Win Shares - Page 2

Jason Moyer · 06-05-2002, 09:54 AM

I'll try and put it somewhere tonight, if I can, probably after midnight tho.

For some unknown reason, I am just completely addicted to Puresim right now. I simmed 2 complete seasons yesterday (i.e. about 8 hours of simming).

Jason

rcbuss · 06-05-2002, 04:11 PM

I'd also be interested in the file.
Would you be able to drag in an appropriately formatted stat file to do blanket calculations for the entire league?

swampdragon · 06-05-2002, 05:00 PM

I have been reading James' books since the early 80's and also have Win Shares. The concept is intriguing, but I have my doubts. I think it rewards players who played for lucky teams that won all their close games. To the extent clutch performance is not luck, that's defensible. However, one of the things James had been trying to teach 15-20 years ago is that luck is a large factor in clutch performance.
Second, the system shortchanges current AL hitters, since an individual hitter will be responsible for a smaller percentage of the offense in the era of the DH. Since the offense-defense balance doesn't change, the individual AL hitters are therefore responsible for fewer wins. That may or may not be true, but it doesn't really help us evaluate players across leagues.

mtw · 06-05-2002, 05:16 PM

<blockquote>quote:<hr />Originally posted by Jason Moyer:
A shortstop or modern second baseman (post-1930) has a much larger effect on the offense that is being generated against his team than James gives credit for, IMHO. I don't think catchers deserve much more of a share than first basemen do. Aside from a marginal effect on the team ERA, because of the way they handle a pitching staff, they're basically useless.

Jason<hr /></blockquote>Here's where the problem with Bill James concepts, such as Win Shares: there are just too many factors to consider. On a team composed of entirely of right-handed flyball pitchers, a shortstop's value would be less than on team of all left-handed groundball pitchers. Also, how good are a team's pitchers at holding runners? That would affect a catcher's ability to prevent steal attempts and throw out basestealers. Also, a good defensive catcher not only throws runners out, he prevents them from running in the first place. What's the value of that, and how does it relate to the frequency of steal attempts if that team had an average catcher or if the pitchers were terrible at holding runners? Also a good defensive catcher can prevent wild pitches and passed balls. What's the value of these "non-events" and how much less valuable is a catcher, if at all, if the pitchers have great control? What value on an OF throw to the right base, or hitting the cutoff man, to prevent runners from advancing?

How much impact do these "non-stats" have on a game's outcome?

Can Bill James tell me?

Covenant47 · 06-05-2002, 10:08 PM

I like Runs Created personally. I'm not sure that Win Shares improve on Runs Created at all.

andy m · 06-05-2002, 10:24 PM

<blockquote>quote:<hr />Originally posted by mtw:
Can Bill James tell me?

<hr /></blockquote>he could, but he is too busy writing his new book on 'run shares', where he attempts to work out who is responsible for every single run scored in the history of baseball. it will include factors such as how tall the grass is, the exact weight of the baseball at the time of pitch AND when it reaches the batter (possibility of moisture adhering to ball in flight), and whether the batter is wearing shoes that are too big. look for it at amazon.com in the year 2020.

OldGiants · 06-06-2002, 06:23 AM

<blockquote>quote:<hr />Originally posted by Jason Moyer:
 <blockquote>quote:<hr />Originally posted by OldGiants:
Can there be negative win shares? If not, I'd have to question the validity of the analysis.<hr /></blockquote>Team's can't win negative games, so no you can't.

Jason<hr /></blockquote>Wrong. A negative win is a loss. I had to cut my post short, or I would have given this example:

Sosa hits a 3-run homer in the top of the ninth to put the Cubs up 5-4. Alfonseca walks a man, then allows a 2-run homer.

IMO, Sosa earns a win share, Alfonseca a loss share. Coming away from this game with a 'zero' for both players because the Cubs lost the game is flat-out wrong.

Forcing 'win shares' to equal the team's win total can never be a valid determination of a player's value. It favors players on better teams far too much, even with the tripling idea.

And it overlooks the way the game is really played, as my simple example points out.

I like James and applaud his body of work, but this idea is too far from the truth to be useful.

