Posted by oriolemagic on 2/21/2012 5:02:00 PM (view original):
Did you read anything Bobzilla wrote?  Did you read what I wrote?  No matter how you look at the advanced stats..

Advanced Stats are only advanced for real life. There's a lot of trial and error that went into them. Assuming that they apply here is a huge leap of faith. In real life, recent FIP is a bettor predictor then recent ERA of near-future ERA. I don't believe this is true in HBD, and in any event, no one has really looked into it, so you're just assuming that it still applies. I'm convinced that BABIP works differently in HBD and real-life based advanced stats that make assumptions about BABIP (as FIP does) are to be dismissed out of hand for HBD.

Posted by kcden on 2/21/2012 5:45:00 PM (view original):

Posted by JFerg on 2/21/2012 5:42:00 PM (view original):

Posted by kcden on 2/21/2012 5:24:00 PM (view original):

Of the three stats that FIP looks at (BB+HBP, K, and HR), home runs are the least stable from year-to-year.
 

Boy, that's a bit of a red flag to me, as far as saying that FIP has great value, given HR is weighted by far the heaviest in the equation.

It's weighted the heaviest because HR have that much more of an impact than a BB or K.  You could also look at xFIP which replaces a pitchers HR allowed with the amount a league average pitcher would have given up with the same amount of flyballs.  SIERA and tERA/tRA are two other advanced metrics that do take into account balls in play.  You can also look at Bill James' component ERA.  I pretty much look at them all.

Bobzilla

Yeah, I get why it is weighted heaviest, but if it is by far the least stable from year to year (which makes sense), and is that heavily weighted, that tells me that equation is less valuable or accurate at actually predicting anything.

Not true.

http://www.insidethebook.com/ee/index.php/site/comments/correlation_of_pitcher_metrics_fip_strikes_again/

Bobzilla

If that's your reasoning, you're using a short posting by the gentleman that created the FIP stat to prove that FIP accurately predicts the future... and still that is just predicting the future, which says nothing of this season's Cy Young

I chatted this discussion with a friend earlier today; he had an interesting take.  I don't necessarily agree with it, but it was interesting. 
First off, he said Shelson is the Cy Young, no question.  
I then poinited out the extra 81 innings for Ducey as the reason folks were arguing he should have won. 
His response was something along the lines of  "[that guy is bullshit.  Completely unrealistic player]"  since no one throws nearly that many innings these days, at a high level of success or not.

So now we determine Cy Young based on how realistic a player is ;)

Posted by oriolemagic on 2/21/2012 6:39:00 PM (view original):
So now we determine Cy Young based on how realistic a player is ;)

I said I don't necessarily agree with it... it is an interesting take, though.  And my friend does play HBD, so I would assume he would take that into account if he was looking at Ducey in voting (although I'm sure he wouldn't reject taking Ducey on his team, due to how unrealistic he is)

Posted by kcden on 2/21/2012 6:29:00 PM (view original):
If that's your reasoning, you're using a short posting by the gentleman that created the FIP stat to prove that FIP accurately predicts the future... and still that is just predicting the future, which says nothing of this season's Cy Young

Tango didn't do the testing.  You mentioned predicting.  What does one predict?  The future.  As for this particular season, you can make a perfectly rational argument that Ducey was better than Sheldon.

Bobzilla

Posted by JFerg on 2/21/2012 7:25:00 PM (view original):

Posted by kcden on 2/21/2012 6:29:00 PM (view original):
If that's your reasoning, you're using a short posting by the gentleman that created the FIP stat to prove that FIP accurately predicts the future... and still that is just predicting the future, which says nothing of this season's Cy Young

Tango didn't do the testing.  You mentioned predicting.  What does one predict?  The future.  As for this particular season, you can make a perfectly rational argument that Ducey was better than Sheldon.

Bobzilla

I agree you can make a rational argument for Ducey; I just disagree with the conclusion that Ducey's season was better.
As far as my point with FIP being predictive, my point is that the way I see FIP being used, it is useful some of the time in predicting if a player is "due for a course correction" but is not a great indicator of how effective or "good" or valuable of a season a player had.... and still we don't know if it is at all valid for HBD.

Posted by JFerg on 2/21/2012 5:59:00 PM (view original):

Posted by kcden on 2/21/2012 5:45:00 PM (view original):

Posted by JFerg on 2/21/2012 5:42:00 PM (view original):

Posted by kcden on 2/21/2012 5:24:00 PM (view original):

Of the three stats that FIP looks at (BB+HBP, K, and HR), home runs are the least stable from year-to-year.
 

