One thing i don't get is why FIP and xFIP try to predict ERA. ERA is crap. Shouldn't it try to predict itself? If xFIP isn't stable year to year, how accurate is it really? I think that's the question.

its on the ERA scale. It tried to predict how well a pitcher pitches regardless of defense

Santiago has always had an era better than fip, and a fip always better than xfip. It's a large enough sample that its probably not a fluke anymore. Citing xfip now is more cherry picking to put him in the worst possible light than a fair analysis of what he brings to the table.

 

His era is high right now, but a big part of that was from his last handful of outings where he clearly was hurt. Give him some time to get healthy and he can be an effective back of the rotation starter down the stretch.

All attempts at uberstats have been failures. FIP variants are fine to get a complete picture of a pitcher, but they are not replacements for looking at everything else.

 

Unfortunately, FIP variants tend to get thrown around in a vacuum without any consideration for other stats, excepting of course a comparison to the pitcher's ERA.

 

Having said that, I would love to see an attempt at an uberstat based purely on pitchf/x data.

 

 

Edited July 7, 20179 yr by Doomtints

its on the ERA scale. It tried to predict how well a pitcher pitches regardless of defense

Ehhh, I've seen it described as a better predictor of future era than current era. By why predict era at all? I haven't seen FIP be all that consistent. If it's truly independent, then it should be consistent regardless of defense. Siera has some promise. In the end there is more variability than we can even hope to calculate. All stats show an incomplete picture. ERA is one stat that actually shows a complete picture. It is the end result of all variables compiled. Outliers or shifts on the normalcy curve are the most interesting parts of statistical analysis in my opinion. Hector Santiago represents that at this point.

Ehhh, I've seen it described as a better predictor of future era than current era. By why predict era at all? I haven't seen FIP be all that consistent. If it's truly independent, then it should be consistent regardless of defense. Siera has some promise. In the end there is more variability than we can even hope to calculate. All stats show an incomplete picture. ERA is one stat that actually shows a complete picture. It is the end result of all variables compiled. Outliers or shifts on the normalcy curve are the most interesting parts of statistical analysis in my opinion. Hector Santiago represents that at this point.

the metric being discussed in the thread is XFIP, not FIP. There is a subtle difference discussed below.

 

Trying to predict ERA makes no sense. Why would fielding INDEPENDENT pitching predict ERA that is a fielding DEPENDENT stat?

 

The purpose of the metric is to predict pitchers performance independent of the defense behind them on an ERA scale for easy comparison to a widely available metric.

 

FIP uses actual home runs given up. Since home runs per fly ball vary a lot in small sample, XFIP uses league average home runs per fly ball rate against the pitchers fly ball rate. On a league wide scale, home runs per fly ball increases the sample size large enough to stabilize.

 

 

I do agree that statistical changes are the interesting part of the game. However isolating the variables is the only way to predict outcomes. Isolating defense from run prevention gives us a better picture of the pitchers effect on the game.

 

In this case Santiago is really interesting. The ongoing correlation of lower than expected homeruns to fly ball rate might lead one to believe he has a skill. It also might lead one to believe he's been lucky.

 

SIERRA has some promise, I plan on doing some research.

 

 

http://www.fangraphs.com/library/pitching/xfip/

Edited July 8, 20179 yr by Sconnie

Please stay on topic, xfip? Fip? dip? Vorp?

 

This topic is about forks, and maybe spoons.

Please stay on topic, xfip? Fip? dip? Vorp?

 

This topic is about forks, and maybe spoons.

what? We can't statistically say he's done?

All attempts at uberstats have been failures. FIP variants are fine to get a complete picture of a pitcher, but they are not replacements for looking at everything else.

 

Unfortunately, FIP variants tend to get thrown around in a vacuum without any consideration for other stats, excepting of course a comparison to the pitcher's ERA.

 

Having said that, I would love to see an attempt at an uberstat based purely on pitchf/x data.

I second this notion. More perfect words have not been spoken.

 

what? We can't statistically say he's done?

 

A spoonful of statistics is acceptable.

A spoonful of statistics is acceptable.

Snepp! It feels like a very long time since I've seen you in the forums. Good to hear from you

It's been at least a week and a half. A glorious holiday weekend at the lake followed by an equally "un"glorious week of catching up at work.

https://media.giphy.com/media/VUOnqJTJZggbC/giphy.gif

 

anyone surprised by how bad Santiago is?

I am. I thought he'd be middling, maybe kinda bad, but not terrible.

 

But if a middling pitcher suffers velocity issues, I guess it's to be expected.

 

I am. I thought he'd be middling, maybe kinda bad, but not terrible.

But if a middling pitcher suffers velocity issues, I guess it's to be expected.

He should be a middling pitcher but I still think he hurt something in that extra inning game. His velocity dropped after that and it cratered in his last start.