PDA

View Full Version : Bill James' The Politics of Glory


538280
04-16-2007, 06:47 PM
I read this book in the summer of 2005 (a little less than two years ago), and lately I've been re-reading parts of it. I like the book, and I generally agree with what James says about arguments for the HOF, I agree that most of the arguments you hear on radio talk shows or on TV or in magazines and newspapers tend to be weak and poorly constructed. I think for people unfamiliar with arguing about players it is a very good read and can help you to sift out what is a legitimate argument from what is a bad one. I think it gives good ways to show evidence to support your claims rather than just throwing them out there. I also think the historical info in the book is great. It was great to read detailed accounts of the lives of baseball figures such as Joe Tinker, George Davis, Vern Stephens, and Phil Rizzuto among others. You come away from the book knowing a lot more about the players as people as well as their HOF credentials.

But I also think the book has a number of flaws which aren't mentioned so often. Upon rereading it, with more knowledge than I used to have, I find myself thinking that many of the methods and arguments he makes in the book aren't particularly useful and can lead you to believe an undeserving candidate is worthy perhaps even more so than the other way around. I also think a lot of the arguments he uses himself in favor of or against a candidate aren't particularly meaningful and he often makes broad general statements (Drysdale benifitted from the environment he played in) rather than taking steps to quantify exactly how much he was helped by those conditions.

To begin with I'll talk about the methods a little bit. I know that I've criticized them enough so many of you are probably already familiar with my problems with them, but of the different statistical tools James introduces in the book to see if a player is worthy or not of the HOF, I don't find any of them to be particularly useful. Black Ink is 1. Primarily weighted on the triple crown stats and other counting stats such as hits and runs which just are not good ways, by themselves, to try to determine the true worth of a player. OBP isn't even included in Black Ink. They also deal with league leaderships which are very arbitrary in nature. You may actually do better in a category one year and not lead the league yet do worse another year and lead the league. Babe Ruth won the batting title with a .378 BA yet didn't when he hit .393. Comparing players to the league average is a formal methology of determining a player's value; that should always be done in preference to looking at league leads. We have many things readily available which do that and it's not hard to do anyway so I don't think it's appropriate to focus on league leads at all. I also feel that this method reinforces what many think is a big problem with the voting from the BBWAA, which is that they often don't realize the inherent value of offense from a key defensive positoin, where that player is making defensive contributions as well, vs. from a corner OF or 1B. This methods doesn't take into account defensive ability or position.

HOF Standards are almost entirely based on counting stats which again I think totally miss the point of what players do to help their teams win games for the most part. I'm almost asking Bill James, Bill, you created measures such as runs created, you know that statistics such as OBP and SLG far better model run scoring than the triple crown stats (and RBI which is situational, dependant on those around you and your lineup spot), why don't your measures to determine HOF worthiness fall more along those lines-with things that you yourself have frequently shown which really equal victories in the standings? They also are based totally on raw numbers, with no attention paid to the environment in which a player played. Because of this players from hitters' eras like the 20s, 30s, 90s, and 2000s come out overrated by this method and players from pitchers eras like the 00s, 10s, 60s, and 70s are underrated (generally). This method does factor in position but ignores defensive ability, which is a problem.

Similarity scores are subject to many of the same criticisms as HOF Standards. They are based only on career counting stats which aren't really good indicators of actual value, and they totally ignore how great a player was at his best. They have no era adjustment at all. On BBRef you can see each player's top 10 most Similar players, you can see each player's OPS+ there. It is clear from this that the similarity lists created for many players form low run environments are not accurate. In his Keltner List chapter James lists Tony Oliva's most similar players as if it is a negative to his HOF case, because the list of those players aren't impressive. Oliva in his career had a 131 OPS+. The average of his comps is 114. It is clear that the comp list is not fair to Oliva, because Oliva played in a pitcher's era, the 60s an 70s. This is just one example of how those methods can (and do), lead you to false conclusions. If the system was put together with context adjustmetns perhaps the system would compare to modern outfielders like Kirby Puckett (124 OPS+), who in the context of their time, are much more similar to Oliva as hitters than players like Garret Anderson (105 OPS+), and Jeff Conine (109). I normally don't support Oliva as a HOFer anyway, so I'm not making his case, I'm just showing how the method is very flawed and it even leads James to what IMO is a false conclusion.

