Friday, September 17, 2010

Murray Chass Versus The Machines

I love picking on Murray Chass. His non-blog is one of my favorite sources of baseball discussion (in my head of course because I have no friends) and I finally bookmarked it, which means I plan on picking on him regularly. Murray is a wealth of baseball knowledge and a wealth of anti-statistics and computer hatred. I prefer to focus on the hatred, rather than his wealth of baseball knowledge. The title of the post (on the non-blog) Murray has written which I will cover here is "Winners Who Have Losers."

If this were a Bill Simmons article I would assume we are going on a 10,000 word trip where he thinly ties in celebrity couples who don't look matched up correctly with NBA teammate pairings that don't seem right either (Did I just give him a column idea. Yup). Since this is Murray Chass, I had to investigate further, then I didn't really understand how he didn't understand winning teams can have pitchers with losing records, and found a pot of gold at the end of the column about how computers suck.

First, let's hear his ruminations on why teams with winning records have pitchers that don't have winning records. How could this be?

This column is dedicated to A. J. Burnett because he is responsible for it. More precisely, his won-lost record gave me the idea for it. The idea centers on a question: how can a starting pitcher on a winning team have a losing record?

1. He's not a good pitcher who plays on a good team.

2. Poor run support.

3. Bad luck.

4. Shit happens.

Would anyone be shocked if Kris Benson joined the Yankees and then had a losing record? Would it shock someone if the Phillies signed Mark Redman and then he had a losing record? This would shock Murray Chass. He must not realize some pitchers just aren't very good. AJ Burnett isn't a bad pitcher, so it must be bad luck or poor run support.

Burnett, whom the Yankees lured to New York with an $82.5 million contract for the 2009 season, has a losing record, 10 wins and 12 losses. With four weeks left in the season, he has been the losing pitcher in 24 percent of the 50 losses the Yankees have.

By comparison CC Sabathia has been the losing pitcher in 10 percent of the team’s losses.

Using the "Chass-o-Meter," this determines that Sabathia is a 14% percent better pitcher than Burnett. See, the Chass-O-Meter doesn't really care about statistics so it just takes numbers that have nothing to do with each other and then uses them to determine a calculation. This is as much as Murray uses numbers except when he wants to calculate at what date the next Social Security check comes in the mail.

Sabathia has also been the winning pitcher in 22 percent of the Yankees’ victories while Burnett has received credit for 11.6 percent.

Chass-o-Meter adjustment required!

Sabathia may be a 14% better pitcher than Burnett, but Sabathia pitches well enough to win the game 10.4% of the time. All of this equals the conclusion that none of these statistics Murray Chass is quoting really mean anything that tells us anything interesting.

Burnett, though, is not the only starter with a losing record for a team with a winning record. He has plenty of company. In fact, seven of the eight division and wild-card leading teams are in those positions despite having pitchers with losing records, and the eighth team has a pitcher with a .500 record as a starter and recently picked up another loss in a 13-inning game.

But I thought every team winning their division had to have pitchers that had winning records? How can this not be true? If you don't use logic or your brain, this conclusion winning teams should have only winning pitchers would make sense. Doesn't every NFL team have two starting wide receivers with 1,000 yards? Doesn't every great NBA team have starters that average in double figures? (sarcasm done)

These seven contending teams are all teams that don't have four quality starters, so they will NEVER be World Series teams in the eyes of Joe Morgan.

The Atlanta Braves, who have been in first place in the National League East since Memorial Day, entered the weekend with three starters with losing records: Derek Lowe 11-12, Tommy Hanson 9-10 and Kenshin Kawakami 1-10.

Hmmm...well that is obviously because these starters suck and not because the Braves offense is the most frustrating offense in the entire history of the world. One day they score 10 runs and then they get shut out, then they score two runs, then get shut out, and then score 12 runs. I hate the 2010 Braves team. Regardless, I am sure these losing records are all a reflection of how bad these pitchers are and not a reflection of any run support they did/did not receive.

Also, Kawakami hasn't made a start in Atlanta since June. He's been out of the rotation and in the minors, so he really shouldn't count as a starter for the Braves.

“I can only say in the case of Kawakami,” Wren said, “it seems every team has a pitcher who doesn’t get support, and he’s that guy for us. I don’t know how to explain that but in his case he didn’t get support. Two years ago it was Jair Jurrjens for us. He’d give up one run and get beat, and he’d give up two runs and get beat.”

