PDA

View Full Version : Rosenthal On Stats in Baseball


Lip Man 1
12-19-2002, 08:36 PM
Gang:

In the print version of The Sporting News this week Ken Rosenthal has a long feature story on the very subject that Andrew and I always argue about, the place in baseball for (for want of a better word...) stat geeks.

Rosenthal talks to old timers like Pat Gillick and Roland Hemond as well as young guns like Theo Epstein to try to see what the trend and role in baseball is for all this.

My take on reading the story is that baseball does have a place and should value what is brought to the table by sabermetrics but that a team can not base it's player decisions solely because of them.

As the story points out, to do so would alienate your scouting department and numbers can't judge heart, guts, determination or the ability to execute fundamentals.

The best teams seem to have struck some sort of balance.

I recommend reading the story.

Lip

kermittheefrog
12-20-2002, 12:28 AM
There are some pretty funny anti-stathead quotes in there, IMO. From what I hear most teams get some imput from a sabermetrician (theres your better term than stat geek). However according to sabermetric consultant Eddie Epstein, who consults many teams, no more than a third of baseball takes sabermetric analysis seriously. This is despite adovating sabermetrics in the media and/or employing a consultant. Lip, from my arguments with you I get the feeling you're like the large persentage of baseball that advocates sabermetrics on some level but doesn't actually have any faith in it.

Also most statheads will tell you that sabermetrics isn't about running an organization solely based on the numbers. It's about gathering as much objective information as possible. And I doubt there is anyone out there saying there is no place in baseball for scouts despite what you may or may not believe. If there is someone saying get rid of scouts altogether thats someone who is a fanatical minority of statheads rather than someone representative of sabermetric belief.

Some people on this board look at me as someone who only looks at numbers but thats not true at all. I seek out and take into consideration a lot of information on players that is non-statistical. I read a ton about baseball everyday from sabermetric research to Baseball America's scouting reports. The only thing I don't really have is contacts within MLB organizations. I'm hoping that changes soon.

Lip Man 1
12-20-2002, 12:49 AM
Andrew:

If it means improving the Sox, hell, I'd even go to bat for you!

If you read the story I get the sense that Hemond doesn't care much for your kind.

Lip

kermittheefrog
12-20-2002, 12:56 AM
Originally posted by Lip Man 1
Andrew:

If it means improving the Sox, hell, I'd even go to bat for you!

If you read the story I get the sense that Hemond doesn't care much for your kind.

Lip

Yeah, doesn't surprise me, he's part of the old guard. I think the problem with the old guard like Hemond isn't that non-statheads can't be good GMs, Brian Sabean isn't a stathead, he's a good GM. However a lot of non-stathead GMs get their jobs out of the good ole boy system rather than merit. Thats why Dave Stewart threw a fit, he thought he was assured a job by the GOB system. Who would have thought the Jays would hire a GM based on who is best for the job? The problem is when guys like Stewart get the job and don't even have a plan for the organization. The Toronto ownership group isn't John Henry, it wasn't a case of a stathead looking for a stathead GM. It was a case of Ricciardi came in with a plan for the organization, clear goals and an organizational strategy. Stewart came in with the attitude that he'd already earned the job.

Dadawg_77
12-20-2002, 01:03 AM
I think Boston will be a great test case on how sabermetrics works in real life. With an owner with a baises in finance and has made numberous quotes supporting sabermetrics, Bill James on board, and new kid on the block

Also for those who don't have TSN here is a link to the story

http://story.news.yahoo.com/news?tmpl=story&u=/sn/20021218/sp_sn/analyze_this__are_instincts_or_stats_more_valuable _

voodoochile
12-20-2002, 08:16 AM
Originally posted by Dadawg_77
I think Boston will be a great test case on how sabermetrics works in real life. With an owner with a baises in finance and has made numberous quotes supporting sabermetrics, Bill James on board, and new kid on the block

Also for those who don't have TSN here is a link to the story

http://story.news.yahoo.com/news?tmpl=story&u=/sn/20021218/sp_sn/analyze_this__are_instincts_or_stats_more_valuable _

You're link doesn't work - the document expired.

Dadawg_77
12-20-2002, 09:36 AM
Originally posted by voodoochile
You're link doesn't work - the document expired.