I had thought for a long time that James really only created these "metrics" to try to predict who would have the HOF as many people have said that since I read the book. I think it is clear from reading it over again that that was not his intent. The only one that was created to predict how the BBWAA would handle the player, rather than their actual worthiness, was the HOF Monitor. Here is a quote from the book, pg. 68:

"The monitor is a amoral system, it is concerned with what does happen in the voting, not what should, and thus it gives various accomplishments a somewhat illogical weight, to make players' records match up with the voting. Getting 200 hits in a season, for some reason, has always impressed Hall of Fame voters (or at least the players who have done that have always done well in HOF voting), so that gets a comparatively heavy weight. Driving in 100 runs or scoring 100 runs appears to mean less, so it gets less weight. This isn't my judgement, it just reflects the voting.

The Standards List doesn't make any attempt to predict the voting, it just evaluates the players' records in a kind of a common-sense way, and then centers the results at 50 percent.

To state this another way, it would disturb the design of the HOF Monitor if a player who actually is in the HOF, like Lloyd Waner or George Kelly, had a very low total, while a contemporary who isn't had a high total. If that happened, I would try to find some way to adjust the system so that the selected player had a better total. The Standards List doesn't care that there are players with 50+ totals who are out and players with 25 who are in, that's just the breaks of the voting."

From that I think it is clear that James created a system such as HOF Standards trying to absolutely not reflect the actual HOF selection process, he basically just directly states that there. I still am wondering why James did create methods like that though that are so flawed for this book. In his other books he creates methods such as WS or RC which really do reflect a player's contributions to their teams winning and scoring runs. That is what the game is all about, that is what should, IMO, become part of whether deserves the HOF or not. Not a big mess of counting stats thrown together to create a totally arbitrary number that really means very little about a player's value and doesn't reflect it in any way. I think we've seen better from James and I certainly have. I still think all the methods created in that book are overused here (though it has been getting better). I don't think they're useful at all.

I also think some arguments made in the book are not particularly useful or convincing. James spends a lot of time ripping those arguments commonly made by others (and I think his criticisms there are legit), but he makes many shaky arguments himself IMO, particularly in his chapter about Don Drysdale. He makes an argument for and against Drysdale, to show both sides of the story, and I think both arguments aren't particularly well constructed and contain many irrelevent points. He makes very general statements, that strikeouts help Drysdale, that the team didn't score many runs for him, that his won-lost record isn't accurate in his actual value, a lot of general statements, but without a lot of data to support that those things really matter or are even true or exactly how much Drysdale benifitted from those things. He also goes on and on about how Drysdale actually pitched better than pitchers with better won-lost records. But to me that is a very flawed argument, just because these pitchers had better won-lost records than Drysdale with higher ERAs doesn't mean that Drysdale deserves those records-and James seems to imply that. On the other side, James makes arguments in response to that Drysdale struck out a lot of batters than pitchers like Sam McDowell did as well and they're not in the HOF. Yes, Bill, but that's the point; the point is that a lot of Ks are a point in Drysdale's favor, not that every pitcher who struck out a lot of batters should also be in the HOF.

Another thing is James' almost total reliance on won-lost records in evaluating pitchers. He goes on and on about what is a HOF caliber won-lost record and even creates a system called Fibonacci Win Points to put won-lost records on a different scale given winning percentages. I think that won-lost records are almost useless. I can't think of one thing won-lost record attempts to quantify that isn't quantified better by something else. And the fact we put wins and losses on pitchers indicates that pitching is mostly responsible for why teams win games but it's not. Pitchign is probalby about 40% of why teams win games. Scoring runs and giving up runs are both half, and pitching doesn't get all the credit for preventing runs becaue fielding actually factors in, so a pitcher has about 40% control by himself of the game yet he's supposed to be responsible for the whole 100% by the logic of won-lost records. They're useless when you also have other things and dont' deserve a big piece in any evaluation of pitchers IMO yet James refers to them probalby more than anything else regarding pitchers Again I think, by his other writing, that James realizes this, yet he seems not to in the this book.

Overall, I like the book for the other things I mentioned, but I think it has a lot of negatives as well that are often ignored. I've seen a lot better from James.

iPod
04-16-2007, 09:52 PM
You gotta remember man, he wrote that book in 1994. He himself has done a lot of growing since then.