[ 06-06-2002, 12:26 PM: Message edited by: OldGiants ]

mtw · 06-06-2002, 08:22 AM

<blockquote>quote:<hr />Originally posted by andy m:
he could, but he is too busy writing his new book on 'run shares', where he attempts to work out who is responsible for every single run scored in the history of baseball. .<hr /></blockquote>

woburn · 06-06-2002, 10:34 AM

[qb] <blockquote>quote:<hr />Originally posted by OldGiants:
Forcing 'win shares' to equal the team's win total can never be a valid determination of a player's value. It favors players on better teams far too much, even with the tripling idea.
QB]<hr /></blockquote>I don't believe that's true. There's a study in the book which matches a pitcher on a good team with a pitcher on a bad team, where their performance was very similar. It showed that the Win Shares were also very similar.

Jason Moyer · 06-06-2002, 01:15 PM

<blockquote>quote:<hr />Originally posted by OldGiants:
[QB][QUOTE]Forcing 'win shares' to equal the team's win total can never be a valid determination of a player's value. It favors players on better teams far too much, even with the tripling idea.[QB]<hr /></blockquote>I think you're missing something here.

If a team goes 60-120, but one of the players has a monster year, creating more runs per out of any other player, that means that the rest of the team, frankly, sucks. As such, that player will get a ridiculously large portion of the win shares, in comparison to his teammates. The team has 180 win shares to divide, and chances are that player will get 30-50 of them, if he had an MVP type season.

If a team goes 120-60, and a player has the same season, it means that his other teammates are pulling their share of the load. However, this team has 360 win shares to divide, so that player can still get his 30-50 even tho some of his teammates will also likely get 20-30 or more.

The entire idea behind win shares is that the players on a good team are better than the players on a bad team. This makes sense. A good player on a bad team is still as valuable to his team as a player on a good team, he'll just make up a larger percentage of the team's value.

Let's look at some examples from the NL in 2001.

Team - Record - Win Share Leader:

St Louis - 90-69 - Jim Edmonds 30
Pittsburgh - 62-100 - Brian Giles 29
San Diego - 79-83 - Phil Nevin 31
New York - 82-80 - Mike Piazza 21
Montreal - 68-94 - Orlando Cabrera 26
San Francisco - 90-72 - Barry Bonds 54
Arizona - 92-70 - Luis Gonzalez 37

See anywhere on that list where a player is doing lousy simply because his team is lousy? I sure don't. Phil Nevin had a damn fine year last year, despite being on a losing team. Jim Edmonds and Brian Giles were equally valuable, even tho Giles played for a team that lost 100 games.

Some neat observations that I think we'd all agree with can be made from this. The San Francisco Giants, I think most people would agree, had awful pitching last year. Shortstop Rich Aurillia had 33 win shares - that's 2 MVP caliber seasons on one team, pretty good for a team with only 90 wins - unfortunately their best SP was Russ Ortiz with 15, which is bad for a #1 SP. By comparison, Randy Johnson had 26 win shares and Curt Schilling had 24, as you would expect, looking at their individual statistics. Why was Arizona better? Gee, let's see. Luis Gonzalez had 37 win shares and Schilling/Johnson combined for 50. They had 13 guys who gained credit for 10 or more win shares, which is pretty good depth. The Giants had 2 MVP type seasons (Bonds/Aurilia) and a good season from Kent (27 win shares) but after that things fall apart quickly. Ther had just 9 guys gain 10 or more win shares, and most of those shares were attributable to 3 players. Russ Ortiz was the 4th best player on the team. The 10th best player on the Diamonbacks was just some nobody named Jay Bell.

Jason

Jason Moyer · 06-06-2002, 01:33 PM

I should also note that Win Shares very rarely tells you something that you don't know. When it does, it is interesting.

For instance, you would assume the Phillies best player last year was Bobby Abreu, right? Well, not according to Win Shares:

1. Scott Rolen 29
2. Bobby Abreu 26
3. Jimmy Rollins 20 (by the way, this is an incredibly high total for a rookie - most rookie of the year winners have somewhere around 10-15 win shares - unfortunately, Jimmy picked up 20 win shares the same season Pujols picked up 29)

Thinking about their hitting stats, that doesn't really seem right. But if you look on a closer level, here are their runs created and outs:

1. Scott Rolen 102 RC 417 Outs
2. Bobby Abreu 126 RC 454 Outs
3. Jimmy Rollins 89 RC 503 Outs

Looking at it in terms of RC/12 outs:

1. Scott Rolen 2.94
2. Bobby Abreu 3.33
3. Jimmy Rollins 2.12

That's pretty close. Jimmy is a full RC/12 behind, but Bobby wasn't that much more productive than Scott was as the plate, in terms of actually creating offense for his team (and offense is measured in runs right? homeruns and batting average don't win games, creating runs does).

