Any statisticians around? Warning: Math!

[Guest]

Many of my critics argue that the flaw in my formula is that a five year sample is simply too small. These critics suggest that a ten year sample would be more reliable.

I doubt any of them are going to spend the time developing a formula for ten years, so I have done the work for them.

What follows is a ranking of pitchers based on WAR using a 10 year sample.

WAR 8.0+ (MVP)
Young, Cy
Johnson, Walter
Nichols, Kid
Clarkson, John

WAR 5.0+ (All-Star)
Mathewson, Christy
Keefe, Tim
Grove, Lefty
Alexander, Pete
Gibson, Bob
Perry, Gaylord
Radbourn, Old Hoss
Seaver, Tom
Galvin, Pud
Niekro, Phil
Spahn, Warren
Carlton, Steve
Welch, Mickey
Jenkins, Fergie
Roberts, Robin
Feller, Bob
Blyleven, Bert
Rusie, Amos
Marichal, Juan
Bunning, Jim
Drysdale, Don
Palmer, Jim
Hubbell, Carl
Plank, Eddie
Willis, Vic
Ryan, Nolan
Walsh, Ed
Vance, Dazzy
Koufax, Sandy
McGinnity, Joe
Coveleski, Stan
Newhouser, Hal
Brown, Mordecai
Waddell, Rube

WAR 2.0+ (Starter)
Lyons, Ted
Sutton, Don
Wynn, Early
Gomez, Lefty
Ford, Whitey
Ruffing, Red
Faber, Red
Rixey, Eppa
Lemon, Bob
Joss, Addie
Dean, Dizzy
Grimes, Burleigh
Hoyt, Waite
Bender, Chief
Pennock, Herb
Chesbro, Jack
Hunter, Catfish
Marquard, Rube
Haines, Jesse

Since these same critics argue that ten good seasons are required for induction into the Hall of Fame, by their lithmus test the players listed in the third section of this ranking are not worthy of inclusion.

19 of 58 pitchers in the Hall of Fame failed to generate ten seasons of All-Star credentials. That's a staggering 33% or 1/3rd of the inductees.

Now that you have the full numbers on five, seven, and ten years for pitchers, do you not now see that five years does in fact represent the actual prime of a baseball player's career?

 

[jbhofmann]

I would argue that most of the last list doesn't need to be in the HOF. I think it kinda proves the point that more than 5 great seasons are needed.

 

[jbhofmann]

I would like to see a formula that is something like:

(p)WAR + (HOF Monitor x .5) + (Black Ink x .1) + (Gray Ink x .2)=

 

[ChasHawk]

I would like to see a formula which makes Chris Levy and his agenda go away.

 

[pigskincardboard]

Chris, here's the problem with your formula and the critics would agree:

You cannot arbitrarily arrive at a number. You cannot say, "I looked at HOFers and saw that most had 5 peak seasons."

If you want to combat the critics, all you do is use math. Your explanation of five years is long-winded; whereas saying, when HOFers are normally distributed according length to productivity, the median amount of greatness was five years.

Greatness was defined by 2-std (or 3) away from the mean -- the top 5 or 0.3 percent, retrospectively -- compared to one's peers. Pitchers and hitters were adjusted & compared independently to account for various "eras" in baseball history.

All career lengths were normalized to a standard length based on the average length of a career for players with more than 3-years playing time.

Based on our data, the margin of error was under one player-divided-year and we're comfortable in our conclusions.

Here's the thing with math: you can't half ass it and you can't arrive at your conclusions (ie. Sandy Koufax is a top-10 player) before you start the modeling.

There's a very large difference between dicking around and wanting to publish something. If you publish your current result set, it'll be torn to absolute pieces. Everything has to have a reason, and that reason must be supported by math.

If you want to know why Sandy Koufax is revered, it's because he's not only great but was Jewish. Stupid things like skin-color, ethnicity, etc., still play a part in determining how popular a player is.

 

[Guest]

I would like to see a formula which makes Chris Levy and his agenda go away.

What agenda is that? Why does this seem to irritate you so much?

 

[ROLLTIDE4LIFE]

I would like to see a formula which makes Chris Levy and his agenda go away.

Stop being so resistant. You are ignorant. Block Levy's post if you want him to go away. If not just realize you and your backwoods thinking is wrong. That is a fact.

 

[phillyfan0417]

Guys,

Levy is presenting an opinion. I'm not really sure why there needs to be venom about it. If you dont agree, its pretty simple to present your side of the debate and move on, this does not need to become personal.

 

[njlw226]

I'm a big math guy myself and love stuff like this. Big thanks to the OP.

I also have to say that I'm amazed at the negative reaction some people in this thread have reacted with. The OP was simply providing some analysis for discussion and it turned into personal attacks. I may not necessarily agree 100% with all of his methods, but I think it's silly to attack him for this. Just my 2 cents...