I also think some arguments made in the book are not particularly useful or convincing. James spends a lot of time ripping those arguments commonly made by others (and I think his criticisms there are legit), but he makes many shaky arguments himself IMO, particularly in his chapter about Don Drysdale. He makes an argument for and against Drysdale, to show both sides of the story, and I think both arguments aren't particularly well constructed and contain many irrelevent points. He makes very general statements, that strikeouts help Drysdale, that the team didn't score many runs for him, that his won-lost record isn't accurate in his actual value, a lot of general statements, but without a lot of data to support that those things really matter or are even true or exactly how much Drysdale benifitted from those things. He also goes on and on about how Drysdale actually pitched better than pitchers with better won-lost records. But to me that is a very flawed argument, just because these pitchers had better won-lost records than Drysdale with higher ERAs doesn't mean that Drysdale deserves those records-and James seems to imply that. On the other side, James makes arguments in response to that Drysdale struck out a lot of batters than pitchers like Sam McDowell did as well and they're not in the HOF. Yes, Bill, but that's the point; the point is that a lot of Ks are a point in Drysdale's favor, not that every pitcher who struck out a lot of batters should also be in the HOF.


I think you missed the point. He was trying to make arguments that were essentially fair, but that, when you read them, you could tell he was intentionally being a bit partisan. He was in character, so to speak. The two massive essays were, in a sense, compilations of all the rational to semi-rational cases you hear for and against Drysdale, and he doesn't begin to digest them until the third massive essay.


Another thing is James' almost total reliance on won-lost records in evaluating pitchers. He goes on and on about what is a HOF caliber won-lost record and even creates a system called Fibonacci Win Points to put won-lost records on a different scale given winning percentages. I think that won-lost records are almost useless. I can't think of one thing won-lost record attempts to quantify that isn't quantified better by something else. And the fact we put wins and losses on pitchers indicates that pitching is mostly responsible for why teams win games but it's not.


Page 271:
"Well, I don't want to oversell the method. It doesn't evaluate anything except the won-lost record, so there could be a thousand reasons why one pitcher with 250 points might not belong in the Hall of Fame, while another pitcher with 150 might deserve to be there."

Page 401:
"For the sake of clarity, I am not arguing that Don Drysdale's won-lost record proves that he should not be in the Hall of Fame. What I am saying is that his won-lost record is not, in itself, a Hall of Fame type of record."

I read the Fibonacci method as being like the condoms in school argument: don't use W/L, but if you're going to use it anyway at least have some sort of order.

Honus Wagner Rules
04-16-2007, 09:56 PM
Chris, how many times do I have to cover this for you? The reason Black Ink is weighted heavily in the triple crown stats is because that is what the HoF voters tend to generally focus on first. James is not saying that is the right way to look at a player but that's how the voters view players. What I keep telling and you don't seem to understand is that voters don't care about "advanced metrics". They don't look at WS, WARP3, PCA, RC, Linear Weights, Secondary Average, VORP, etc. They only look at traditional stats and most likely will for the foreseeable future. This is what James is getting at with Black Ink. Winning a batting titles will ALWAYS impress the HoF voters. How many times have you heard Tony Gwynn called the best "pure" hitter of his time? "Pure" is a code word for high batting average and not much else offensively. If a player has 4-6 batting titles he will get sustantial support even if he has no power, no speed and medicore defense. Also, is it better to lead the league in BA with a .378 BA instead of hitting .393 and not lead the league? In terms of the HoF voters it is better.

jalbright
04-17-2007, 06:59 AM
If James wrote that book now, I'm sure he'd use some work based on Win Shares--but he hadn't done that work back when the book was written. It's an open guess how much he'd rely on some of the measures he used/introduced in that book. Certainly, he'd rely more on Win Shares than those.

Personally, I think some of the measures like the inks and HOF standards contain useful information that can help an evaluation of a candidate in close calls. That said, I recognize the flaws in them (though every statistical measure has some flaws) and give them comparatively little weight.

