Posted by kcden on 2/21/2012 11:12:00 AM (view original):
Alright, just read the definition of "FIP"... how the hell is an HR the only type of hit that a pitcher is responsible for? Similarly, how does a pitcher take no responsibility for any out other than a K?


If FIP is all you want to take into account when evaluating a pitcher, you are a ******* retard.

"(HR*13+(BB+HBP-IBB)*3-K*2)/IP, plus a league-specific factor (usually around 3.2) to round out the number to an equivalent ERA number

In real life, A pitcher's ERA next year is more likely to mirror his FIP last year than his ERA last year.  This has been tested and is true.  I prefer xFIP or SIERRA to FIP, actually.  But all of these are attempts to strip out factors like luck (above or below average rate of seeing eye singles, or hard-hit shots right at a defender), as well as non-pitcher factors (mostly defense, which, studies have shown, has a much lasrger than generally understood impact on pitching outcomes), and come up with a predictor for future performance.

FIP, xFIP, and SIERA are all called "ERA Estimators".  The idea is that if a pitcher were to pitch 100 seasons of the same quality as the season in question, with a neutral defense behind him, against league average hitters, in a neutral park context, this is what his ERA would likely work out as, over that very large sample size.

And most stat guys would tell you that xFIP is much more accurate of a predictor than FIP. 

Of course, neither is useful in HBD.

Posted by kcden on 2/21/2012 11:18:00 AM (view original):

Also, I don't know how it would change the evaluation of FIP, but I think it is incredibly meaningful within WIS (if you are going to use FIP) the order in which outcomes are determined.  IIRC, it goes:  Is it a hit?  Yes> Was the hit taken away by a good/great play, Was the hit a Single/Double/Triple/HR, Was Double/Triple/HR reduced to something less by a great play, etc... or something like that.
I may be thinking of this incorrectly, but the fact that the determination of "Hit" is made before "HR" makes an HR less damning to a pitcher's production... i.e. it is quite possible that a pitcher that gives up a lot of HR can "always" have so few runners on base that an ERA/WHIP like Sheldon's with a high HR rate makes sense, and is not just luck.

This is sort of what I was getting at.  I don't think formulas like FIP, xFIP, or SIERA, or on the offensive side, wOBA or OPS+, have much bearing on HBD because a lot of the assumptions behind them, which have been tested in real life and work well when evaluating real life, don't apply to the way a PA is broken down by the HBD Sim.  I actually don't think hit is determined before HR in the sim, by the way.  I think the Sim for a PA works roughly in the order, HBP, BB, K, HR, type of ball in play, outcome of ball in play.   But the point is, we don't really know and we would need to understand this better to come up with a Sabermetrics for HBD.     

ERA Higher Than FIP in 2009
Player	ERA	FIP	Diff
Ricky Nolasco	5.06	3.35	1.71
Carl Pavano	5.10	4.00	1.10
Livan Hernandez	5.44	4.44	1.00
Justin Verlander	3.45	2.80	0.65
Mike Pelfrey	5.03	4.39	0.64
Jason Hammel	4.33	3.71	0.62
Derek Lowe	4.67	4.06	0.61
Paul Maholm	4.44	3.83	0.61
Cole Hamels	4.32	3.72	0.60
Jorge De La Rosa	4.38	3.91	0.47

 

ERA Lower Than FIP in 2009
Player	ERA	FIP	Diff
J.A. Happ	2.93	4.33	-1.40
Kevin Millwood	3.67	4.80	-1.13
Jair Jurrjens	2.60	3.68	-1.08
Matt Cain	2.89	3.89	-1.00
Bronson Arroyo	3.84	4.78	-0.94
Randy Wells	3.05	3.88	-0.83
John Danks	3.77	4.59	-0.82
John Lannan	3.88	4.70	-0.82
Rick Porcello	3.96	4.77	-0.81
Ross Ohlendorf	3.92	4.72	-0.80

This chart makes me a bit dubious of FIP... largely due to Nolasco, as I recall him ALWAYS being listed as something like "unlucky" in Fantasy Baseball articles.  I have no doubt that FIP has its uses (Fantasy Baseball seems an appropriate use, as well as an additional (not "The Key") tool in evaluating a player for an award/contract/etc.), but disregarding the fact that a pitcher gets batters out other than with a K and allows hits other than HR that are partially or entirely their "fault" is silly (also, for that formula to follow it's own intent, Inside the Park HR's need to be removed).

Posted by gjello10 on 2/21/2012 11:45:00 AM (view original):

Posted by kcden on 2/21/2012 11:18:00 AM (view original):

Also, I don't know how it would change the evaluation of FIP, but I think it is incredibly meaningful within WIS (if you are going to use FIP) the order in which outcomes are determined.  IIRC, it goes:  Is it a hit?  Yes> Was the hit taken away by a good/great play, Was the hit a Single/Double/Triple/HR, Was Double/Triple/HR reduced to something less by a great play, etc... or something like that.
I may be thinking of this incorrectly, but the fact that the determination of "Hit" is made before "HR" makes an HR less damning to a pitcher's production... i.e. it is quite possible that a pitcher that gives up a lot of HR can "always" have so few runners on base that an ERA/WHIP like Sheldon's with a high HR rate makes sense, and is not just luck.