"Murray, please leave me alone. I am currently attempting to find another mediocre washed-up bat to add to the Braves lineup. I currently am looking to trade for Lance Berkman to back up Derrek Lee and Troy Glaus at first base. What? It isn't 2003?"

Indeed, according to Elias Sports Bureau, Kawakami, before his most recent start, last Friday against Florida, had one of the weakest run supports in the majors, 3.47 runs per game. But run support can be misleading.

As opposed to the mystery as to why winning teams should only have pitchers with winning records. Acting as if all winning teams should have pitchers with winning records, that's not misleading at all.

In that game, his first start since June 26, the Braves scored only one run, but Kawakami left after giving up 5 runs in 3 innings.

Well, that one cherry picked example proves that there is no way a team could ever not offer a pitcher enough run support to get him the win!

Kawakami has been mostly terrible this year, there's very little about that, but he has pitched a few starts where he hasn't gotten much run support. Mostly, he has been terrible though. Still, the point is that if Kawakami gives up 3 runs over 6 innings (which he has done multiple times), that is a 3.00 ERA, which should be good enough to win a baseball game. He doesn't get enough support for the bullpen to give up one run or for him to give up more than 3 runs.

Wren noted that since June 7, a span of 14 starts, Lowe had not allowed more than four runs in a game. His record in that time, though, was 3-7.

BUT YOU DON'T GET IT! ONE TIME KENSHIN KAWAKAMI GAVE UP 5 RUNS IN 3 INNINGS! THAT PROVES RUN SUPPORT IS MISLEADING MUCH MORE THAN YOUR 14 GAME SAMPLE SIZE!

Then Murray goes on to list the pitchers for contending teams that have losing records, as if this is a huge mystery that needs to be solved. Winning teams can have pitchers with losing records for a variety of reasons, none of which Murray Chass figures out here. Sometimes it is run support, sometimes the pitchers pitches poorly, and sometimes the pitcher just has bad luck.

Sometimes losing teams have pitchers with winning records. How does that happen? It just happens sometimes and there isn't one given reason for it, though run support is a factor that (much to Murray's chagrin) can't be overlooked.

Now let's get to Murray's fear of machines and computers. Later in this post (on this non-blog) Murray wrote about a new way to figure out postseason awards. Naturally, he hates this idea.

As secretary-treasurer of the Baseball Writers Association, Jack O’Connell is responsible each year for finding 240 BBWAA members to vote on the four awards the association gives each November.

(Begins searching nursing homes in Florida for viable candidates, settles on using voters who think Lou Piniella still plays for the Yankees and WHIP is something you used to be able to discreetly buy in the sex shop around the corner no one in the neighborhood talked about but everyone frequented.)

As the number of newspapers dwindles and their impact fades, O’Connell finds his task increasingly tougher.

Because we all know the only qualified voters naturally come from a dying format that is losing influence. There's no way any qualified voter could come from anything but a newspaper.

But now comes a fellow who is not only going to make O’Connell’s job easier, but he will also eliminate it altogether.

Why would anyone want to make a job easier or eliminate the need for that job altogether!? What is this? The Industrial Revolution?

His name is Sean Forman, and he is the creator of the Web site, Baseball-Reference.com. He also has a method of selecting the most valuable players and Cy Young award winners that he believes is far more reliable than the writers’ voting.

He puts 10 pitchers names in a hat, has a monkey draw each name out of a hat, lay the name on the ground and whichever name the monkey defecates on last is the Cy Young award winner?

Forman presented his system in two articles that ran on consecutive Fridays in The New York Times,

This is the part where Murray Chass doesn't link the two articles that ran in the New York Times because that is how old-school baseball reporting works. You discuss and criticize someone else's work without giving them credit for the work or allowing the reader easy access to make up their own minds on whether what you are saying is valid or not. Why post a link to the source when a person can go Googlize it on the Interwebs himself?

I will provide the links for you. Here and here.

which apparently is jumping on the statistics bandwagon as a way of trying to lure younger readers to the fading newspaper.

I like how Murray is acting like the New York Times has a deal with Playboy to show exclusive photos before the magazine comes out or is having a porn star write a weekly sex column. It's so seedy of an old newspaper trying to stay in business to use new methods to present information to the reader. The current methods of using old-school analysis is working so well that readers just can't get enough joy out of canceling their newspaper subscriptions. Stick with what isn't working! That's the way to success and the best way to stay in business!