Throw in the defensive win shares, and this is a no brainer. Whether you use win shares or just watch baseball, you know that Scott is the best defensive third baseman in baseball right now and over the last 4-5 years, if not ever. Bobby is a good right fielder, but he's not the outfield equivalent of Scott Rolen. Likewise, Jimmy gets a boost because he was much better with the glove than even the Philadelphia scouts and management had expected (Bowa and several analysts even said this several times last summer, so it's no surprise that Win Shares backs this up).

Even something in Win Shares that seems kinda odd, at first glance, comes back looking pretty good when you sit and analyze it.

Jason

Jason Moyer · 06-06-2002, 02:18 PM

<blockquote>quote:<hr />Originally posted by mtw:
Here's where the problem with Bill James concepts, such as Win Shares: there are just too many factors to consider. On a team composed of entirely of right-handed flyball pitchers, a shortstop's value would be less than on team of all left-handed groundball pitchers. Also, how good are a team's pitchers at holding runners? That would affect a catcher's ability to prevent steal attempts and throw out basestealers. Also, a good defensive catcher not only throws runners out, he prevents them from running in the first place. What's the value of that, and how does it relate to the frequency of steal attempts if that team had an average catcher or if the pitchers were terrible at holding runners? Also a good defensive catcher can prevent wild pitches and passed balls. What's the value of these "non-events" and how much less valuable is a catcher, if at all, if the pitchers have great control? What value on an OF throw to the right base, or hitting the cutoff man, to prevent runners from advancing?

How much impact do these "non-stats" have on a game's outcome?

Can Bill James tell me?

<hr /></blockquote>Yes. Have you read Win Shares? All of the factors you list are taken into account in his system.

This isn't some 5 line formula we're talking about here. The formula itself takes up pages 1 to 77 in the book, and makes adjustments for everything that it is possible to adjust for.

Team has a lot of left handed pitchers? The system assumes your third baseman should receive more chances and your first baseman should receive fewer independent chances. A lot of flyball pitchers? Adjusts for that too. A first baseman who prefers to flip to the pitcher instead of just tagging the base himself, making his stats look more impressive? Adjusts for that too. A catcher who plays a lot of innings and allows fewer stolen bases than he would be expected to allow? Adjusts for that. A team with a groundball staff should have how many double plays assuming a certain number of balls were put in play against the pitcher? Adjusts for that.

Jason

OldGiants · 06-06-2002, 03:03 PM

Thanks for all your posts, Jason. I'll definitely have to do more than browse the book.

I'm not convinced about win shares, but I'll read more.

Jason Moyer · 06-06-2002, 03:13 PM

<blockquote>quote:<hr />Originally posted by OldGiants:
Thanks for all your posts, Jason. I'll definitely have to do more than browse the book.

I'm not convinced about win shares, but I'll read more.<hr /></blockquote>I don't mean to be redundant but I do want to say that:

1. I don't agree with them 100% (mostly because I'd like to see the data he's basing the intrinsic weights on)

2. Because so much of the data has to be figured by estimating, they aren't completely accurate on a historical level. I believe James says that a difference of 3 or less is not worth arguing over because of the estimations that are required in the system. Hopefully this work will mean that STATS at least will track more meaningful stats (independent putouts by first basemen, non-pitcher assists by first basemen, independent putouts by catchers, runners on first base, etc). That way, at least in real baseball, we could calculate win shares without estimating anything.

Jason

mking55 · 06-07-2002, 03:50 AM

Jason,

Did you ever finish your WinShares spreadsheets? I am curious to try them out. Although I am not convinced of it's validity, I do admit that I am intrigued by the whole idea.

I saw that you posted a couple of days ago that you were going to make them available for download, but I don't see the link. Is there?

Thanks.

eneubaue · 06-07-2002, 04:27 AM

Jason,

Good points on Win Shares - I, too, have been an avid reader/follower of James' work since 1981 (I have a copy of one of his early abstracts he sold via mail-order).

I think people should be aware that James' himself said that his Win Shares system is a first attempt to try to account for all elements in baseball (batting, fielding, pitching) in trying to evaulate/analyze players and how much each player contributes to his team winning. Once others start to play around with his system and make modifications/improvements, then I think we'll begin to see some of the benefits of his Win Shares system. I'd like to see the folks over at the Baseball Prospectus get their hands on it and run with it, making changes/modifications as necessary.