Jim Albright

538280
04-17-2007, 07:58 AM
Chris, how many times do I have to cover this for you? The reason Black Ink is weighted heavily in the triple crown stats is because that is what the HoF voters tend to generally focus on first. James is not saying that is the right way to look at a player but that's how the voters view players. What I keep telling and you don't seem to understand is that voters don't care about "advanced metrics". They don't look at WS, WARP3, PCA, RC, Linear Weights, Secondary Average, VORP, etc. They only look at traditional stats and most likely will for the foreseeable future. This is what James is getting at with Black Ink. Winning a batting titles will ALWAYS impress the HoF voters. How many times have you heard Tony Gwynn called the best "pure" hitter of his time? "Pure" is a code word for high batting average and not much else offensively. If a player has 4-6 batting titles he will get sustantial support even if he has no power, no speed and medicore defense. Also, is it better to lead the league in BA with a .378 BA instead of hitting .393 and not lead the league? In terms of the HoF voters it is better.

I don't think that was James' intent. His intent for the HOF Monitor was to model the trends of the voting, but I don't think that was what he was trying to do with other methods introduced in that book, such as HOF Standards and Black Ink. I already showed the quote for HOF Standards, and he doesn't come out and say that directly for Black Ink but I find no part in the book where he references that he only put it together the way he did to try to model the BBWAA's voting. He says it is a "counterpoint" to HOF Standards which he says are clearly NOT meant to model the voting. He refers to it as a method to tell whether a player is a HOFer (not to predict if a player will be). Do you have any quotes or anything that show his intent was to model the voting because from reading the part of the book where he introduces it I don't get that impression at all.

I am rather puzzled as to why James would come up that method as well though. I think his other writings show that he could do much better.

If James wrote that book now, I'm sure he'd use some work based on Win Shares--but he hadn't done that work back when the book was written. It's an open guess how much he'd rely on some of the measures he used/introduced in that book. Certainly, he'd rely more on Win Shares than those.

James had already created runs created and offensive winning percentage by this point for his annual Abstracts and his Historical Abstract. I would have liked to see more use of those things, which do actually model what players do to help their team score runs and win games, rather than the methods he developed like Black Ink and HOF Standards. He never references OWP in the book and he references RC once in the Phil Rizzuto/Vern Stephens chapter.

KCGHOST
04-17-2007, 08:05 AM
Bill James, like all of us, has his weak moments. The Fibonacci thing was nothing more than a math exercise to see if he could make better use of the primary stat voters were using to put pitchers into the HoF.

jalbright
04-17-2007, 12:10 PM
James had already created runs created and offensive winning percentage by this point for his annual Abstracts and his Historical Abstract. I would have liked to see more use of those things, which do actually model what players do to help their team score runs and win games, rather than the methods he developed like Black Ink and HOF Standards. He never references OWP in the book and he references RC once in the Phil Rizzuto/Vern Stephens chapter.

The simple problem James had at the time was those measures didn't address defense. He knew his attempts at evaluating that side of the game were not very good, and approaching it your way would have left him vulnerable to criticisms he had gotten way too much before. The new measures didn't solve the problem, to be sure--but he was writing a different book. Besides, his first historical Abstract had already addressed things in those terms. I'm sure that he wanted to sell books, and repeating the previous book wasn't going to do that.

Jim Albright

538280
04-19-2007, 09:20 AM
The simple problem James had at the time was those measures didn't address defense. He knew his attempts at evaluating that side of the game were not very good, and approaching it your way would have left him vulnerable to criticisms he had gotten way too much before. The new measures didn't solve the problem, to be sure--but he was writing a different book. Besides, his first historical Abstract had already addressed things in those terms. I'm sure that he wanted to sell books, and repeating the previous book wasn't going to do that.

Jim Albright

You might be right about the selling books part. Perhaps he thought he had to come up with some sort of new measure to sell the book, even if it wasn't very good. I don't see how the new measures solved the defense problem though. None of the measures mentioned factor in defense except just primary defensive position.

