Buck: "part of developing pitching is having guys who can defend"

[malkusm]

Ignore all the unnecessary quotes at the beginning of my post. Gonna go get a Plus membership now :laughlol:

[OFFNY]

Buck: "Part of developing pitching is having guys who can defend."

I suppose that the antithesis of that would be having Bucky LaGrange behind you.

[video=youtube;yxx8SY72E_U]https://www.youtube.com/watch?v=yxx8SY72E_U

[Beef Supreme]

Because the types of things measured are not really dependant on location?

Exactly, weams! Drungo brought up the most appealing aspect of FIP -- predicting pitcher value for those who change teams. Seems like K/9 and BB/9 should hold rather steady for the majority of pitchers from year to year. HR/9 can fluctuate, of course. So ballpark should not affect FIP much. But in practice, does FIP show a tendency to reflect ballpark effects? Has this analysis been pursued? Could ballparks actually have an effect on Ks and BBs?

[weams]

Exactly, weams! Drungo brought up the most appealing aspect of FIP -- predicting pitcher value for those who change teams. Seems like K/9 and BB/9 should hold rather steady for the majority of pitchers from year to year. HR/9 can fluctuate, of course. So ballpark should not affect FIP much. But in practice, does FIP show a tendency to reflect ballpark effects? Has this analysis been pursued? Could ballparks actually have an effect on Ks and BBs?

Maybe what he is stating is that it is team effects. Not ballpark effects.

[weams]

But in practice, does FIP show a tendency to reflect ballpark effects? Has this analysis been pursued? Could ballparks actually have an effect on Ks and BBs?

I'd say no, no and nono.

[weams]

I could be wrong, but logic would suggest that walk rates would be lower in a park like San Diego than they would be in a band box like Cincinatti. A pitcher would be more likely to challenger a hitter when behind in the count if he's got a big ball park to work worth. Seems logical to me but the stats might say otherwise.

You may have a point. Or home runs up in places with good hitters background. I suspect stuff like that is mostly data noise though.

[Beef Supreme]

I'd say no, no and nono.

Typo, weams, or are you enjoying Sunday morning watching the hair removal infomercial? :-)

[weams]

Typo, weams, or are you enjoying Sunday morning watching the hair removal infomercial? :-)

But in practice, does FIP show a tendency to reflect ballpark effects?

No.

Has this analysis been pursued?

No

Could ballparks actually have an effect on Ks and BBs?

NoNo

[Can_of_corn]

I could be wrong, but logic would suggest that walk rates would be lower in a park like San Diego than they would be in a band box like Cincinatti. A pitcher would be more likely to challenger a hitter when behind in the count if he's got a big ball park to work worth. Seems logical to me but the stats might say otherwise.

Maybe, a bit. And a stadium like Oakland could effect both since more foul pop outs will be converted into outs so there will be less three ball and two strike counts.

I am guessing that the ballpark effects on K and BB rate would be so small as not worth noting.

[weams]

Maybe, a bit. And a stadium like Oakland could effect both since more foul pop outs will be converted into outs so there will be less three ball and two strike counts.

I am guessing that the ballpark effects on K and BB rate would be so small as not worth noting.

As I said, I think it would be statistical noise. Only.

[CA-ORIOLE]

Origanally posted by Malksum: My point is that the burden of proof is on people to tell me why I should rely on FIP as a projection tool, since they are the ones using it to project. The scientific method allows any and all criticism of a theory, as far as I know. Critics of FIP don't need to prove that soft contact is a repeatable skill, with batted ball data that doesn't yet exist. No more than critics of flat-earth theory needed to prove that the earth was round, back in the day.

Sure, I could say quantum physics is crap by your logic. It's only theory after all. The statistical analysis has been done, re-validated, tweaked and scrutinized by numerous professional statisticians and studies over the years. Many, perhaps set out, or at least skeptical to prove McCracken's original theory as wrong. The r2 For FIP is better that ERA, better than SIERA, xFIP or any other metric that has been out out there to date. I have seen ZERO studies to the contrary. You want to point me to one? Who says batted ball data doesn't exist? What do you think the statisticians used to come up with DIPS/FIP and continually re-validate it? Batted ball data is plentiful. As far as I know their have been only minor changes and tweaks to FIP/DIPS theory over the years: Among them a special category for Knuckle ballers, a hit profile for popups, and incorporation of HBP's.

I couldn't care less if you want to use FIP is a projection tool or not, but the fact is FIP's r2 remains the best there is of what we have. FIP is a valuation method. It is another form of ERA. In fact, if it were referred to as "FIP based ERA" I think there would be less confusion about it.

FIP is a useful tool. It's also probably flawed in some ways. To argue that critics of FIP need to prove that FIP is wrong, is taking the statistic farther than the statistic was ever intended to go.