I, like you, don't agree with everything James' has to offer (what's up with his modifications on the Runs Created formulas?), but at least he has provided a place to start.

Malleus Dei · 06-07-2002, 04:44 AM

I didn't like the concept of Win Shares when I first heard about it.

Then I looked into it and found that it was an excellent and a very useful statistic.

IMO it has a poorly-chosen name that turns some people off to it right away.

But do take the time to take a real hard look at it before you start to criticize it. There's more to it than you might think.

Jason Moyer · 06-08-2002, 05:03 PM

My biggest overall beef with win shares, the more think about it, are the defensive adjustments.

I think defensive win shares are absolutely amazing at showing who is the best player. But I'm not sure I agree that it shows who contributed more.

For instance, let's say you have Steve Garvey at first on one team and Bill Buckner on another. Ok, we know Garvey is a better first baseman, altho Buckner is still good. However, if the conditions they're playing under cause the first baseman of Bucnker's team to have a lot more chances than Garvey, shouldn't Buckner get more Win Shares? I mean, you can have a defensive wizard playing shortstop for instance, but on a flyball team maybe he isn't contributing as much as your so-so center fielder.

It's weird. I almost think that defensive win shares are great for evaluating a player's talent, but that the total baseball-style linear weights (for defense) are almost better at gauging actual win contribution...

Am I messing something up logically here? If I were on the SABR list I'd ask them.

Jason

inverarity · 06-09-2002, 08:29 AM

Great thread - I just want to throw in another YES to Jason's question of whether anyone would want his win shares spreadsheets. Definitely! Jason, are you still considering making them available?

[ 06-09-2002, 02:32 PM: Message edited by: inverarity ]

Jason Moyer · 06-09-2002, 03:36 PM

Yes, I apologize it's taking so long.

Unfortunately job hunting/GTA3 have been eating my time. I should get a chance to work on them tommorrow after I send out a few more resumes.

Glad you're interested.

Jason

06-05-2002, 09:54 AM	#21
Jason Moyer Hall Of Famer Join Date: Feb 2002 Location: Pittsburgh Posts: 5,108	I'll try and put it somewhere tonight, if I can, probably after midnight tho. For some unknown reason, I am just completely addicted to Puresim right now. I simmed 2 complete seasons yesterday (i.e. about 8 hours of simming). Jason __________________ "I pretty much popped everything cold turkey. We were doing steroids they wouldn't give to horses." -- Tom House "I was very fortunate to have a pitching coach by the name of Tom House...Tom, I really miss those days that we spent in the weight room and out on the field working together." -- Nolan Ryan's HoF Induction Speech

06-05-2002, 04:11 PM	#22
rcbuss Minors (Triple A) Join Date: Apr 2002 Location: Mauston, WI Posts: 226	I'd also be interested in the file. Would you be able to drag in an appropriately formatted stat file to do blanket calculations for the entire league? __________________ Robert C Buss FOBL Mauston Mad Cows

06-05-2002, 05:00 PM	#23
swampdragon Hall Of Famer Join Date: May 2002 Location: The Lonely Mountain Posts: 2,506	I have been reading James' books since the early 80's and also have Win Shares. The concept is intriguing, but I have my doubts. I think it rewards players who played for lucky teams that won all their close games. To the extent clutch performance is not luck, that's defensible. However, one of the things James had been trying to teach 15-20 years ago is that luck is a large factor in clutch performance. Second, the system shortchanges current AL hitters, since an individual hitter will be responsible for a smaller percentage of the offense in the era of the DH. Since the offense-defense balance doesn't change, the individual AL hitters are therefore responsible for fewer wins. That may or may not be true, but it doesn't really help us evaluate players across leagues. __________________ “Of all tyrannies, a tyranny sincerely exercised for the good of its victims may be the most oppressive. It would be better to live under robber barons than under omnipotent moral busybodies." -- C.S. Lewis