Honus Wagner Rules
04-19-2007, 09:55 AM
I don't think that was James' intent. His intent for the HOF Monitor was to model the trends of the voting, but I don't think that was what he was trying to do with other methods introduced in that book, such as HOF Standards and Black Ink. I already showed the quote for HOF Standards, and he doesn't come out and say that directly for Black Ink but I find no part in the book where he references that he only put it together the way he did to try to model the BBWAA's voting. He says it is a "counterpoint" to HOF Standards which he says are clearly NOT meant to model the voting. He refers to it as a method to tell whether a player is a HOFer (not to predict if a player will be). Do you have any quotes or anything that show his intent was to model the voting because from reading the part of the book where he introduces it I don't get that impression at all.
That's the impression I got. I don't have a specific quote but the main theme of the book, to me at least, was to make some sense of the HoF voting. Look at how Black Ink is weighted. The catagories that get four points are typically what the voters look at first. When I was akid growing up I watched al lot of baseball on TV. When a player came to bat their always showed the player's BA-HR-RBI and that's it. For a player like Ozzie Smith those stats tell you nothing about his value


I am rather puzzled as to why James would come up that method as well though. I think his other writings show that he could do much better.
I think James wanted to focus more on the history of the HoF and how the voting evolved through the years.



James had already created runs created and offensive winning percentage by this point for his annual Abstracts and his Historical Abstract. I would have liked to see more use of those things, which do actually model what players do to help their team score runs and win games, rather than the methods he developed like Black Ink and HOF Standards. He never references OWP in the book and he references RC once in the Phil Rizzuto/Vern Stephens chapter.
That's a good question. I think James was trying to give a reader a contemporary view of Stephens and Rizzuto as players focusing more on the players and how they were perceived when they were active. James mentioned how the Red Sox were viewed as a team of great hitters with poor defense and few intangibles. The Priddy story was also very interesting. I got to learn more about Joe Gordon as well. I never realized how truly a great ballpayer he was. If not for WWII and Yankee Stadium he could have been a top 5 all-time second baseman.

538280
04-19-2007, 07:55 PM
That's the impression I got. I don't have a specific quote but the main theme of the book, to me at least, was to make some sense of the HoF voting. Look at how Black Ink is weighted. The catagories that get four points are typically what the voters look at first. When I was akid growing up I watched al lot of baseball on TV. When a player came to bat their always showed the player's BA-HR-RBI and that's it. For a player like Ozzie Smith those stats tell you nothing about his value

How Black Ink is weighted is exactly my main criticism of it. It is clear if you read the book over that his intent with Black Ink and HOF Standards was not to predict the voters. I'm not going to show quotes for it anymore, but if you read what he says about HOF Standards he pretty much directly states that he's NOT trying to model the voting the tendencies. And he says a number of times that Black Ink is meant to be used with HOF Standards as a way to measure dominance while HOF Standards more looks at longevity. It is clear from what he says that his intent in creating those metrics was not to model the voting.

I think James wanted to focus more on the history of the HoF and how the voting evolved through the years.

He does focus a lot on the history of the HOF, those were some parts of the book I really did like.

Honus Wagner Rules
04-19-2007, 11:48 PM
How Black Ink is weighted is exactly my main criticism of it. It is clear if you read the book over that his intent with Black Ink and HOF Standards was not to predict the voters. I'm not going to show quotes for it anymore, but if you read what he says about HOF Standards he pretty much directly states that he's NOT trying to model the voting the tendencies. And he says a number of times that Black Ink is meant to be used with HOF Standards as a way to measure dominance while HOF Standards more looks at longevity. It is clear from what he says that his intent in creating those metrics was not to model the voting.
We all know Black Ink is NOT the best way to evaluate a player's career.


He does focus a lot on the history of the HOF, those were some parts of the book I really did like.
I found the Stephens/Rizzuto and Gordon/Doerr sections most interesting.

jalbright
04-20-2007, 02:26 PM
Chris,

If I've shared this with you before, I apologize for repeating myself. However, one reason I have some affinity for the Inks is that, when I first rated Japanese players, it was before win shares, so I had to come up with something. I used the inks as part of that process, and the results were actually pretty good. Certainly, it wasn't perfect, but it was useful, and at that point, I would have been hard pressed to do as well with what I had available to me. The arrival of "short-form" win shares represented an improvement in my ability to deal with the data, so I worked with that instead. Truthfully, though, the two lists were more similar than I would have expected. That experience taught me that these measures had some actual value, and all your criticism of the measures won't erase that practical experience.

