Sabermetrics

[clement]

Just a quick reply and then I go to bed...

Kovash didn't get the same results because he didn't do that study. From the way I read it, he was just comparing his results to Bradbury's and arguing that his approach is more sound.

I certainly appreciate your intellectual curiosity on all of this - the part I sort of take exception to in the sabre approach is the apparent reliance solely on quantitative methods. I'm a staunch believer in anthropological or qualitative research as well as survey data and other methodologies that don't just rely on numbers. For example, if you survey 1000 former major leaguers and they tell you by an overwhelming margin that team chemistry or hustle or protection or leadership or clutch hitting have an impact on the game, I consider that a type of evidence. Sabremetricians not only do not engage in this kind of research, but they seem to be dismissive this type of evidence as unscientific. It's not.

On the Albert example, you don't need research to deal with it. You just look at how he performs. If he expands the zone and is still productive, you say, "Atta boy, go after those pitches." If he expands the zone and he starts striking out or pulling ground balls to the shortstop, you tell him to be more selective. And your advice to him might depend on whether you have Manny Ramirez, Troy Glaus or Cesar Izturis in the on-deck circle. (i.e., protection)

[skmsw]

the part I sort of take exception to in the sabre approach is the apparent reliance solely on quantitative methods. I'm a staunch believer in anthropological or qualitative research as well as survey data and other methodologies that don't just rely on numbers.

Oh, so am I. Some of the stuff I'm not able to share is very much in keeping with that.

For example, if you survey 1000 former major leaguers and they tell you by an overwhelming margin that team chemistry or hustle or protection or leadership or clutch hitting have an impact on the game, I consider that a type of evidence.

Murky waters. I hear you, and it'd be nice to be able to take those kind of results at face validity. But there is a fair amount of evidence that people are not always the best reporters of why things happen to them. The risk is pretty high of returning a result indicating what former major leaguers believe, or what former major leaguers have heard from coaches since second grade, as opposed to indicating what actually affected former major leaguers.

Sabremetricians not only do not engage in this kind of research, but they seem to be dismissive this type of evidence as unscientific. It's not.

You're making a bit of an overgeneralization here, but I understand why and it is probably a valid criticism.

[clement]

For example, if you survey 1000 former major leaguers and they tell you by an overwhelming margin that team chemistry or hustle or protection or leadership or clutch hitting have an impact on the game, I consider that a type of evidence.

Murky waters. I hear you, and it'd be nice to be able to take those kind of results at face validity. But there is a fair amount of evidence that people are not always the best reporters of why things happen to them. The risk is pretty high of returning a result indicating what former major leaguers believe, or what former major leaguers have heard from coaches since second grade, as opposed to indicating what actually affected former major leaguers.

Of course what you say is true. I'm just saying it's one piece of the puzzle - and an important one not to neglect. There are measurement accuracy problems and limitations in a statistical approach as well; those waters can be just as murky, but I don't see enough people raising those kinds of questions about the methodology (at least not on these sites, that may not necessarily be the case in the community of people actually conducting the research if they are understanding and acknowledging the limitations of their approaches).

The thing that frustrates me among many people who do the research, and even more the people who read it, is that there is a prevalent belief that qualitative/subjective = inaccurate/false and quantitative/objective = accurate/truth. In my view, the best understanding of phenomena happens when you tackle them using multiple methodologies, and where possible, with an unbiased lens.

[Popeye_Card]

The way I've always looked at it is like degreed engineers vs. bootstrappers. One goes by the numbers and what they've learned in school and what time they've been around, and the other draws from years of experience.

When it goes as planned, one can thank the numbers while the other thanks the experience. When it doesn't, one can say, "well that's not what should have happened", and the other says "that's not what happened last time" (whether it did or not).

Not surprisingly, white collar and blue collar engineers butt heads every day too.

[AWvsCBsteeeerike3]

The way I've always looked at it is like degreed engineers vs. bootstrappers. One goes by the numbers and what they've learned in school and what time they've been around, and the other draws from years of experience.

When it goes as planned, one can thank the numbers while the other thanks the experience. When it doesn't, one can say, "well that's not what should have happened", and the other says "that's not what happened last time" (whether it did or not).

Not surprisingly, white collar and blue collar engineers butt heads every day too.

I'm a civil engineer and, as you can probably tell, don't go by the numbers.

[Popeye_Card]

I'm a civil engineer and, as you can probably tell, don't go by the numbers.

I'm sure you can probably attest to what I am talking about then. Booksmarts vs. experience.

I'm a mechanical, and most times I will trust the guy with the experience. But I'm going to check the numbers to be sure.

[AWvsCBsteeeerike3]

I'm a civil engineer and, as you can probably tell, don't go by the numbers.

I'm sure you can probably attest to what I am talking about then. Booksmarts vs. experience.

I'm a mechanical, and most times I will trust the guy with the experience. But I'm going to check the numbers to be sure.