06-05-2002, 05:16 PM	#24
mtw All Star Reserve Join Date: Dec 2001 Posts: 653	</font><blockquote><font size="1" face="Verdana, Arial">quote:</font><hr /><font size="2" face="Verdana, Arial">Originally posted by Jason Moyer: <strong>A shortstop or modern second baseman (post-1930) has a much larger effect on the offense that is being generated against his team than James gives credit for, IMHO. I don't think catchers deserve much more of a share than first basemen do. Aside from a marginal effect on the team ERA, because of the way they handle a pitching staff, they're basically useless. Jason</strong></font><hr /></blockquote><font size="2" face="Verdana, Arial">Here's where the problem with Bill James concepts, such as Win Shares: there are just too many factors to consider. On a team composed of entirely of right-handed flyball pitchers, a shortstop's value would be less than on team of all left-handed groundball pitchers. Also, how good are a team's pitchers at holding runners? That would affect a catcher's ability to prevent steal attempts and throw out basestealers. Also, a good defensive catcher not only throws runners out, he prevents them from running in the first place. What's the value of that, and how does it relate to the frequency of steal attempts if that team had an average catcher or if the pitchers were terrible at holding runners? Also a good defensive catcher can prevent wild pitches and passed balls. What's the value of these "non-events" and how much less valuable is a catcher, if at all, if the pitchers have great control? What value on an OF throw to the right base, or hitting the cutoff man, to prevent runners from advancing? How much impact do these "non-stats" have on a game's outcome? Can Bill James tell me? __________________ Over-Zealous Apologist

06-05-2002, 10:24 PM	#26
andy m Minors (Double A) Join Date: Dec 2001 Location: UK Posts: 183	</font><blockquote><font size="1" face="Verdana, Arial">quote:</font><hr /><font size="2" face="Verdana, Arial">Originally posted by mtw: <strong>Can Bill James tell me? </strong></font><hr /></blockquote><font size="2" face="Verdana, Arial">he could, but he is too busy writing his new book on 'run shares', where he attempts to work out who is responsible for every single run scored in the history of baseball. it will include factors such as how tall the grass is, the exact weight of the baseball at the time of pitch AND when it reaches the batter (possibility of moisture adhering to ball in flight), and whether the batter is wearing shoes that are too big. look for it at amazon.com in the year 2020. __________________

06-05-2002, 10:08 PM	#25
Covenant47 Bat Boy Join Date: Mar 2002 Posts: 8	I like Runs Created personally. I'm not sure that Win Shares improve on Runs Created at all.

06-06-2002, 06:23 AM	#27
OldGiants Registered User Join Date: Apr 2002 Posts: 211	</font><blockquote><font size="1" face="Verdana, Arial">quote:</font><hr /><font size="2" face="Verdana, Arial">Originally posted by Jason Moyer: <strong> </font><blockquote><font size="1" face="Verdana, Arial">quote:</font><hr /><font size="2" face="Verdana, Arial">Originally posted by OldGiants: <strong>Can there be negative win shares? If not, I'd have to question the validity of the analysis.</strong></font><hr /></blockquote><font size="2" face="Verdana, Arial">Team's can't win negative games, so no you can't. Jason</strong></font><hr /></blockquote><font size="2" face="Verdana, Arial">Wrong. A negative win is a loss. I had to cut my post short, or I would have given this example: Sosa hits a 3-run homer in the top of the ninth to put the Cubs up 5-4. Alfonseca walks a man, then allows a 2-run homer. IMO, Sosa earns a win share, Alfonseca a loss share. Coming away from this game with a 'zero' for both players because the Cubs lost the game is flat-out wrong. Forcing 'win shares' to equal the team's win total can never be a valid determination of a player's value. It favors players on better teams far too much, even with the tripling idea. And it overlooks the way the game is really played, as my simple example points out. I like James and applaud his body of work, but this idea is too far from the truth to be useful. <small>[ 06-06-2002, 12:26 PM: Message edited by: OldGiants ]</small>

06-06-2002, 08:22 AM	#28
mtw All Star Reserve Join Date: Dec 2001 Posts: 653	</font><blockquote><font size="1" face="Verdana, Arial">quote:</font><hr /><font size="2" face="Verdana, Arial">Originally posted by andy m: <strong>he could, but he is too busy writing his new book on 'run shares', where he attempts to work out who is responsible for every single run scored in the history of baseball. .</strong></font><hr /></blockquote><font size="2" face="Verdana, Arial"> __________________ Over-Zealous Apologist

06-06-2002, 10:34 AM	#29
woburn Registered User Join Date: Jan 2002 Posts: 191	[qb] </font><blockquote><font size="1" face="Verdana, Arial">quote:</font><hr /><font size="2" face="Verdana, Arial">Originally posted by OldGiants: Forcing 'win shares' to equal the team's win total can never be a valid determination of a player's value. It favors players on better teams far too much, even with the tripling idea. QB]</font><hr /></blockquote><font size="2" face="Verdana, Arial">I don't believe that's true. There's a study in the book which matches a pitcher on a good team with a pitcher on a bad team, where their performance was very similar. It showed that the Win Shares were also very similar.