Part of the idiocy of the non-statistics, old-school mainstays that abhor anything beyond home runs or RBI's to judge a player is that even giving a "statistics-based" calculation a forum to be heard in public is "jumping on the statistics bandwagon." Even given lip service to the idea statistics can be used to figure out anything baseball related draws an angry and dramatic response from the old-school mainstays.

It is as of Murray is so afraid the statistics will influence the public they should not even be heard and should only be whispered about in the backrooms of darkly lit night clubs (or a basement in the house of someone's mom) where they can't have any pull over the reading public. It is not that Murray Chass doesn't like statistics, he doesn't even think a newspaper should give those who use statistics that aren't Murray-approved a forum for their opinion. It's like a little anti-statistics Nazi Germany and Murray is glad to play the angry dictator with the mustache who wants all seeds of rebellion pushed down immediately.

Without getting into details,

We don't get into details because Murray doesn't understand the details. That's what this all stems from. Murray doesn't understand statistics so he naturally fears them and just assumes they are bad. They threaten everything that he knew and made his living off of. If there is a new way to determine which player should win the Cy Young or MVP award that may mean Murray and those who voted for those awards in the past could be wrong. It's all about ego and the fear of being wrong.

Forman presents a formula by which m.v.p. and Cy Young candidates can be ranked. The result is a number that is assigned to each player – Robinson Cano 6.3, Evan Longoria 6.2, Miguel Cabrera 6.0, using A.L. m.v.p. as an example.

Integral to those numbers is something called WAR, which stands for wins above replacement. What replacement? A replacement player, of course, but he’s mythical.

He is mythical, but he is based on a normal Triple-A call up who is far below league average in all aspects of his skill level. This replacement player is used as the standard to compare all other players, so the criteria in comparing the players doesn't change. So the replacement player is mythical, but based on a mathematical formula that Murray Chass doesn't understand, much like the idea that human beings can be completely unbiased and objective when choosing postseason awards.

Statistics zealots apparently love to deal with mythical or hypothetical players.

"Statistics zealots?"

As of believing one side of a discussion on an issue should get absolutely NO FORUM to discuss their side isn't zealotry in itself.

The problem for those of us who prefer dealing with reality and actual human beings is we can’t buy into the idea of using mathematical formulas instead of real players.

Even using mathematical formulas to compare players is comparing the real players to each other. All of the real players are being compared using the exact same mathematical formula. No one is comparing Albert Pujols to Fake Player A, while Joey Votto is being compared to Fake Player B. Real players are being compared to real players...just through a mathematical formula to determine the statistical differences in the players instead of using human bias and preference to compare players.

However, given that many BBWAA voters seemed to buy into the formula stuff in the Cy Young voting last year, it may not be long before they vote for all of the awards on the basis of WAR (no intangibles needed).

I am pretty sure this is a Zack Greinke reference. Just like Murray Chass is too stupid to understand how winning teams can have losing pitchers, there's no way in his mind the best pitcher in the American League can't be on the best team in the American League. It's inconceivable to Murray's brittle and apparently non-three dimensional thinking mind this could happen.

The best scientists in the world aren't at the biggest company in the world. The smartest person in the world isn't in the smartest country in the world (I don't think). The tallest person in the world doesn't come from the country with the tallest average height of its citizenry. It's possible the best pitcher isn't from the best team in the American League.

And another thing..."Intangibles" are defined as,

"1. Incapable of being perceived by the senses.
2. Incapable of being realized or defined."

So why would anyone take something that can't be defined or perceived by the senses into account when measuring how valuable a player is to this team? It can't be defined, so the idea of defining it in terms of whether Player A is more valuable than Player B based on any type of intangible not only goes against the definition of the word, but is pointless.

O’Connell can simply ask Forman to e-mail him the final WAR numbers, and then he can stand on the dais at the New York baseball dinner each January and present the m.v.p. and Cy Young awards to a computer.

That was really the last sentence of this post (on this non-blog). I don't think Sean Forman was saying these WAR numbers should be used primarily to determine the MVP and Cy Young winners, but they are a TOOL to be used.