 

[pigskincardboard]

I think anyone that's studied baseball and it's effect on popular american culture and it's growth in popularity during specific events could take a gander at why old-timey baseball people hate advanced statistics.

I know a lot of you guys hate getting philosophical, but baseball means a lot more than simply statistics and determining a player's worth.

I'm sure there are plenty of good books out there that tell you exactly why old-timers hate stat nerds.

To defeat the statistics guru, simply ask for his margin of error.

Ask him for his correlation coefficient.

Not everything is quantifiable and not every statistic is bullet proof.

Any statistician that's not willing to admit that there's an X% chance that he's wrong is a moron.

 

[hofautos]

^^^
I did say I thought it would be interesting to see your 10 year numbers that I would give them more credence over ANY 5 year formula, but never suggested it should be considered for one's admission into the HOF or even that I would consider it as a single formula to define one's greatness.

I never suggested any formula to determine who should get in the HOF, so I don't know who you are referring to.
I don't think WAR or any single stat should be used.

I believe a player's entire career, including all of their stats, as well as their character should be considered.

I am suggesting that no single formula can be created to determine one's greatness, and if one could, that baseball references HOF monitor does a far better job than your 5 year one, and if you don't recognize a flaw in your formula, that you are only fooling yourself.

PSS- I was hoping your 10 year numbers would be for their best 10 years, not only if they had 10 years over 5 WAR.
Simply total their best 10 years.

 

[Guest]

^^^
I did say I thought it would be interesting to see your 10 year numbers that I would give them more credence over ANY 5 year formula, but never suggested it should be considered for one's admission into the HOF or even that I would consider it as a single formula to define one's greatness.

I never suggested any formula to determine who should get in the HOF, so I don't know who you are referring to.
I don't think WAR or any single stat should be used.

I believe a player's entire career, including all of their stats, as well as their character should be considered.

I am suggesting that no single formula can be created to determine one's greatness, and if one could, that baseball references HOF monitor does a far better job than your 5 year one, and if you don't recognize a flaw in your formula, that you are only fooling yourself.

I'm sorry, but I just don't see how career numbers can be used to accurately compare one player to another.

It seems unfair to credit someone who accomplished a task in eighteen seasons the same as another who accomplished it in fifteen. You seem perfectly content to say to a player "go out there and try as long as your body holds up and if you reach 500 home runs, 3,000 hits, or 300 wins you'll be a great player."

I can't understand how you, or anyone for that matter, is comfortable comparing one player to another based on a variable sample size.

 

[hofautos]

^^^
I did say I thought it would be interesting to see your 10 year numbers that I would give them more credence over ANY 5 year formula, but never suggested it should be considered for one's admission into the HOF or even that I would consider it as a single formula to define one's greatness.

I never suggested any formula to determine who should get in the HOF, so I don't know who you are referring to.
I don't think WAR or any single stat should be used.

I believe a player's entire career, including all of their stats, as well as their character should be considered.

I am suggesting that no single formula can be created to determine one's greatness, and if one could, that baseball references HOF monitor does a far better job than your 5 year one, and if you don't recognize a flaw in your formula, that you are only fooling yourself.

I'm sorry, but I just don't see how career numbers can be used to accurately compare one player to another.

It seems unfair to credit someone who accomplished a task in eighteen seasons the same as another who accomplished it in fifteen. You seem perfectly content to say to a player "go out there and try as long as your body holds up and if you reach 500 home runs, 3,000 hits, or 300 wins you'll be a great player."

I can't understand how you, or anyone for that matter, is comfortable comparing one player to another based on a variable sample size.

It seems unfair to suggest someone is better based on 5 years than someone who had almost similar 10 years numbers.

EXAMPLE your PLAYER A is NOT better than my Player B.

year 1 = 9
year 2 = 9
year 3 = 9
year 4 = 9
year 5 = 9
year 6 = 5
year 7 = 4
year 8 = 2


year 1 = 9
year 2 = 9
year 3 = 9
year 4 = 9
year 5 = 8
year 6 = 8
year 7 = 8
year 8 = 8
year 9 = 8
year 10 = 8

I just can't accept a 5 year. Perhaps if you could come up with something that used your formula IN ADDITION to serious bonus points for great longevity.
AND I DO give someone bonus points who can sustain for long periods, especially in breaking records. I believe that does add some factor to their greatness.
Using your same answers, you could say only one year is necessary? What in your mind is wrong with just using 1 year instead of 5 years to determine one's greatness?

 

[reljac]

As an actuary, any attempts to quantify a players 'primary value' would be more subjective than objective. The subjectivity would be inherent in most of the choices you make (similar to the subjectivity in WAR).

If you choose to look at each year or era as independant then you would not weigh players equally. This would result in bias since it assumes the game is different.

