I Used to be FIP, but now I'm xFIP
Apr 23, 2018 18:17:02 GMT -5
Texas Rangers, Tim_GiantsGM, and 5 more like this
Post by sansterre - Milwaukee Brewers on Apr 23, 2018 18:17:02 GMT -5
So I don't think it's a secret that I use FIP as my pitching metric of choice. And, going through StatsPlus, my team is doing excellently. The virtuous, respectful, apple pie-eating boys of Oakland are #2 in FIP right now. Confirming the stat's validity, the degenerate, foul-mouthed, don't-call-their-mothers-on-mothers'-day wastrels in Colorado have the #31 FIP in the league.
Seems like a pretty fair stat to me.
But another stat caught my eye. xFIP. I don't like this stat as much. Oakland is only #5 in it; Colorado is #17. What gives?
Some quick definitions:
FIP is basically a pitcher metric that only cares about four things: strikeouts, walks allowed, home runs allowed and innings pitched. It uses linear weights, the formula is basically (HRA * 13 + BB * 3 - K * 2) / IP + 3.2. It completely ignores defense and clutch pitching (coincidental or not).
xFIP is the same as FIP, but it basically assumes that every flyball is equally likely to be a home run, so it charges pitchers for every fly ball they allow, not for the home runs they allow.
So. League average in the PBL right now is that 12.93% of all flyballs become home runs.
Here are the teams that have the lowest HR / FB rates:
1) Washington (6.75%)
2) Chicago Cubs (7.94%)
3) Cleveland (8.21%)
4) Oakland (8.65%)
5) San Diego (9.13%)
Here are the teams that have the highest HR / FB rates:
1) Colorado (19.89%)
2) Montreal (18.82%)
3) White Sox (17.11%)
4) Pittsburgh (16.95%)
5) Seattle (16.75%)
That's kind of a weird pair of lists isn't it? I mean, I don't pretend to know every team's pitching quality, but I know that Pittsburgh has one of the top 2 rotations in the league.
So, as you may not be surprised to learn, I set out to learn how to predict the HR/FB rates for every team in the league. I grabbed these numbers and compared them to Park Factors for home runs allowed. There is a correlation there, but it's only .12, which means that (at least over 25 games) only 1.44% of a team's HR/FB rate is attributable to its park factors. Colorado has a home-run favoring park sure, as the White Sox do, but Seattle and Montreal are neutral and Pittsburgh has the second lowest HR park effect in the league. Not a huge indicator here.
What about Movement? Movement is made of two things: groundball rate and hr/fb, so surely there should be some sort of connection there. I mean, we've controlled for groundball rate (because we're only looking at flyballs) so movement should clearly be the controlling factor.
So I generated a stat I called IAM, Innings Adjusted Movement, which is basically every team's movement based on how many innings were pitched at what movement (an 8 movement pitcher with 40 innings is more heavily weighted than an 8 movement pitcher with 3 innings). And I ran this for every team.
Here are the same ten teams from the top and bottom of HR/FB lists, but with IAM rankings as well:
Colorado: #32 HR/FB, #3 IAM
Montreal: #31 HR/FB, #9 IAM
White Sox: #30 HR/FB, #29 IAM
Pittsburgh: #29 HR/FB, #1 IAM
Seattle: #28 HR/FB, #24 IAM
Washington: #1 HR/FB, #12 IAM
Cubs: #2 HR/FB, #8 IAM
Cleveland: #3 HR/FB, #18 IAM
Oakland: #4 HR/FB, #2 IAM
San Diego: #5 HR/FB, #10 IAM
Do you see a connection there? Because I don't. While Seattle and the White Sox have low team movement, Colorado and Pittsburgh (teams #3 and #1) in movement are two of the four worst teams at giving up home runs on flyballs.
I ran a correlation between IAM and HR/FB and it was literally zero. Over 25 games, for 32 teams, movement has meant exactly nothing with regards to what percentage of flyballs you allow turn into home runs.