Try This one (http://news.yahoo.com/news?tmpl=story2&cid=95&ncid=95&e=11&u=/sn/20021218/sp_sn/analyze_this__are_instincts_or_stats_more_valuable _)

voodoochile
12-20-2002, 09:52 AM
Originally posted by Dadawg_77
Try This one (http://news.yahoo.com/news?tmpl=story2&cid=95&ncid=95&e=11&u=/sn/20021218/sp_sn/analyze_this__are_instincts_or_stats_more_valuable _)

Nope, but I fixed the problem in the link (you had two "http:\\" at the beginning) still it leads me to an expired webpage...

idseer
12-20-2002, 10:49 AM
i would think there's a relatively simple way to prove how valuable sabermatrics actually is.

i've seen dozens of programs regarding facial restructure gleaned only from a bare skull. it's used quite extensively by police and other agencies. but what i've always wanted to see is a test of taking a known person's skull and giving it to one of these sculpture specialists who is unaware of the identity of the (now deceased) person. i would love to see if the result, in fact, looks like the actual person looked! (i have my doubts)

similarly, with all years and names of players being kept hidden, pick a previous year, or small series of years (whatever it takes to make predictions on future performance) and a large number of players as a starting point. have someone do the math and show what the future of this large group of individual players will be. compare this result with the actual future of these players. has anyone ever really done this? and i don't mean a couple players that prove someone's point. will the calculations be any different than estimating a players worth based on the basic stats ( avg. hr, rbi, era) as was always done in the past?
unless it can be shown in such a way there will always be a large segment that feels grinding all those numbers to the n'th degree is just a gigantic waste of time ... interesting perhaps, but a worthless tool.

kermittheefrog
12-20-2002, 12:34 PM
Originally posted by idseer
i would think there's a relatively simple way to prove how valuable sabermatrics actually is.

i've seen dozens of programs regarding facial restructure gleaned only from a bare skull. it's used quite extensively by police and other agencies. but what i've always wanted to see is a test of taking a known person's skull and giving it to one of these sculpture specialists who is unaware of the identity of the (now deceased) person. i would love to see if the result, in fact, looks like the actual person looked! (i have my doubts)

similarly, with all years and names of players being kept hidden, pick a previous year, or small series of years (whatever it takes to make predictions on future performance) and a large number of players as a starting point. have someone do the math and show what the future of this large group of individual players will be. compare this result with the actual future of these players. has anyone ever really done this? and i don't mean a couple players that prove someone's point. will the calculations be any different than estimating a players worth based on the basic stats ( avg. hr, rbi, era) as was always done in the past?
unless it can be shown in such a way there will always be a large segment that feels grinding all those numbers to the n'th degree is just a gigantic waste of time ... interesting perhaps, but a worthless tool.

Sabermetrics isn't all about predicting future performance of specific players. It comes down to learning information about the game in general to provide a guideline for acquiring players or making a system to predict future performance. And these methods are well tested. I believe people like Branch Rickey, Sandy Alderson and Billy Beane have already proven the usefulness of sabermetrics. I really feel like if you don't believe in sabermetrics by now you either haven't seen the information out there that proves things or you're just stubborn.

idseer
12-20-2002, 02:21 PM
Originally posted by kermittheefrog
. I really feel like if you don't believe in sabermetrics by now you either haven't seen the information out there that proves things or you're just stubborn.

kermit, i'm not looking for a fight, and i'm not being stubborn. i was asking this with an open mind. i've never seen any verifyable proof that this system does any better than the older methods used. maybe i'm asking for something that, like the theory of relativity, only a few can really understand. the best way to show someone like myself is a practical experiment that shows a player is better understood with this system than any other method. maybe that's why the majority of baseball people still have little faith in it?

does such an experiment exist?

kermittheefrog
12-21-2002, 02:48 AM
Originally posted by idseer
kermit, i'm not looking for a fight, and i'm not being stubborn. i was asking this with an open mind. i've never seen any verifyable proof that this system does any better than the older methods used. maybe i'm asking for something that, like the theory of relativity, only a few can really understand. the best way to show someone like myself is a practical experiment that shows a player is better understood with this system than any other method. maybe that's why the majority of baseball people still have little faith in it?

does such an experiment exist?

You're talking really general. There is no blanket test of every bit of sabermetric thought in part because there isn't consensus on every bit of sabermetric thought any more than there is consensus on everything in traditional baseball thought. If you want to get down to things like tradional numbers like BA, RBI and HR versus OBP and SLG. OBP and SLG correlate a lot better with teams runs scored than batting average. I think the weakness of looking at RBIs is fairly obvious. RBI has a lot to do with the guys around you in the lineup. If you hit third in a lineup with two guys who get on base in front of you, you'll get lots of RBIs. Just looking at HR is silly because you don't know how many at bats its done in. Theres just a whole lot of info missing if you just look at BA-HR-RBI. I don't know what else to say here.