If you look at the overal picture devoid of eras/years, then you add bias in assuming the game is the same. obviously this would eliminate most Pre-War players from the hall and heavily weight the Hall with late 80's and 90's players.

If you try to quantify the value above the mean that a player brings to a team, you make the mistake of weighing all stats equally. A top stolen base guys SBs are a much greater variance from the mean than a top HR guy guys HRs.

The other difficulty in any statistical model is that it will never capture fan appeal. Fan Appeal comes into play in many ways, the team the player plays for (think Yankees vs. Royals), the 'hero status' of the player (Think Dale Murphy), off the field contributions (Think Roberto Clemente), steroid usage even if not proven (think Mark McGwire)... these are all entirely subjective portions that become part of voters decisions.

Koufax's value above mean was exceptional in his dominant years. But if you look at both his early career seasonal stats and his career stats. You may not have seen a HOFer in the making.


In the end to statistically come up with something is a significant undertaking. I've done similar projects during my actuarial exam work.

If you goal is to create a short list of guys to collect, why not just start listing your favorite HOFers.

 

[ChasHawk]

Mike - he doesn't get it, and won't. Longevity & the ability to avoid/recover from injury should always be considered when measuring a players career, and he just doesn't understand that at all.

These are the same type of people who call Jim Brown the greatest NFL player of all time because of his 5.2 yds per carry.
Nobody EVER takes into consideration the fact that he retired after only 9 seasons. He was 29 years old when he retired, so he never played through his decline years. Jim Brown loves his yds/carry record, but a big reason he has it is because of the small sample size.

Sandy Koufax was a great pitcher, for 6 years. He also had 6 terrible years at the beginning of his career and retired at 30.

He was a great pitcher, but he's not even the best left-handed pitcher of all time.

You have to look at the entire career, period.

 

[jbhofmann]

To be fair, I don't think any respected person would claim that Emmitt Smith was greater than Jim Brown.

 

[hofautos]

Mike - he doesn't get it, and won't. Longevity & the ability to avoid/recover from injury should always be considered when measuring a players career, and he just doesn't understand that at all.

These are the same type of people who call Jim Brown the greatest NFL player of all time because of his 5.2 yds per carry.
Nobody EVER takes into consideration the fact that he retired after only 9 seasons. He was 29 years old when he retired, so he never played through his decline years. Jim Brown loves his yds/carry record, but a big reason he has it is because of the small sample size.

Sandy Koufax was a great pitcher, for 6 years. He also had 6 terrible years at the beginning of his career and retired at 30.

He was a great pitcher, but he's not even the best left-handed pitcher of all time.

You have to look at the entire career, period.

To some point, I agree that declining years shouldn't be used in defining one's greatness. Example many great players ERA and BA go down because they strive to make milestones, or simply because they love the game. I just think that there is no "magic formula", but if one was created, it would be a combination of all stats, and different weights given to different stats, but no matter what, longevity is an important one. I don't discount anything Levy has done, I just think he needs to modify his efforts if he wants to gain more acceptance. I noticed that Baseball reference is now including WAR numbers too. I like baseball's reference formula, but if somehow they included WAR into their HOF monitor, they could make their formula even better.

 

[ChasHawk]

To be fair, I don't think any respected person would claim that Emmitt Smith was greater than Jim Brown.

Nope. I don't even think that Emmit Smith was greater than Barry Sanders or Walter Payton.

What I do know is that Brown retired at the height of his career, which is why his avg is so high.
I also know that he was bigger than most defensive lineman that he ran against.

You can't just quantiy everything in life with numbers.

 

[RL24]

I would like to see a formula which makes Chris Levy and his agenda go away.

It's pretty easy actually. You foe him. It's so easy in fact, I think I'm going to give it a go right now so I don't have to see your 12 posts in this thread. Goodbye chashawk's agenda!

 

[phillyfan0417]

To some point, I agree that declining years shouldn't be used in defining one's greatness. Example many great players ERA and BA go down because they strive to make milestones, or simply because they love the game. I just think that there is no "magic formula", but if one was created, it would be a combination of all stats, and different weights given to different stats, but no matter what, longevity is an important one. I don't discount anything Levy has done, I just think he needs to modify his efforts if he wants to gain more acceptance. I noticed that Baseball reference is now including WAR numbers too. I like baseball's reference formula, but if somehow they included WAR into their HOF monitor, they could make their formula even better.

Longevity is something to take into consideration but stats are far more important. There are many players who hang around for years who are nothing more than role players and minor stars.

I love watching the game and seeing the ins and outs and for years I would argue you cant tell much without watching the player but it simply isnt true. Watching the game is emotional, stats are stats.

There is room for both schools but stats win in the end because they show your body of work. You cant use emotion to argue numbers.