In other words, teams allowing a lot of HR/FB are just getting unlucky. And teams allowing a low rate of HR/FB are just getting lucky.
Seriously, who knew Colorado had the 3rd lowest movement in the league? Not I.
Sorry, thought all this was interesting and wanted to share.
Seems like a pretty fair stat to me.
But another stat caught my eye. xFIP. I don't like this stat as much. Oakland is only #5 in it; Colorado is #17. What gives?
Some quick definitions:
FIP is basically a pitcher metric that only cares about four things: strikeouts, walks allowed, home runs allowed and innings pitched. It uses linear weights, the formula is basically (HRA * 13 + BB * 3 - K * 2) / IP + 3.2. It completely ignores defense and clutch pitching (coincidental or not).
xFIP is the same as FIP, but it basically assumes that every flyball is equally likely to be a home run, so it charges pitchers for every fly ball they allow, not for the home runs they allow.
So. League average in the PBL right now is that 12.93% of all flyballs become home runs.
Here are the teams that have the lowest HR / FB rates:
1) Washington (6.75%)
2) Chicago Cubs (7.94%)
3) Cleveland (8.21%)
4) Oakland (8.65%)
5) San Diego (9.13%)
Here are the teams that have the highest HR / FB rates:
1) Colorado (19.89%)
2) Montreal (18.82%)
3) White Sox (17.11%)
4) Pittsburgh (16.95%)
5) Seattle (16.75%)
That's kind of a weird pair of lists isn't it? I mean, I don't pretend to know every team's pitching quality, but I know that Pittsburgh has one of the top 2 rotations in the league.
So, as you may not be surprised to learn, I set out to learn how to predict the HR/FB rates for every team in the league. I grabbed these numbers and compared them to Park Factors for home runs allowed. There is a correlation there, but it's only .12, which means that (at least over 25 games) only 1.44% of a team's HR/FB rate is attributable to its park factors. Colorado has a home-run favoring park sure, as the White Sox do, but Seattle and Montreal are neutral and Pittsburgh has the second lowest HR park effect in the league. Not a huge indicator here.
What about Movement? Movement is made of two things: groundball rate and hr/fb, so surely there should be some sort of connection there. I mean, we've controlled for groundball rate (because we're only looking at flyballs) so movement should clearly be the controlling factor.
So I generated a stat I called IAM, Innings Adjusted Movement, which is basically every team's movement based on how many innings were pitched at what movement (an 8 movement pitcher with 40 innings is more heavily weighted than an 8 movement pitcher with 3 innings). And I ran this for every team.
Here are the same ten teams from the top and bottom of HR/FB lists, but with IAM rankings as well:
Colorado: #32 HR/FB, #3 IAM
Montreal: #31 HR/FB, #9 IAM
White Sox: #30 HR/FB, #29 IAM
Pittsburgh: #29 HR/FB, #1 IAM
Seattle: #28 HR/FB, #24 IAM
Washington: #1 HR/FB, #12 IAM
Cubs: #2 HR/FB, #8 IAM
Cleveland: #3 HR/FB, #18 IAM
Oakland: #4 HR/FB, #2 IAM
San Diego: #5 HR/FB, #10 IAM
Do you see a connection there? Because I don't. While Seattle and the White Sox have low team movement, Colorado and Pittsburgh (teams #3 and #1) in movement are two of the four worst teams at giving up home runs on flyballs.
I ran a correlation between IAM and HR/FB and it was literally zero. Over 25 games, for 32 teams, movement has meant exactly nothing with regards to what percentage of flyballs you allow turn into home runs.
In other words, teams allowing a lot of HR/FB are just getting unlucky. And teams allowing a low rate of HR/FB are just getting lucky.
Seriously, who knew Colorado had the 3rd lowest movement in the league? Not I.
Sorry, thought all this was interesting and wanted to share.