This is sort of what I was getting at.  I don't think formulas like FIP, xFIP, or SIERA, or on the offensive side, wOBA or OPS+, have much bearing on HBD because a lot of the assumptions behind them, which have been tested in real life and work well when evaluating real life, don't apply to the way a PA is broken down by the HBD Sim.  I actually don't think hit is determined before HR in the sim, by the way.  I think the Sim for a PA works roughly in the order, HBP, BB, K, HR, type of ball in play, outcome of ball in play.   But the point is, we don't really know and we would need to understand this better to come up with a Sabermetrics for HBD.     

+ 1 Million.

I'm not sure i would have K separate from type of out though, if it was evaluated separately i would expect high K pitchers to be more dominant than an equivalent soft tosser, and i dont think you can definitively show that.

Posted by kcden on 2/21/2012 11:55:00 AM (view original):

ERA Higher Than FIP in 2009
Player	ERA	FIP	Diff
Ricky Nolasco	5.06	3.35	1.71
Carl Pavano	5.10	4.00	1.10
Livan Hernandez	5.44	4.44	1.00
Justin Verlander	3.45	2.80	0.65
Mike Pelfrey	5.03	4.39	0.64
Jason Hammel	4.33	3.71	0.62
Derek Lowe	4.67	4.06	0.61
Paul Maholm	4.44	3.83	0.61
Cole Hamels	4.32	3.72	0.60
Jorge De La Rosa	4.38	3.91	0.47

 

ERA Lower Than FIP in 2009
Player	ERA	FIP	Diff
J.A. Happ	2.93	4.33	-1.40
Kevin Millwood	3.67	4.80	-1.13
Jair Jurrjens	2.60	3.68	-1.08
Matt Cain	2.89	3.89	-1.00
Bronson Arroyo	3.84	4.78	-0.94
Randy Wells	3.05	3.88	-0.83
John Danks	3.77	4.59	-0.82
John Lannan	3.88	4.70	-0.82
Rick Porcello	3.96	4.77	-0.81
Ross Ohlendorf	3.92	4.72	-0.80

This chart makes me a bit dubious of FIP... largely due to Nolasco, as I recall him ALWAYS being listed as something like "unlucky" in Fantasy Baseball articles.  I have no doubt that FIP has its uses (Fantasy Baseball seems an appropriate use, as well as an additional (not "The Key") tool in evaluating a player for an award/contract/etc.), but disregarding the fact that a pitcher gets batters out other than with a K and allows hits other than HR that are partially or entirely their "fault" is silly (also, for that formula to follow it's own intent, Inside the Park HR's need to be removed).

So you find ERA and WHIP works better in evaluating a pitchers performance?

The question then is what stats do you want to use to judge value in HBD

I think the combination of ERA/WHIP in a given year is a pretty good indication of how good they were THAT YEAR.  For predicting future performance, it is all over the charts, just like FIP appears to be (probably more-so, but I haven't studied large groups of players for FIP vs. ERA, and don't plan to do so... maybe I'll look at it a bit more before my fantasy auction this season and see where it gets me).

I judge value in HBD based on ratings.  If I see players that consistently produce below what I think their ratings say they should, I re-evaluate.

Posted by MikeT23 on 2/21/2012 11:25:00 AM (view original):
Yeah, I said the same thing about FIP too.    Plus batters can be just as responsible, if not moreso, for a K/BB/HR. 

In real life, the idea is that over a large sample pitchers tend to face an aggregate batter who is of roughly league average quality.  So you ignore that part of the equation because it should be (roughly) equal for most pitchers.  What you're really trying to tease out here is how much the good/bad defense behind a pitcher is impacting the top-line ERA number.  The key understanding here is that, in real life, almost all pitchers work in a fairly narrow range of BABIP over a large enough sample (something like .280-.310), and also that LOB% tends to normalize for all pitchers over a large enough sample. 

So when you see Verlander with a career-best 2.40 ERA in 2011, but a FIP of 2.99 and an xFIP of 3.12, you can ask yourself what's going on there.  You then see that his BABIP for 2011 was .236, compared to a career BABIP of .285, and that his LOB% was 80.3% in 2011, compared to a career LOB% of 73.3% (which is roughly normal for a good pitcher).  So he was getting lucky in the placement of Balls In Play against him, he was getting a bit lucky in the way the hits against him spread out through the order, allowing the higher Strand Rate (LOB%), and the defense likely played well behind him.  For reference, the best modern-era SP in terms of career BABIP is Catfish Hunter, at .243, so we can all expect Verlander's ERA for 2012 to regress towards his FIP/xFIP/SIERA numbers, which all show that he pitched at a 2.99-3.12 ERA rate, from the 2.40 ERA he posted last year based on an unsustainable BABIP.