06-06-2002, 01:15 PM	#30
Jason Moyer Hall Of Famer Join Date: Feb 2002 Location: Pittsburgh Posts: 5,108	</font><blockquote><font size="1" face="Verdana, Arial">quote:</font><hr /><font size="2" face="Verdana, Arial">Originally posted by OldGiants: [QB][QUOTE]Forcing 'win shares' to equal the team's win total can never be a valid determination of a player's value. It favors players on better teams far too much, even with the tripling idea.[QB]</font><hr /></blockquote><font size="2" face="Verdana, Arial">I think you're missing something here. If a team goes 60-120, but one of the players has a monster year, creating more runs per out of any other player, that means that the rest of the team, frankly, sucks. As such, that player will get a ridiculously large portion of the win shares, in comparison to his teammates. The team has 180 win shares to divide, and chances are that player will get 30-50 of them, if he had an MVP type season. If a team goes 120-60, and a player has the same season, it means that his other teammates are pulling their share of the load. However, this team has 360 win shares to divide, so that player can still get his 30-50 even tho some of his teammates will also likely get 20-30 or more. The entire idea behind win shares is that the players on a good team are better than the players on a bad team. This makes sense. A good player on a bad team is still as valuable to his team as a player on a good team, he'll just make up a larger percentage of the team's value. Let's look at some examples from the NL in 2001. Team - Record - Win Share Leader: St Louis - 90-69 - Jim Edmonds 30 Pittsburgh - 62-100 - Brian Giles 29 San Diego - 79-83 - Phil Nevin 31 New York - 82-80 - Mike Piazza 21 Montreal - 68-94 - Orlando Cabrera 26 San Francisco - 90-72 - Barry Bonds 54 Arizona - 92-70 - Luis Gonzalez 37 See anywhere on that list where a player is doing lousy simply because his team is lousy? I sure don't. Phil Nevin had a damn fine year last year, despite being on a losing team. Jim Edmonds and Brian Giles were equally valuable, even tho Giles played for a team that lost 100 games. Some neat observations that I think we'd all agree with can be made from this. The San Francisco Giants, I think most people would agree, had awful pitching last year. Shortstop Rich Aurillia had 33 win shares - that's 2 MVP caliber seasons on one team, pretty good for a team with only 90 wins - unfortunately their best SP was Russ Ortiz with 15, which is bad for a #1 SP. By comparison, Randy Johnson had 26 win shares and Curt Schilling had 24, as you would expect, looking at their individual statistics. Why was Arizona better? Gee, let's see. Luis Gonzalez had 37 win shares and Schilling/Johnson combined for 50. They had 13 guys who gained credit for 10 or more win shares, which is pretty good depth. The Giants had 2 MVP type seasons (Bonds/Aurilia) and a good season from Kent (27 win shares) but after that things fall apart quickly. Ther had just 9 guys gain 10 or more win shares, and most of those shares were attributable to 3 players. Russ Ortiz was the 4th best player on the team. The 10th best player on the Diamonbacks was just some nobody named Jay Bell. Jason __________________ "I pretty much popped everything cold turkey. We were doing steroids they wouldn't give to horses." -- Tom House "I was very fortunate to have a pitching coach by the name of Tom House...Tom, I really miss those days that we spent in the weight room and out on the field working together." -- Nolan Ryan's HoF Induction Speech