Flawed? No it's just perfect in every way. Please stop with ridiculous and condescending comments. FIP is pitcher valuation method like ERA. You wanna say FIP is a useful tool, then say ERA is just a useful tool. FIP may be flawed in many ways. I'm not sure if it's more flawed than ERA, but since it has a better r2 my guess is that almost any statistician associated with the game would point out ERA as more flawed, certainly on aggregate. Does ERA, xFIP and SIERA do a better job of capturing some of the outliars? Almost certainly yes and all that said, I generally prefer ERA+ and ERA based WAR for starting pitchers over FIP based WAR myself.

Also, exactly where do you think the McCracken's "Defense Independent Pitching" was "intended to go"? An arbitrary peripheral stat I suppose? The FIP formula was developed by Tom Tango (a fairly bright guy) with close consultation and advice from McCracken. It was meant to be a pitcher valuation method. It was converted to ERA format to make the valuation more intuitive. That's why it's incorporated into Fangraphs WAR. Like it or hate it, there is a lot of data to support it and continuing to support and re-validate it. The FIP critics here don't need to prove anything, but they could at least provide some critical analysis, studies, or other data to dispute the data that is out there and that has been presented here. I've seen almost none of that. At the very least point me to a semi-respected sabermetric critic.

If your bottom line to this discussion is that any theory can be wrong, I'm not really sure why you want to discuss sabermetrics of any kind on a message board.

From Wikipedia:

Bill James also expressed some skepticism but recognized the potential value of McCracken's findings if further research bore them out. He argued that "the research really should be done, for several reasons. First, if McCracken turns out to be correct, this has important consequences, even allowing us, to a certain extent, to predict movements in pitcher's records. . . ." In his New Historical Baseball Abstract in 2001, James acknowledged that McCracken was correct, that the results were significant, and that James himself felt "stupid for not having realized it 30 years ago."[9] Rob Neyer also noted the impact of McCracken's discovery on James' subsequent work.[10] McCracken's discovery and its influence on baseball analysis is outlined in Moneyball: The Art of Winning an Unfair Game by Michael Lewis.

Writer Jeff Passan wrote:

A decade after Baseball Prospectus let McCracken spread the gospel in a story that popularized DIPS across the sport, it remains among the most seminal theories developed by sabermetrics, the nickname given to quantitative baseball study. It's almost certainly the most revolutionary. Nothing before or since has so upended an entire line of thought and forced teams to assess a wide breadth of players in a different fashion.[21][22]

[CA-ORIOLE]

No.

Is ERA adjusted for Park Factors?

[malkusm]

Not necessary to quote all of the above, but I work as a professional statistician/data analyst so it's not necessary to explain R-squared. I apologize if anything I said came off as "ridiculous and condescending" as that wasn't my intent.

Every theory is flawed, and always will be, since luck is a variable that cannot be controlled for. Nowhere did I state that FIP wasn't an improvement over ERA; for what it is trying to accomplish, it certainly is. That doesn't preclude us from saying that we can do better. But, much like the problems with evaluating defense, there are a lot of variables and a lot of data is needed. That's why even saber guys suggest taking several seasons of UZR into consideration before making a judgment on a player's defense. Similarly, it might take several seasons for a pitcher to show that they have some skill that isn't adequately captured by FIP or other metrics available; that skill may only be meaningful when combined with other factors, such as good infield defense, leading analysts who are looking at league-wide data to form the conclusion that the effect of such a skill has a negligible impact.

When you look at things like this:

Batted Ball Type: xBABIP, wOBAcon, % of batted balls

Groundball – Weak: .151, .112, 31.4%

Groundball – Medium: .461, .416, 9.5%

Groundball – Well-Hit: .647, .610, 3.8%

Line Drive – Weak: .622, .579, 2.3%

Line Drive – Medium: .650, .638, 7.3%

Line Drive – Well-Hit: .719, .815, 11.1%

Flyball – Weak: .078, .074, 18.5%

Flyball – Medium: .069, .081, 8.2%

Flyball – Well-Hit: .641, 1.168, 7.8%

It's hard for me to stand behind the theoretical underpinnings of FIP, that a pitcher controls 100% of HR, BB, and K rate and 0% of everything else. Maybe those numbers should be 90% and 25% or something like that. I don't personally have the time to run the analysis, but I believe that it's coming, from the people who do, and that it might already be developed by a few teams who are using it to exploit small, yet meaningful, market inefficiencies in the pitching market. It's hard for me to stand behind the results of FIP as a pure projection tool when you have cases like Jim Palmer whose actual results outperformed FIP's expectation for 17 consecutive seasons.

The point is not to bash FIP because, as I said, it IS better than ERA. It's a model, and models necessarily use simplifying assumptions to approximate reality. I'm just saying that a better model is probably possible, and I'm not going to be surprised when it becomes established.