That's another thing about the idiocy of the non-statistics, old-school mainstays that abhor anything beyond home runs or RBI's to judge a player. They think any suggestion to use a statistic as a tool is actually an attempt to use statistics to take over the entire game of baseball and replace actual human beings with computers to play the game of baseball.

Murray Chass treats statistics like they are a disease that will infect baseball if used too frequently and ruin everything good about the sport.

The entire disconnect between new-fangled statistics users and the old-guard like Murray Chass is seen in the last sentence. If WAR was the only way these awards were chosen, Murray says they should present the Cy Young and MVP award to a computer when the award would actually still go to a human being. Sabermetric-inclined individuals use statistics to evaluate baseball players, they do not use statistics to replace actual, living baseball players. Murray still for some reason thinks computers are going to replace the actual baseball players when statistics are merely a different way to evaluate actual baseball players. For some reason, this realization still hasn't gotten into Murray's head yet.

10 comments:

  1. the funny part about this, at least for me, is that what do the voters from the BBWAA use to make their votes? do they watch every player play every game and then decide, "hey, that albert pujols was the best player in baseball this year. i bet he had like 30+ hr's and batted over .300 (not that i'm gonna look).

    no, they use STATISTICS. there's no way you can watch all the top players play all 162 games. so they look at their HR's, RBI's, BA, W, and ERA. anything wrong with that? no, of course not.

    what they old guard is missing is that stats have evolved. WAR isn't the be all end all stat. it uses imperfect defensive metrics to help determine overall value. but it is a tool. it's better than watching 3 games played at short by jimmy rollins where he had an error and didn't get to a ball in the hole. UZR takes a full season's worth of data, even if the data is imperfect.

    to not use wOBA (or OPS) to help determine who was had the best bat in the league is choosing to be uniformed. same with some of the baserunning metrics. same w/ UZR. and go ahead and use your eyeballs some for good measure.

    it's amazing how intentionally ignorant someone like Murray Chass or Joe Morgan chooses to be.

    ReplyDelete
  2. Paraphrasing myself & Keith Law's response, "Statistics describe performance that happens on the field. Except errors."

    ReplyDelete
  3. The whole replacement player concept seems to be so scary to anti-statsists. It needs to be explained to them that they already make this comparison, but WAR makes it more precise. When Chass tries to pick the most valuable player, what does he do? At some point in the analysis he has to decide what would have happened to the team without the player - otherwise how can you gauge his value? That's what WAR does, except that it sets a statistical baseline to which the player is compared. So instead of comparing Pujols to a hole in the lineup, we compare him to what he'd likely be replaced with. The difference is Pujols' value, and everyone makes this analysis the same way. We just put numbers on it.

    ReplyDelete
  4. 3 runs in 6 innings is a 4.50 era. :)

    Chase is obviously a card carrying member of Joe Morgan's "Wins are all I need to determine how good a pitcher is" club.

    Of all people, Simmons ahs a good point about who they are letting vote for sports awards nowadays. He mentions that he writes about the NBA, and watches more games then the guys voting for the awards, so why can't he even be considered as a voter? They let the tv guys for each team vote, and those guys might be lucky to see 10 games a year that doesn't involve their team. At some point, the sports awards are going to have to bring in bloggers and others who cover the game from what is not an old style format.

    ReplyDelete
  5. how can a starting pitcher on a winning team have a losing record?

    Cole Hamels. 3.01 ERA and a 1.16 WHIP. He's 11-10. That's pretty fucked up.

    In fact, seven of the eight division and wild-card leading teams are in those positions despite having pitchers with losing records

    Most of the division winners have solid rotations. Now, someone has to be accounting for the team's losses. I really don't care who you are, if your fifth starter is around .500 you should be happy.

    But run support can be misleading.

    How? How in the fuck can run support be misleading? If a pitcher gets 3 runs a game in support he's not going to win nearly as many games as an equally talented pitcher who gets 4 runs a game.

    Kawakami left after giving up 5 runs in 3 innings.

    This is not a flaw with run support, this is why it's stupid to only look at one statistic. Just because a pitcher gets low run support doesn't mean he has 1001289983 losses because the team can't score. It's why there are a shitload of statistics that paint a complete picture.

    What replacement? A replacement player, of course, but he’s mythical.

    It's called a fucking baseline you moron.