06-06-2002, 01:33 PM	#31
Jason Moyer Hall Of Famer Join Date: Feb 2002 Location: Pittsburgh Posts: 5,108	I should also note that Win Shares very rarely tells you something that you don't know. When it does, it is interesting. For instance, you would assume the Phillies best player last year was Bobby Abreu, right? Well, not according to Win Shares: 1. Scott Rolen 29 2. Bobby Abreu 26 3. Jimmy Rollins 20 (by the way, this is an incredibly high total for a rookie - most rookie of the year winners have somewhere around 10-15 win shares - unfortunately, Jimmy picked up 20 win shares the same season Pujols picked up 29) Thinking about their hitting stats, that doesn't really seem right. But if you look on a closer level, here are their runs created and outs: 1. Scott Rolen 102 RC 417 Outs 2. Bobby Abreu 126 RC 454 Outs 3. Jimmy Rollins 89 RC 503 Outs Looking at it in terms of RC/12 outs: 1. Scott Rolen 2.94 2. Bobby Abreu 3.33 3. Jimmy Rollins 2.12 That's pretty close. Jimmy is a full RC/12 behind, but Bobby wasn't that much more productive than Scott was as the plate, in terms of actually creating offense for his team (and offense is measured in runs right? homeruns and batting average don't win games, creating runs does). Throw in the defensive win shares, and this is a no brainer. Whether you use win shares or just watch baseball, you know that Scott is the best defensive third baseman in baseball right now and over the last 4-5 years, if not ever. Bobby is a good right fielder, but he's not the outfield equivalent of Scott Rolen. Likewise, Jimmy gets a boost because he was much better with the glove than even the Philadelphia scouts and management had expected (Bowa and several analysts even said this several times last summer, so it's no surprise that Win Shares backs this up). Even something in Win Shares that seems kinda odd, at first glance, comes back looking pretty good when you sit and analyze it. Jason __________________ "I pretty much popped everything cold turkey. We were doing steroids they wouldn't give to horses." -- Tom House "I was very fortunate to have a pitching coach by the name of Tom House...Tom, I really miss those days that we spent in the weight room and out on the field working together." -- Nolan Ryan's HoF Induction Speech

06-06-2002, 02:18 PM	#32
Jason Moyer Hall Of Famer Join Date: Feb 2002 Location: Pittsburgh Posts: 5,108	</font><blockquote><font size="1" face="Verdana, Arial">quote:</font><hr /><font size="2" face="Verdana, Arial">Originally posted by mtw: <strong>Here's where the problem with Bill James concepts, such as Win Shares: there are just too many factors to consider. On a team composed of entirely of right-handed flyball pitchers, a shortstop's value would be less than on team of all left-handed groundball pitchers. Also, how good are a team's pitchers at holding runners? That would affect a catcher's ability to prevent steal attempts and throw out basestealers. Also, a good defensive catcher not only throws runners out, he prevents them from running in the first place. What's the value of that, and how does it relate to the frequency of steal attempts if that team had an average catcher or if the pitchers were terrible at holding runners? Also a good defensive catcher can prevent wild pitches and passed balls. What's the value of these "non-events" and how much less valuable is a catcher, if at all, if the pitchers have great control? What value on an OF throw to the right base, or hitting the cutoff man, to prevent runners from advancing? How much impact do these "non-stats" have on a game's outcome? Can Bill James tell me? </strong></font><hr /></blockquote><font size="2" face="Verdana, Arial">Yes. Have you read Win Shares? All of the factors you list are taken into account in his system. This isn't some 5 line formula we're talking about here. The formula itself takes up pages 1 to 77 in the book, and makes adjustments for everything that it is possible to adjust for. Team has a lot of left handed pitchers? The system assumes your third baseman should receive more chances and your first baseman should receive fewer independent chances. A lot of flyball pitchers? Adjusts for that too. A first baseman who prefers to flip to the pitcher instead of just tagging the base himself, making his stats look more impressive? Adjusts for that too. A catcher who plays a lot of innings and allows fewer stolen bases than he would be expected to allow? Adjusts for that. A team with a groundball staff should have how many double plays assuming a certain number of balls were put in play against the pitcher? Adjusts for that. Jason __________________ "I pretty much popped everything cold turkey. We were doing steroids they wouldn't give to horses." -- Tom House "I was very fortunate to have a pitching coach by the name of Tom House...Tom, I really miss those days that we spent in the weight room and out on the field working together." -- Nolan Ryan's HoF Induction Speech

06-06-2002, 03:03 PM	#33
OldGiants Registered User Join Date: Apr 2002 Posts: 211	Thanks for all your posts, Jason. I'll definitely have to do more than browse the book. I'm not convinced about win shares, but I'll read more.

