most of what you said in the thread already, but it's nice to see here summarized.
I agree, I am not sold on either. You don't have all the data to calculate players with 4 WAR 8.0 + 7 5.0+ seasons.
I would agree you can define the best players as a combination of both, but certainly WAR has its problems with weights, especially since the formula can't even define 1 great catcher.
The ultimate formula will be a combination of more than just WAR, and I still don't think you can compare players of different positions.
Any formula that would suggest that Gaylord Perry is the 35th Greatest player ever, and Johnny Bench doesn't make top 100, is flawed.
Keep trying, and create your own formula, using WAR.

To define BEST, Maybe .8(5 year WAR) + .4(7 year WAR) + .2(10 year WAR) + bonus points awarded for different things like position (somehow you need it to define great catchers, right?)

To define GREATEST, add bonus points for things like Accomplishments and Character. (e.g., Nolan Ryan was far from one of the Best, but imho, is one of the greatest...longevity, character, and records count in my book, maybe not yours)

Well, if you look at catchers, historically they have not performed on the level of other position players because of the enormous toll the position takes on their bodies.

Please tell me why you think (using statistics) Gaylord Perry is not a 'great' player.

Also, I have never published a list of Top 100 players, so you cannot say that Johnny Bench is off (or on) such list. He is not on my list of Top 50, no.

WAR is already adjusted for position. So if you take a player's stats and the only thing you change is his position, his WAR will change.

You can't define something like character. Ty Cobb may have been a murderer. Babe Ruth was a drunk who may have been involved with women we now consider "underage." Cap Anson is almost single-handedly responsible for segregation in baseball in the nineteenth century. Jackie Robinson turned his back on the game to become a "race man." Or how about all the reports we have of Willie Mays snubbing people and being a "jerk"? As for Nolan Ryan, you say he has 'character' but while he may have 'true grit' that some admire, others think qualities such as that are condemnable, not laudible.

I did not say that Gaylord Perry is not great, nor did I infer that. I meant by comparison. IMHO, you cannot consider GAYLORD #35 and BENCH not make the list at all. In EVERY TOP 50 LIST I have ever seen Johnny Bench was on it, and yet you suggest there is not even 1 great catcher, yet the war adds and subtracts points based on position....that is reason for me by itself to suggest WAR needs "tweaking".

 

[Guest]

Did I say "how much" ones character defines their greatness. I just said it is in the formula...at least in determining HOF'ers...like it or not, it's a fact...bonk on.

You can't put a mathmatical value on character. What would you propose? +3 for Clemente dying in a plane crash doing missionary work? -0.5 for Mays for not shaking hands at autograph signing? -1 for Ruth for being a fall down drunk? -3 for Anson for crusaging to keep African-Americans out of the National League? It's proposterous.

It is not a fact. In 1936 when they first voted on the Hall of Fame the man with the highest votes was Ty Cobb. More votes than Ruth, Wagner, Johnson, Mathewson ... more than all of them. This man was a murderer, a wife beater, a racist who committed violated hate crimes, and even beat a disabled man in the stands! He is arguably the worst human being to ever play the game. And guess what? He was the first man in.

Did I say "how much" ones character defines their greatness. I just said it is in the formula...at least in determining HOF'ers...like it or not, it's a fact...bonk on.

Well most of the major records I don't feel will be broken, or at least no more frequently then someone's WAR stats are beaten by someone else...so what's the difference?

The difference is when someone reaches X number of average WAR 8.0 seasons they are simply added to the group, with no one in the group being removed.

I have no problem with that logic....I have a problem when it is suggested that you can accurately state how great someone is based by their war numbers.

e.g. everyone with 5+ 8.0 WAR numbers are great players, but player with war value 9 is not necessarily greater than player with war value 8.8

Well, WAR was designed (and has been implemented) to tell us exactly how valuable a player is. Obviously one major league team felt so strongly about WAR that Sean Smith (the leading force behind WAR) was signed by a team and promptly gagged. It has completely transformed Baseball-Reference.com (the #1 stat site on the internet) and launched FanGraphs (the most cutting edge stat site). It has effected the Hall of Fame voting (Blyleven) and even MLB award voting (Felix Hernandez' 2010 Cy Young). It has entered the vernacular and been used in articles in Sports Illustrated and referred to on ESPN.

WAR is hear to stay and it's transforming how we view the game today, and how we revisit the game.

A player with a WAR of 9.0 does have greater value than a player with a WAR of 8.8. Granted, the differential is only 0.2, but it is still a measurable difference. Just like $9 is greater than $8.80.