    Murray does realize that most things in life are actually given compared to a baseline reading right? Like money. Money has no intrinsic value until a baseline tells you how much that's worth. If I gave you 48 pesos would you have any idea what the value of 48 pesos is? Likewise 5 bucks is only valuable in so much as what that 5 can get us.

    The problem for those of us who prefer dealing with reality and actual human beings is we can’t buy into the idea of using mathematical formulas instead of real players.

    They should change the stat to WADE: Wins Above David Eckstein.

    If I told you a guy hit .290 with 30 HRs and 100 RBI how would you know if he's the MVP or not? You can't, because all those statistics are meaningless without comparisons. That's what WAR does, it makes those comparisons and gives you a mathematical result.

    no intangibles needed

    If you vote for a guy on "intangibles" you deserve to be kicked in the crotch. Repeatedly. David Eckstein has "intangibles," Albert Pujols is really fucking good at baseball.

    To add to Matt and HH, statistics can't tell you everything. They're simply tools to be used. No one needs statistics to tell you that Albert Pujols is really good at baseball. They just give you evidence as to what he's good at and at what rate he's good at it. Statistics also tell you that Jimmy Rollins sucks (which he does) despite the fact that the media thinks he's awesome.

    ReplyDelete
  6. Matt, they use their viewing of 1-2 games they see of each player, check how many HR, RBI's each player has and then check to see where his team is in the standings. I am afraid that's how they do it.

    Exactly, WAR is a tool. It isn't perfect and it doesn't represent anything perfect. No one is asking them to bow to the alter of statistics, but Murray should at least learn them and then dismiss them. I would have no problem with that. If he has a reason to dismiss the statistics being used, go for it.

    HH, you put that in layman's terms but what sucks is that Murray Chass doesn't even want to hear it. He would rather not use the statistics because they represent something it seems he is afraid of. WAR is a baseline, that's it. I don't even know if I like it that much, but I understand it enough to know whether I like it or not...which is more than Murray can say.

    Martin, I'm an idiot. I hit the wrong button on my calculator it seems. I no good at math.

    I don't agree with some things Bill Simmons says, but I agree 100% with that. Not that Bill Simmons should get a vote because he has a Boston bias, but at least he is more informed and watches the games. Half these sportswriters don't have time to watch the players play, so how can they vote for them or not vote for them?

    Rich, that's called bad luck for Hamels, which apparently Murray doesn't believe in.

    Run support could be misleading, but the bottom line is that a pitcher who gets 7 runs per game of run support will most likely win more games than a guy who gets 5 runs per game. It's not rocket science, it just shows the guy with 5 runs of support has less room to pitch poorly.

    If a good team has a 5th starter that goes .500 that's still a very good team. The worst starter wins as many games as he loses.

    What's interesting is that Murray doesn't use a baseline to compare players because he looks at three statistics. HR, RBI, BA. That's probably it. It's hard to understand baselines if you don't compare players on an even playing field, which is what WAR is. You pesos/dollars example is a good one.

    Murray sees that guy who hits 30 HR and 100 RBI and compares him to every other player with statistics in those categories. The problem is the players are all on different teams with different players around them. So not using a baseline tends to cause a more uneven playing field.

    I love intangibles in a player, but I won't vote for a player based on that. They are not measurable.

    ReplyDelete
  7. This comment has been removed by the author.

    ReplyDelete
  8. I think Murray is also taking a shot at the idea of E-mail. He mentions that Forman should "e-mail" the final WAR results to the BBWAA. But we already know that WAR is easily found on Forman's web site without any need for e-mail. I think Murray is just taking a potshot at e-mail for no particular reason other than it represents something else related to computers that Murray probably despises or feels threatened by.

    By the way, email communication is losing steam these days with the surge of social networking, so I think this is even funnier that Murray thinks E-mail is one of those dirty seedy things that should be hidden in a dark alley.

    ReplyDelete
  9. Intangibles Murray? Really? That makes more sense than WAR? What will you call it VAJ? (Value over Jeeter)

    ReplyDelete
  10. Anon, so I guess that means Murray is almost two steps behind the best communication of today? I didn't even catch that "email" part, but it is funny. It's like he thinks Forman keeps his results in a secret vault in his house so that no one can find out how they come about the statistics and how they are formed. I think he believes we all sit around our computers and wait for the results of the computers to be published.

    Of course intangibles! That's how you measure how good a player is...with something that can't be measured!

    ReplyDelete