Posted by kcden on 2/21/2012 12:05:00 PM (view original):
I judge value in HBD based on ratings.  If I see players that consistently produce below what I think their ratings say they should, I re-evaluate.

Sorry, for awards purposes.  I judge players solely on ratings as well.

Posted by kcden on 2/21/2012 12:03:00 PM (view original):
I think the combination of ERA/WHIP in a given year is a pretty good indication of how good they were THAT YEAR.  For predicting future performance, it is all over the charts, just like FIP appears to be (probably more-so, but I haven't studied large groups of players for FIP vs. ERA, and don't plan to do so... maybe I'll look at it a bit more before my fantasy auction this season and see where it gets me).

So back the original question, how do you value the extra 81 innings when ERA/WHIP do not take them into account?

Posted by gjello10 on 2/21/2012 12:05:00 PM (view original):

Posted by MikeT23 on 2/21/2012 11:25:00 AM (view original):
Yeah, I said the same thing about FIP too.    Plus batters can be just as responsible, if not moreso, for a K/BB/HR. 

In real life, the idea is that over a large sample pitchers tend to face an aggregate batter who is of roughly league average quality.  So you ignore that part of the equation because it should be (roughly) equal for most pitchers.  What you're really trying to tease out here is how much the good/bad defense behind a pitcher is impacting the top-line ERA number.  The key understanding here is that, in real life, almost all pitchers work in a fairly narrow range of BABIP over a large enough sample (something like .280-.310), and also that LOB% tends to normalize for all pitchers over a large enough sample. 

So when you see Verlander with a career-best 2.40 ERA in 2011, but a FIP of 2.99 and an xFIP of 3.12, you can ask yourself what's going on there.  You then see that his BABIP for 2011 was .236, compared to a career BABIP of .285, and that his LOB% was 80.3% in 2011, compared to a career LOB% of 73.3% (which is roughly normal for a good pitcher).  So he was getting lucky in the placement of Balls In Play against him, he was getting a bit lucky in the way the hits against him spread out through the order, allowing the higher Strand Rate (LOB%), and the defense likely played well behind him.  For reference, the best modern-era SP in terms of career BABIP is Catfish Hunter, at .243, so we can all expect Verlander's ERA for 2012 to regress towards his FIP/xFIP/SIERA numbers, which all show that he pitched at a 2.99-3.12 ERA rate, from the 2.40 ERA he posted last year based on an unsustainable BABIP.

<sarcasm> But how many games did he win? </sarcasm>

Thats the type of question your going to have to deal with when trying to explain FIP or xFIP or any other advanced stat to this group

The objective of baseball:  Run scoring/prevention.   Secondary objective(which helps with the primary):  Prevent/get baserunners.

So, yeah, I think ERA/WHIP are pretty good indicators of how a pitcher performed.   You'll seldom find a low WHIP/high ERA or vice versa.

I don't think all hits, or homers, are created equal.   A 500 ft bomb appears the same as a ball off the pole.    To use FIP as some fantastic evaluator of pitching, you have to suspend all rational thought that pitchers can make a perfect pitch and have it deposited in the seats or have that same pitch be a weak grounder back to the mound. 

Posted by oriolemagic on 2/21/2012 12:08:00 PM (view original):

Posted by gjello10 on 2/21/2012 12:05:00 PM (view original):

Posted by MikeT23 on 2/21/2012 11:25:00 AM (view original):
Yeah, I said the same thing about FIP too.    Plus batters can be just as responsible, if not moreso, for a K/BB/HR. 

In real life, the idea is that over a large sample pitchers tend to face an aggregate batter who is of roughly league average quality.  So you ignore that part of the equation because it should be (roughly) equal for most pitchers.  What you're really trying to tease out here is how much the good/bad defense behind a pitcher is impacting the top-line ERA number.  The key understanding here is that, in real life, almost all pitchers work in a fairly narrow range of BABIP over a large enough sample (something like .280-.310), and also that LOB% tends to normalize for all pitchers over a large enough sample. 

So when you see Verlander with a career-best 2.40 ERA in 2011, but a FIP of 2.99 and an xFIP of 3.12, you can ask yourself what's going on there.  You then see that his BABIP for 2011 was .236, compared to a career BABIP of .285, and that his LOB% was 80.3% in 2011, compared to a career LOB% of 73.3% (which is roughly normal for a good pitcher).  So he was getting lucky in the placement of Balls In Play against him, he was getting a bit lucky in the way the hits against him spread out through the order, allowing the higher Strand Rate (LOB%), and the defense likely played well behind him.  For reference, the best modern-era SP in terms of career BABIP is Catfish Hunter, at .243, so we can all expect Verlander's ERA for 2012 to regress towards his FIP/xFIP/SIERA numbers, which all show that he pitched at a 2.99-3.12 ERA rate, from the 2.40 ERA he posted last year based on an unsustainable BABIP.

<sarcasm> But how many games did he win? </sarcasm>

Thats the type of question your going to have to deal with when trying to explain FIP or xFIP or any other advanced stat to this group

Yes, we're obviously not as smart as you.

Sarcasm.