I meant, on your 10 year list, if a player has these 10 pwar values would he be considered:

10.0
10.0
10.0
10.0

4.8
4.8
4.8
4.8
4.8
4.8

I mean did a player HAVE to have x amount of PWAR+8 or x amount of PWAR+5 to be considered on the 10 year list you made?

That player would have an impressive pWAR of 9.0 and obviously would have been included. Every player that I find with a pWAR of 5.0+ is expanded to 7 and 10 year values.

 

[hofautos]

Here's an example of why I think WAR is a very questionable statistic even on just a single season basis:

Albert Pujols
2006 WAR 8.3
plate appearances 634 BA .331 Runs 119 HRs 49 RBI 137 SB 7 Fielding% .996 Total Bases 359
Slugging% .671

2008 WAR 9.6
plate appearances 641 BA .357 Runs 100 HRs 37 RBI 116 SB 7 Fielding% .996 Total Bases 342
Slugging% .653


There are a hundreds of anomalies like this, which in my opinion debunks it's measure of greatness. In 2008 he was a better singles and doubles hitter. His real production was down, in more plate appearances he touched fewer bases, scored fewer runs, knocked in fewer players, and had fewer extra base hits.

Yet despite the decrease in production, Pujols 2008 WAR is considerably higher. That honestly makes no sense... he's 2006 season was much stronger on a stat basis. Also if you were a fan of the game you'd much prefer his 2006 hitting.


Here's a really defining stat for 2006 verse 2008 for Albert Pujols... Clutch hitting:

2006:
RISP BA .397
RISP / 2 Outs BA .435

2008
RISP BA .339
RISP / 2 Outs BA .326

Even his OBP with RISP was down in 2008, which indicates it wasn't due to walks. He was just a better clutch hitter in 2006. Yet he is rewarding more heavily by WAR for getting singles and doubles with the bases empty... That's just plain silly.

I never got an answer from Chevy on clutch/situation hitting. If it is not in WAR, then besides Longevity, it's inability to define a great catcher, i would add that as a THIRD MAJOR flaw. Clutch hitting in my mind should carry HUGE weight. Some people can get hits when there is no pressure but unable to come through when it really matters, and other players are just the opposite. I know, because I used to be able to shoot a great game of pool, but when there was pressure, anxiety or stress or something would always get to me, and i would choke...which is also why I always shot a better game after 1 or 2 beers....3-4 and i declined again...ANYWAY, point being, CLUTCH HITTING ABSOLUTELY NEEDS TO CARRY A HUGE WEIGHT IN A FORMULA I WOULD ACCEPT.

 

[hofautos]

Here's an example of why I think WAR is a very questionable statistic even on just a single season basis:

Albert Pujols
2006 WAR 8.3
plate appearances 634 BA .331 Runs 119 HRs 49 RBI 137 SB 7 Fielding% .996 Total Bases 359
Slugging% .671

2008 WAR 9.6
plate appearances 641 BA .357 Runs 100 HRs 37 RBI 116 SB 7 Fielding% .996 Total Bases 342
Slugging% .653


There are a hundreds of anomalies like this, which in my opinion debunks it's measure of greatness. In 2008 he was a better singles and doubles hitter. His real production was down, in more plate appearances he touched fewer bases, scored fewer runs, knocked in fewer players, and had fewer extra base hits.

Yet despite the decrease in production, Pujols 2008 WAR is considerably higher. That honestly makes no sense... he's 2006 season was much stronger on a stat basis. Also if you were a fan of the game you'd much prefer his 2006 hitting.


Here's a really defining stat for 2006 verse 2008 for Albert Pujols... Clutch hitting:

2006:
RISP BA .397
RISP / 2 Outs BA .435

2008
RISP BA .339
RISP / 2 Outs BA .326

Even his OBP with RISP was done in 2008, which indicates it wasn't due to walks. He was just a better clutch hitter in 2006. Yet he is rewarding more heavily by WAR for getting singles and doubles with the bases empty... That's just plain silly.

WAR is re-adjusted each year, based on the performance of the entire league.

Therefore, if the same player has the same stats for three consecutive years, each year he will have a different WAR based on the performance of the rest of the league.

WAR is not a formula that says H + RBI + HR + BB = X

It measures a player's performance in these regards versus the rest of the league.

Therefore, when you have McGwire and Sosa both hitting a ton of home runs in the same seasons, it does not create the same WAR shockwave as when Ruth was hitting more home runs than the entire league.

As an example when Maris hit 61, Mantle was "close." Not a dramatic WAR. When Ruth hit 60, Gehrig was "close." Not Ruth's best WAR.

I actually agree with that concept...but how do they determine how much weight. They must have been drawing at straws on some of the formulas (grin).
Seriously though, I hope it considers "situational hitting"..i still haven't seen a response on that one.