Unfortunately, I can verify what you're saying. My entire job is basically designing roads and storm sewers - to carry storm runoff - for cities. Obviously, the cheaper the better but you still want high quality on the streets and for the storm sewer to be able to carry the water that a storm is going to produce. There are guys working for the cities that swear I'm an idiot. And, at 26 I couldn't possibly have learned what took them their whole lives to know. Well, they're just flat out wrong.

Long story short, I was in charge of one road and these couple guys swore I was an idiot. Every meeting we had they swore up and down the pipes I had weren't big enough to keep the road from flooding. After the first meeting, I was like okay I'll look into it to make sure what i did was correct. And, of course, I was So, I told them that and they were like "no, no, no, no, no, no,no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no. There is a 8' wide ditch there now that's 3' tall. That's what we need."

After explaining that concrete is a [expletive] smoother than tubmbleweeds with dollar general sacks stuck in them, they approved my design but weren't about to stop saying that there is no way it would work. And it wasn't just them....the Contractor told me yesterday he was surprised as hell that it was working and even the City Engineer said he had his doubts. But, and it couldn't have worked out any better, we had two HUGE rains so I went out and took pictures of my design working in the middle of a downpour and after just so everyone knew. There were floods all across town, everywhere, in total I think 7 roads were under water and had to be closed. Not mine though. The system worked BEAUTIFULLY. It was about 80-90% full for the 25 year storm. Right where the City wants it. Oh yeah, and 4 million dollars cheaper than what the City guys proposed. Boo-Ya to me!!!!! (No, I'm not humbe)

Edit: Don't quote this and say.....see, numbers don't lie.

This is a scientifically proven method.....correct 100% of the time.

[ghostrunner]

Of course what you say is true. I'm just saying it's one piece of the puzzle - and an important one not to neglect. There are measurement accuracy problems and limitations in a statistical approach as well; those waters can be just as murky, but I don't see enough people raising those kinds of questions about the methodology

Regarding this, I still have concerns with some of the defensive stats. I recognize what's being attempted in most of them - rather than go on errors or assists, the idea is to go by where the hits go. And in the aggregate, I imagine the numbers are pretty solid. But it seems like errors in the evaluator's judgement on zones (location) could easily have an adverse effect on a players numbers. How are sharp liners that reach the outfield accounted for? How are they judging the zones? What about bad hops? I know places like BIS supply data, but should we assume that's reliable enough on something like the location of a grounder? Are computers and cameras involved in any way or is it all an eyeball judgment of one person? Probably across a wide swath it's accurate, but a small collection of numbers for one player seems more prone to error. I imagine years of data would correct some of this, but I'd think it would take longer to get an accurate read than for hitting stats which are easier to get a handle on, and by that point injuries and age would have to factor in.

[cards2468]

I'm a civil engineer and, as you can probably tell, don't go by the numbers.

I'm sure you can probably attest to what I am talking about then. Booksmarts vs. experience.

I'm a mechanical, and most times I will trust the guy with the experience. But I'm going to check the numbers to be sure.

Unfortunately, I can verify what you're saying. My entire job is basically designing roads and storm sewers - to carry storm runoff - for cities. Obviously, the cheaper the better but you still want high quality on the streets and for the storm sewer to be able to carry the water that a storm is going to produce. There are guys working for the cities that swear I'm an idiot. And, at 26 I couldn't possibly have learned what took them their whole lives to know. Well, they're just flat out wrong.

Long story short, I was in charge of one road and these couple guys swore I was an idiot. Every meeting we had they swore up and down the pipes I had weren't big enough to keep the road from flooding. After the first meeting, I was like okay I'll look into it to make sure what i did was correct. And, of course, I was So, I told them that and they were like "no, no, no, no, no, no,no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no no, no, no, no, no, no. There is a 8' wide ditch there now that's 3' tall. That's what we need."

After explaining that concrete is a [expletive] smoother than tubmbleweeds with dollar general sacks stuck in them, they approved my design but weren't about to stop saying that there is no way it would work. And it wasn't just them....the Contractor told me yesterday he was surprised as hell that it was working and even the City Engineer said he had his doubts. But, and it couldn't have worked out any better, we had two HUGE rains so I went out and took pictures of my design working in the middle of a downpour and after just so everyone knew. There were floods all across town, everywhere, in total I think 7 roads were under water and had to be closed. Not mine though. The system worked BEAUTIFULLY. It was about 80-90% full for the 25 year storm. Right where the City wants it. Oh yeah, and 4 million dollars cheaper than what the City guys proposed. Boo-Ya to me!!!!! (No, I'm not humbe)

Tell them to give you the extra 4 million and share the wealth with your fellow GRBers

[AWvsCBsteeeerike3]

Tell them to give you the extra 4 million and share the wealth with your fellow GRBers

actually, now that I think about it, the 4 mil was how much their way would have cost, mine was 2 mil so the savings was ~ 2 mil. But, yeah, that's what I said.

To be honest, my neck wasn't on the line, my bosses was. But, he trusts me