06-06-2002, 03:13 PM	#34
Jason Moyer Hall Of Famer Join Date: Feb 2002 Location: Pittsburgh Posts: 5,108	</font><blockquote><font size="1" face="Verdana, Arial">quote:</font><hr /><font size="2" face="Verdana, Arial">Originally posted by OldGiants: <strong>Thanks for all your posts, Jason. I'll definitely have to do more than browse the book. I'm not convinced about win shares, but I'll read more.</strong></font><hr /></blockquote><font size="2" face="Verdana, Arial">I don't mean to be redundant but I do want to say that: 1. I don't agree with them 100% (mostly because I'd like to see the data he's basing the intrinsic weights on) 2. Because so much of the data has to be figured by estimating, they aren't completely accurate on a historical level. I believe James says that a difference of 3 or less is not worth arguing over because of the estimations that are required in the system. Hopefully this work will mean that STATS at least will track more meaningful stats (independent putouts by first basemen, non-pitcher assists by first basemen, independent putouts by catchers, runners on first base, etc). That way, at least in real baseball, we could calculate win shares without estimating anything. Jason __________________ "I pretty much popped everything cold turkey. We were doing steroids they wouldn't give to horses." -- Tom House "I was very fortunate to have a pitching coach by the name of Tom House...Tom, I really miss those days that we spent in the weight room and out on the field working together." -- Nolan Ryan's HoF Induction Speech

06-07-2002, 03:50 AM	#35
mking55 All Star Starter Join Date: Jun 2002 Location: Boading, China Posts: 1,249	Jason, Did you ever finish your WinShares spreadsheets? I am curious to try them out. Although I am not convinced of it's validity, I do admit that I am intrigued by the whole idea. I saw that you posted a couple of days ago that you were going to make them available for download, but I don't see the link. Is there? Thanks. __________________ Why do you live in Boading? Somebody has to. http://adventuresinjurong.wordpress.com/

06-07-2002, 04:27 AM	#36
eneubaue Minors (Double A) Join Date: Dec 2001 Posts: 156	Jason, Good points on Win Shares - I, too, have been an avid reader/follower of James' work since 1981 (I have a copy of one of his early abstracts he sold via mail-order). I think people should be aware that James' himself said that his Win Shares system is a first attempt to try to account for all elements in baseball (batting, fielding, pitching) in trying to evaulate/analyze players and how much each player contributes to his team winning. Once others start to play around with his system and make modifications/improvements, then I think we'll begin to see some of the benefits of his Win Shares system. I'd like to see the folks over at the Baseball Prospectus get their hands on it and run with it, making changes/modifications as necessary. I, like you, don't agree with everything James' has to offer (what's up with his modifications on the Runs Created formulas?), but at least he has provided a place to start.

06-08-2002, 05:03 PM	#38
Jason Moyer Hall Of Famer Join Date: Feb 2002 Location: Pittsburgh Posts: 5,108	My biggest overall beef with win shares, the more think about it, are the defensive adjustments. I think defensive win shares are absolutely amazing at showing who is the best player. But I'm not sure I agree that it shows who contributed more. For instance, let's say you have Steve Garvey at first on one team and Bill Buckner on another. Ok, we know Garvey is a better first baseman, altho Buckner is still good. However, if the conditions they're playing under cause the first baseman of Bucnker's team to have a lot more chances than Garvey, shouldn't Buckner get more Win Shares? I mean, you can have a defensive wizard playing shortstop for instance, but on a flyball team maybe he isn't contributing as much as your so-so center fielder. It's weird. I almost think that defensive win shares are great for evaluating a player's talent, but that the total baseball-style linear weights (for defense) are almost better at gauging actual win contribution... Am I messing something up logically here? If I were on the SABR list I'd ask them. Jason __________________ "I pretty much popped everything cold turkey. We were doing steroids they wouldn't give to horses." -- Tom House "I was very fortunate to have a pitching coach by the name of Tom House...Tom, I really miss those days that we spent in the weight room and out on the field working together." -- Nolan Ryan's HoF Induction Speech

06-09-2002, 08:29 AM	#39
inverarity Minors (Single A) Join Date: Mar 2002 Posts: 95	Great thread - I just want to throw in another YES to Jason's question of whether anyone would want his win shares spreadsheets. Definitely! Jason, are you still considering making them available? <small>[ 06-09-2002, 02:32 PM: Message edited by: inverarity ]</small>

06-09-2002, 03:36 PM	#40
Jason Moyer Hall Of Famer Join Date: Feb 2002 Location: Pittsburgh Posts: 5,108	Yes, I apologize it's taking so long. Unfortunately job hunting/GTA3 have been eating my time. I should get a chance to work on them tommorrow after I send out a few more resumes. Glad you're interested. Jason __________________ "I pretty much popped everything cold turkey. We were doing steroids they wouldn't give to horses." -- Tom House "I was very fortunate to have a pitching coach by the name of Tom House...Tom, I really miss those days that we spent in the weight room and out on the field working together." -- Nolan Ryan's HoF Induction Speech