April 15, 2013 posted by Patrick DiCaprio

Closer Identifier Algorithm Results for April 15

Closer Identifier Algorithm Results for April 15
Print Friendly

All statuses are based on stats through Sunday’s games.

On twitter, Eric Mack of raised an issue that evinced a misunderstanding of what this experiment is about, but his question was a good one, worthy of an explanation. The predictions are put out on a weekly basis. But that does not mean that CIA predicts a player will lose a job that week. It is a convenience borne out of the fact that a) I am not going to do this every day because I am not that dedicated, and b) I have a column to write every week. These are not “predictions” but descriptions of a pitcher’s status as a closer, if that makes sense.

The driving idea behind CIA is that it is a substitute for a clinical analysis, and the thesis is that it will make better predictions than a traditional clinical analysis, as proven in the Philip Tetlock Study. There is no difference between Fantasy Baseball pundits and political/economics/business/government pundits. So why should we view ourselves as being able to make better predictions than these so-called “experts?” CIA is a way to challenge the idea that fantasy baseball pundits can make better predictions that the masses.

When you or I think to ourselves after reading an article or watching a game that “Mitchell Boggs is going to lose his job,” do you mean that he will lose it this week? Clearly not. And your opinion will change over time as Boggs’, or anyone else’s, performance changes. This is what CIA does; it responds to changing information on a regular basis in the same way that we do and makes predictions the same way that we do. It is not a compilation of weekly projections for that week.

If it were a set of weekly projections its record would be deceiving, just as yours would be, and that is not what it is about. At the end of the year, for example, there were 35 possible statuses among relievers. It would be silly to say “each week it got 29 right and so in September its record was 116-24.” That is not what we do in our own clinical analyses and it would be very easy to approach 100% if that were the methodology and the goal. But it would be wrong, not a reflection of reality and would defeat the stated purpose above.

The columns have done a lot of explanations of what this experiment is about and this point was addressed early on, but it is worth mentioning again. As for trying to predict who will get a job, my view is that it is a fool’s errand for the most part, and I am not going to make bad predictions or give bad advice based on speculation merely because readers may want it. I leave that to the foolish. I try to add players in when they are discussed as having a share of the job, but do not rate them as a “hit” or “miss” unless they actually have a share. So, I guess I am a little foolish as well.

On to the current statuses:

Pitcher CIA Result Comments
Jim Johnson Hold Just keeps on converting saves like clockwork.
Joel Hanrahan Lose Yeah right, his hamstring is hurt.
Andrew Bailey Hold
Addison Reed Hold
Chris Perez Lose SV/SVO only 50% but probably safe
Joaquin Benoit Lose Off to a slow start but nominally the closer. Status change likely.
Phil Coke Lose Not ready to call this a “hit” yet for CIA
Jose Veras Lose Do you really care?
Greg Holland Lose No longer automatic “lose” based on SV/SVO but in short sample anything can happen. Not ready to call it a “hit.”
Kelvin Herrera Hold Right now he looks like Kimbrel and Chapman.
Ernesto Frieri Hold
Glen Perkins Hold
Mariano Rivera Hold
Grant Balfour Hold SV/SVO is 100% but only one opportunity
Tom Wilhelmsen Hold Right now he is as good as anyone.
Fernando Rodney Lose SV/SVO is 50% so still has a ‘lose” status.
Joe Nathan Hold
Casey Janssen Hold
National League
JJ Putz Hold
Craig Kimbrel Hold
Kyuji Fujikawa Hold 4.00 K/BB and 66% SV/SVO
Aroldis Chapman Hold
Rafael Betancourt Hold A perfect 5/5 in SV/SVO
Brandon League Hold Gets an auto “hold” at 4/4 in SV/SVO but skills still terrible.
Steve Cishek Lose Still zero saves, but skill set is fine so he will likely get a “hold” shortly.
Jim Henderson Hold So far so good.
Bobby Parnell Hold Zero BB so far.
Jonathan Papelbon Hold
Jason Grilli Hold Off to a surprising start
Huston Street Hold
Sergio Romo Hold As good as there is in baseball
Mitchell Boggs Lose Hasn’t officially lost the job as of this writing.
Trevor Rosenthal Hold A 10.00 K/BB (!)
Rafael Soriano Hold Off to a strong start.
Carlos Marmol
John Axford

You must be logged in to post a comment.