[Frobby]

Very, very desirable!

There are many exceptionally talented pitchers topping the FIP leaderboard. A little surprised to see Phil Hughes so high (#6).

Check out Hughes' BB/9 last year and you'll know why he's at #6.

[CA-ORIOLE]

Not necessary to quote all of the above, but I work as a professional statistician/data analyst so it's not necessary to explain R-squared. I apologize if anything I said came off as "ridiculous and condescending" as that wasn't my intent.

Every theory is flawed, and always will be, since luck is a variable that cannot be controlled for. Nowhere did I state that FIP wasn't an improvement over ERA; for what it is trying to accomplish, it certainly is. That doesn't preclude us from saying that we can do better. But, much like the problems with evaluating defense, there are a lot of variables and a lot of data is needed. That's why even saber guys suggest taking several seasons of UZR into consideration before making a judgment on a player's defense. Similarly, it might take several seasons for a pitcher to show that they have some skill that isn't adequately captured by FIP or other metrics available; that skill may only be meaningful when combined with other factors, such as good infield defense, leading analysts who are looking at league-wide data to form the conclusion that the effect of such a skill has a negligible impact.

When you look at things like this:

Batted Ball Type: xBABIP, wOBAcon, % of batted balls

Groundball – Weak: .151, .112, 31.4%

Groundball – Medium: .461, .416, 9.5%

Groundball – Well-Hit: .647, .610, 3.8%

Line Drive – Weak: .622, .579, 2.3%

Line Drive – Medium: .650, .638, 7.3%

Line Drive – Well-Hit: .719, .815, 11.1%

Flyball – Weak: .078, .074, 18.5%

Flyball – Medium: .069, .081, 8.2%

Flyball – Well-Hit: .641, 1.168, 7.8%

It's hard for me to stand behind the theoretical underpinnings of FIP, that a pitcher controls 100% of HR, BB, and K rate and 0% of everything else. Maybe those numbers should be 90% and 25% or something like that. I don't personally have the time to run the analysis, but I believe that it's coming, from the people who do, and that it might already be developed by a few teams who are using it to exploit small, yet meaningful, market inefficiencies in the pitching market. It's hard for me to stand behind the results of FIP as a pure projection tool when you have cases like Jim Palmer whose actual results outperformed FIP's expectation for 17 consecutive seasons.

The point is not to bash FIP because, as I said, it IS better than ERA. It's a model, and models necessarily use simplifying assumptions to approximate reality. I'm just saying that a better model probably exists, and I'm not going to be surprised when it becomes established.

Thanks for the intelligent reply.

1. I'm not sure why you keep referring to FIP in terms such as a "pure projection tool".

2. No one is stating or implying pitchers babip data is 100% random.

3. Luck and/or randomness should be controlled time and sample size. FIP and ERA can vary widely from season to season but generally regress towards each other.

4. With respect to the data you submitted, are saying that pitchers can control the quality and distribution of line drives, groundballs and flyballs to some statistical significance? What percentage and do you think this is and by how much? What is the difference with respect to the published reports that Drungo provided with respect to variation of babip data.

4. Since you're a professional statistician, then I'm sure you understand why a significant amount of time and energy might not be devoted to capturing outliars from aggregate data. I'm open to the idea that FIP does not capture "everything" and have readily admitted in this thread that other metrics like ERA or SIERA or xFIP may capture the outliars better. Since you understand r2, then you know the dilemma with that.

5. With respect to Palmer, I think even the most diehard fans would readily admit that Palmer would not have been as good with out (arguably) one of the best defensive teams ever assembled over a very long time. If you look at Palmer's RA9 defense numbers on BBREF I doubt you'd find anywhere near that number for any pitcher over a career. Those types of situations don't often happen. Is that the only aspect affecting Palmers FIP/ERA divergence? Probably not since it's (still very large delta between rWAR and FWAR). but it's a pretty significant factor. Palmer's road ERA was quite a bit higher than his home ERA (more than should be expected imo) so it's possible that he was adept at gaining a marginal advantage from his park. Certainly his home park suppressed home runs. He also never gave up a grand slam (I'm sure you've heard) and was able to mantain a pretty high LOB% among other factors. I didn't go through each year, but in 1971 Palmers FIP based WAR (fWAR) outperformed his rWAR. So I would say that he really did not "outpich his FIP" in 1971 at least.

The point is not to bash FIP because, as I said, it IS better than ERA. It's a model, and models necessarily use simplifying assumptions to approximate reality. I'm just saying that a better model probably exists, and I'm not going to be surprised when it becomes established.

Probably so. SIERA and xFIP were attempts to do so. I think you can utilize the data we have now and get a pretty good idea of what is going on.

Again, definitely appreciate the reply.