Jim Albright

538280
08-15-2009, 07:48 PM
I've been rereading this book again a little lately. I don't think I had looked at it since I started this thread. I still think much of the statistical reasoning in the book is not the best at all (I've seen much better from James since and before the book was written in '94), but I do think the historical parts are very interesting. There are some arguments in the book that I just completely disagree with though and think James really misses the point in many cases.....like for example he says Pee Wee Reese and Phil Rizzuto are not comparable players, and then cites their career counting stats. I just find that to be a laughable argument. When people say Rizzuto and Reese are similar players they are talking the types of players they were, their skills. They were both good fielding SSs who were also solid but not great hitters at the position. They have similar hitting rates. Reese has more counting stats simply because he had a longer career, that doesn't mean they aren't similar players.

STLCards2
08-15-2009, 09:06 PM
The stuff about Drysdale seems the most outdated, as James finished the writing concluding that Drysdale is not a worthy HOFer. Every more recent, highly respected value-stat (WPA, WAR, Matt's DNRA marker, WSAB) all have Drysdale as an easy HOFer.

What drove me crazy was James would state that Drysdale was helped by his park, for example, and is overrated because of it (okay, some truth there), but then he would talk about Drysdales' hitting as being a smallish factor in his ranking. All data since points to Drysdale's offense being as much of a positive as his park was a negative. He talked very minimally about his massive IP accumulation in such a relatively short career too.

538280
08-16-2009, 08:55 PM
I think the Drydale argument is a result of what I talked about in the opening post-the extreme overreliance in the book on won-lost records. James mentions many times in the book how Drysdale's won-lost record to him was not HOF worthy and how he needed to do other things outside of that to make him a HOFer with that record. This is very ironic because in his annual Abstracts, of which I have many of them, James always wrote about how won-lost record shouldn't be payed much attention to in evaluations vs. other much more telling stats, but in this HOF book he seemed to place a lot of stock in won-lost records for pitchers. He also consistently gave weight to things like championships won by players as a credential which I don't understand at all.

Ubiquitous
08-17-2009, 09:57 AM
To be fair Drysdale got elected in 1984 and some of the head scratchers for election came after him. Then Bill wrote his book in 1993 and got it published in 1994. AFter writing this book Sutton, Niekro, and Bunning got sent in. After Drysdale's election those guys plus Hunter, Jenkins, Newhouser, and Perry got in.

Who looks like a HoF'er and who doesn't evolves over time based on who gets in.

ol' aches and pains
08-17-2009, 11:55 AM
I haven't read the book since it came out, but I seem to remember James stating that Drysdale's won/loss record was almost identical to Milt Pappas, in fact Pappas was actually better by a couple of games. And nobody in their right mind thinks Milt Pappas should be in the Hall of Fame, therefore Drysdale shouldn't be, either. A strange statement for the Godfather of Sabremetrics to make, it almost sounds like my logic. :laugh

Cougar
08-17-2009, 12:06 PM
I haven't read the book since it came out, but I seem to remember James stating that Drysdale's won/loss record was almost identical to Milt Pappas, in fact Pappas was actually better by a couple of games. And nobody in their right mind thinks Milt Pappas should be in the Hall of Fame, therefore Drysdale shouldn't be, either. A strange statement for the Godfather of Sabremetrics to make, it almost sounds like my logic. :laugh

You're only remembering the straw man...James then went on to demonstrate through a number of simulations that Drysdale accumulating that W-L mark with a number of "big seasons" (I forget precisely how he defined it; say 20 wins) is more valuable than Pappas' longer career with a bunch of solid but unspectacular 14-10 records.

KCGHOST
08-17-2009, 03:13 PM
For all its flaws "The Politics of Glory" is still seminal piece of writing on baseball. There aren't too many of us who aren't more knowledgeable for having read it.

538280
08-17-2009, 05:32 PM
For all its flaws "The Politics of Glory" is still seminal piece of writing on baseball. There aren't too many of us who aren't more knowledgeable for having read it.

I would agree with this, because of the great historical topics touched on in the book. I loved reading the facts and profiles of players such as Rizzuto, Jerry Priddy, George Davis, and Joe Tinker. I also think James did a great job squashing the logic of many illogical arguments made about players in the HOF (but also giving some of his own illogical arguments-albeit not as bad as the ones he criticized). However, as far as statistics are concerned I don't think this book is very good even compared to work James had done beforehand, I might be holding too high a standard as it is far better than what anyone on the BBWAA has ever done but I think the statistical measures such as Black Ink, HOF Standards and such just add to the thousands of statistics out there and they really cloud the picture (at least IMO) more than anything else.