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

[Frobby]

My comment did not dismiss FIP. Nor was I comparing it to ERA or any other metric.

But since you brought it up, though ERA is flawed, I think FIP is as flawed or more flawed than ERA. Just because FIP is newer doesn't make it better. To make an analogy, FIP is putting blinders on a race horse: maybe helpful to the horse while on the course, but would mean death for the horse in the wild.

For example, I would choose to have 1975 Jim Palmer, with his 2.96 FIP, 5.24 K/9, 25 complete games and 10 shutouts with a 2.09 ERA, on my squad over 2013 Max Scherzer, with his 2.74 FIP, 10.9 K/9 and zero complete games with a 2.90 ERA. With Scherzer, he and the team are much more reliant on the bullpen to secure a win. I'd take Palmer; FIP declares Scherzer more valuable.

Not sure I like that analogy. Complete games and shutouts are a function of how the game is played now. Even Palmer would tell you that -- he has made the point during broadcasts many times. So, while I might prefer 1975 Jim Palmer to 2013 Max Scherzer, that wouldn't be my reason.

[weams]

Not sure I like that analogy. Complete games and shutouts are a function of how the game is played now. Even Palmer would tell you that -- he has made the point during broadcasts many times. So, while I might prefer 1975 Jim Palmer to 2013 Max Scherzer, that wouldn't be my reason.

I prefered that Jim Palmer to this Max Scherzer.

[waroriole]

I believe that Buck likes strikeouts. I believe he knows we will never pay for them. We would need to grow them.

Nah, Buck's not a fascist. Groundballs are democratic.

[Stotle]

But Kershaw and Felix wouldn't see as MUCH improvement with a good defense behind them because not as many balls would be in play.

So improving the defense can in fact narrow the difference between a staff of Verlanders and a staff of Chen's. And can probably be done alot more cheaply than going out and paying for high K pitchers.

Right. This is the part of team building that gets tricky (and is supremely interesting). It all comes down to resources and in-house options. Your in-house options give you your starting point, and your resources dictate the potential paths you can take in constructing a winner. Obviously the more resources you have the greater number of paths to constructing a winner will be available to you.

To beat a dead horse, this is why I thought the Jimenez signing was particularly unfortunate. It simultaneously reduced the resources available to club while plugging in a volatile profile type that didn't (to me) seem to mesh with the type of team Baltimore has been putting together.

[oceanicvoyage]

I don't think this is revolutionary, but the O's are making it more prominent. ERA, FIP, and xFIP are flawed and I think good teams realize that and use that to their advantage.

Over the past 5 years, the A's, Giants, Braves have the highest positive difference between xFIP and ERA (ERA lower than xFIP). These are teams who have historically had quality pitching and continued success with young pitchers, but not necessarily cy young caliber pitchers every year.

I remember there was a lot of buzz over the Gio Gonzalez trade from the A's to the Nats where people were claiming he wouldn't be as good because the A's had better defense and park factor. I think the A's were willing to flip him because they knew they could manipulate ERA with their park and defense and continue to develop minor leaguers and young players around defense. Plus, his performance didn't really change perhaps because he learned to pitch to the good defenses.

[Pheasants]

I don't think I've ever seen pitchers equal to the early 1990s Atlanta Braves at getting a stream of weak ground balls. I haven't checked the stats at all, maybe my memory is gone, but between them and the current O's pitchers, I do believe pitchers (combined with fielders) can significantly effect results that FIP doesn't count.

After a few years of FIP not being predictive for specific reasons, shouldn't there be an admission that it isn't a great predictive tool in cases where those reasons exist? Shouldn't the designers be busy correcting their formula, rather than insisting that once again the Orioles will have horrible pitching? After all that's what being a stathead is all about, finding a stat that truly shows what the results are going to be.

[Stotle]

I don't think I've ever seen pitchers equal to the early 1990s Atlanta Braves at getting a stream of weak ground balls. I haven't checked the stats at all, maybe my memory is gone, but between them and the current O's pitchers, I do believe pitchers (combined with fielders) can significantly effect results that FIP doesn't count.

After a few years of FIP not being predictive for specific reasons, shouldn't there be an admission that it isn't a great predictive tool in cases where those reasons exist? Shouldn't the designers be busy correcting their formula, rather than insisting that once again the Orioles will have horrible pitching? After all that's what being a stathead is all about, finding a stat that truly shows what the results are going to be.