Boy, that's a bit of a red flag to me, as far as saying that FIP has great value, given HR is weighted by far the heaviest in the equation.

It's weighted the heaviest because HR have that much more of an impact than a BB or K.  You could also look at xFIP which replaces a pitchers HR allowed with the amount a league average pitcher would have given up with the same amount of flyballs.  SIERA and tERA/tRA are two other advanced metrics that do take into account balls in play.  You can also look at Bill James' component ERA.  I pretty much look at them all.

Bobzilla

Yeah, I get why it is weighted heaviest, but if it is by far the least stable from year to year (which makes sense), and is that heavily weighted, that tells me that equation is less valuable or accurate at actually predicting anything.

Not true.

http://www.insidethebook.com/ee/index.php/site/comments/correlation_of_pitcher_metrics_fip_strikes_again/

Bobzilla

From fangraphs:

- FIP does a better job of predicting the future than of measuring the present, meaning there can be a lot of fluctuation in smaller samples.

- xFIP has the highest correlation with future ERA of all the pitching metrics.

Posted by oriolemagic on 2/21/2012 6:39:00 PM (view original):
So now we determine Cy Young based on how realistic a player is ;)

No.  We just take your word on it while ignoring WHIP, ERA, BB9, K9 and OAV.

Posted by jvford on 2/21/2012 8:23:00 PM (view original):

Posted by JFerg on 2/21/2012 5:59:00 PM (view original):

Posted by kcden on 2/21/2012 5:45:00 PM (view original):

Posted by JFerg on 2/21/2012 5:42:00 PM (view original):

Posted by kcden on 2/21/2012 5:24:00 PM (view original):

Of the three stats that FIP looks at (BB+HBP, K, and HR), home runs are the least stable from year-to-year.
 

Boy, that's a bit of a red flag to me, as far as saying that FIP has great value, given HR is weighted by far the heaviest in the equation.

It's weighted the heaviest because HR have that much more of an impact than a BB or K.  You could also look at xFIP which replaces a pitchers HR allowed with the amount a league average pitcher would have given up with the same amount of flyballs.  SIERA and tERA/tRA are two other advanced metrics that do take into account balls in play.  You can also look at Bill James' component ERA.  I pretty much look at them all.

Bobzilla

Yeah, I get why it is weighted heaviest, but if it is by far the least stable from year to year (which makes sense), and is that heavily weighted, that tells me that equation is less valuable or accurate at actually predicting anything.

Not true.

http://www.insidethebook.com/ee/index.php/site/comments/correlation_of_pitcher_metrics_fip_strikes_again/

Bobzilla

From fangraphs:

- FIP does a better job of predicting the future than of measuring the present, meaning there can be a lot of fluctuation in smaller samples.

- xFIP has the highest correlation with future ERA of all the pitching metrics.

Is this supposed to contradict something I said?  Otherwise I don't get the point of your post.

Bobzilla

Yep

Posted by jvford on 2/21/2012 8:54:00 PM (view original):
Yep

But my post and your post are saying the same thing.  FIP is predictive of a pitcher's ERA the following season.  xFIP was slightly lower than FIP in the test that I linked. tERA was slightly higher, SIERRA was slightly lower.  However, they were all just about equal.  What is great about FIP is it's simplicity.

Bobzilla

Posted by JFerg on 2/21/2012 9:15:00 PM (view original):

Posted by jvford on 2/21/2012 8:54:00 PM (view original):
Yep

But my post and your post are saying the same thing.  FIP is predictive of a pitcher's ERA the following season.  xFIP was slightly lower than FIP in the test that I linked. tERA was slightly higher, SIERRA was slightly lower.  However, they were all just about equal.  What is great about FIP is it's simplicity.

Bobzilla

Why are you involved in this discussion? You don't play HBD, right? Advanced Metrics are not applicable here, because the sim runs differently than real life in hundreds of little ways, and we have no Advanced Metrics developed and tested for HBD.

But why use a "predictive" stat to compare a season that is already complete? They are useless to evaluate what ACTUALLY OCCURRED. Those are the stats that matter in award voting.

That is the problem. What the "advanced stats" show ME, is that Ducey had the season of his career, while Sheldon had a slightly "down" year. 

I vote for Ducey, due to his park and I think if he doesn't win it this year he may never win it at all. I am not the norm though. Most owners who bother to vote, look at straight line stats. I reward Ducey for having his career year.