There's no formula to correct. It is exactly what it claims to be -- a measure of a pitcher's performance based on the variables the pitcher can control.

[Pheasants]

There's no formula to correct. It is exactly what it claims to be -- a measure of a pitcher's performance based on the variables the pitcher can control.

Then it should not be stated as having predictive power. So the formula is good, but is misued.

[weams]

There's no formula to correct. It is exactly what it claims to be -- a measure of a pitcher's performance based on the variables the pitcher can control.

Aside from nuanced issues. Like the Maddux strike or the ability to shove a baseball deep into a hand between two fingers and make it dive. Those things are not really included in the data, yet do have an effect on the result. Or even fielding their position themselves. Like Jake Arrieta did not do as an Oriole.

[Stotle]

Aside from nuanced issues. Like the Maddux strike or the ability to shove a baseball deep into a hand between two fingers and make it dive. Those things are not really included in the data, yet do have an effect on the result. Or even fielding their position themselves. Like Jake Arrieta did not do as an Oriole.

I don't understand the distinction. Maddux strike or ability to throw a pitch are included in the data. They manifest via strikeout/walk rate. They are characteristics unique to the pitcher.

Pitcher fielding has a de minimis result on the data. Case in point, Arrieta's least valuable defensive season was last year -- his most valuable season overall by a wide margin. In fact, his -3 DRS was worse than any season in Baltimore.

[weams]

I don't think this is revolutionary, but the O's are making it more prominent. ERA, FIP, and xFIP are flawed and I think good teams realize that and use that to their advantage.

Over the past 5 years, the A's, Giants, Braves have the highest positive difference between xFIP and ERA (ERA lower than xFIP). These are teams who have historically had quality pitching and continued success with young pitchers, but not necessarily cy young caliber pitchers every year.

I remember there was a lot of buzz over the Gio Gonzalez trade from the A's to the Nats where people were claiming he wouldn't be as good because the A's had better defense and park factor. I think the A's were willing to flip him because they knew they could manipulate ERA with their park and defense and continue to develop minor leaguers and young players around defense. Plus, his performance didn't really change perhaps because he learned to pitch to the good defenses.

All good points. Thanks for the input.

[weams]

I don't understand the distinction. Maddux strike or ability to throw a pitch are included in the data. They manifest via strikeout/walk rate. They are characteristics unique to the pitcher.

Pitcher fielding has a de minimis result on the data. Case in point, Arrieta's least valuable defensive season was last year -- his most valuable season overall by a wide margin. In fact, his -3 DRS was worse than any season in Baltimore.

The fact that Maddux can throw a ball and make it be a strike is not included in the data. The fact that the ball that drops down dramatically is likely to be beat into the dirt is not either.

[weams]

I don't understand the distinction. Maddux strike or ability to throw a pitch are included in the data. They manifest via strikeout/walk rate. They are characteristics unique to the pitcher.

Pitcher fielding has a de minimis result on the data. Case in point, Arrieta's least valuable defensive season was last year -- his most valuable season overall by a wide margin. In fact, his -3 DRS was worse than any season in Baltimore.

I did not bother looking at Jake's data. (Who knew!) I just always felt that he made his fielders uncomfortable with his stiffness and the quality of his timing and throws to the bases.

[weams]

Then it should not be stated as having predictive power. So the formula is good, but is misued.

It does appear to have good predictive qualities from season to season for pitchers.

[Stotle]

The fact that Maddux can throw a ball and make it be a strike is not included in the data. The fact that the ball that drops down dramatically is likely to be beat into the dirt is not either.

You're going to need to explain that. That's like claiming the data doesn't account for Aroldis Chapman throwing the ball so hard. Maddux's "ability to make a ball a strike" is absolutely reflected in his strikeout/walk rates, and is ignored inasmuch as it produces balls in play. Ditto the ability for Pedro to throw crazy pitches because of long fingers. All of that is taken into account because it manifests in the production.

All FIP does is isolate the portion of the production that the pitcher most controls, which is what allows it to more accurately predict future performance (because this aspect of production is less likely to be subject to external variables that can vary widely year-to-year).