THE BOOK--Playing The Percentages In Baseball

http://www.baseballthinkfactory.org/files/newsstand/discussion/fangraphs_cameron_morgan_dunn/

http://www.baseballthinkfactory.org/files/newsstand/discussion/d_magazine_rangers_borrowed_15m_from_mlb_to_make_payroll/

http://www.baseballthinkfactory.org/files/newsstand/discussion/rany_on_the_royals_banned/

I thought that the best thing there was a discussion about Dave Cameron’s post at Fangraphs about Nyjer Morgan and Adam Dunn.

I linked to that, GGC; it’s stuck in moderation with a few other things.

Cool, thanks, I’ll check those out.

Welcome back, sure was quiet here without you.  Now we know what WOWY looks like at the Inside The Book Blog.

In which state or country did you spend your vacation money?

vr, Xei

Re: Dunn v Morgan.

I don’t have a specific comment about Morgan, but just the general offense v defense debate.

You can have the same discussion about Franklin Gutierrez as well, and even more so.  Gutierrez’s fielding is hailed even higher than Morgan.  He’s got one more full season under his belt while being 3 years younger.  And he’s a bit better as a hitter.  And you can reasonably call Gutierrez = or better than Dunn.

This is the same discussion I had two years ago re: Jeter v Polanco.  You can easily argue that the fielding gap between Gutierrez and Dunn is 3 wins.  You can reasonably argue it’s a 4 win gap.  You can possibly stretch that out to even a 5-win gap, but that’s stretching it.  It’s certainly far more than a 2 win gap and I would believe a 5-win gap far more than I would believe a 2-win gap.  That makes it closer to 4 wins than 3 wins.

At the same time, what’s the offensive gap between Dunn and Gutierrez?  Three wins?  Maybe 4?

You simply can’t say “he’s got a great glove” and then say “but it can’t be worth that much”.  And you can’t say “he’s got a terrible glove” and then say “but, it’s can’t be that costly”.

Yes it can.  A run saved is a run earned. 

Since 2002, according to Fangraphs, Mike Cameron is +28 WAR, while Adam Dunn is +21 wins.  Mike Cameron has been way better than Adam Dunn, and yet most people wouldn’t believe it.

Xei: Montreal.

***

The other thread, about Hicks: Gillette just sold the Montreal Canadiens and arena back to Molson’s.  Gillette and Hicks are part owners of a soccer team in England.  Borrowing money from the league is SOP in the NHL.  Coyotes are doing it.  Penguins did it last decade.  It’s not unusual.  I’m just surprised Hicks hasn’t sold the Rangers yet.

Thanks for the Rany link.  As it turned out, it may as well have been a dream, or some weekly TV show since there was no change, as it was a misunderstanding between Jack, Chrissy and Mr. Roper.

Well, I’m not sure this is one of the “best” primer threads, but it’s one where I thought of this blog and audience while reading it, wondering how they’d handle the issues raised:

http://www.baseballthinkfactory.org/files/newsstand/discussion/bleacher_report_is_luke_hochevar_a_bust/

The author uses some of the more recently available stats (LD rates, FIP, etc.) to look at Hochevar, and concludes H has been “unlucky.”

Let’s see.  His LOB rate is 64% (on 800 batters faced):

http://www.fangraphs.com/statss.aspx?playerid=6943&position=P

The average is around 70% or so, and the better you are, the higher LOB you have (naturally).  The worse you are, the more runners you allow to score.

The problem is that Hochevar has an average number of walks, HR, and BABIP (all more or less).  The only way for him to have such an obscenely low LOB rate is for him to be terrible with men on base:

http://www.baseball-reference.com/players/split.cgi?n1=hochelu01&year=Career&t=p#bases

And lookie, lookie: his tOPS+ is 78 with bases empty and 127 with men on base (where “100” is set as his overall tOPS+, regardless of what his OPS+ actually is… i.e., his 78 and 127 is being compared to himself).

So, he has been HORRIBLE with men on base.  If there is a physical reason for this (stretch v windup), I’d like to hear it.  If there is a mental reason for this (he panics with men on base), I’d like to hear it.  Otherwise, if there is no prior reason for him to have this horrible a split, then it must be bad luck.  Or it could be a combination of all three.  Since we don’t know, the only thing we can do is regress, and presume that he comes from the population of ballplayers that in fact does show actual split differences (i.e., he is human), but that those split differences are not very pronounced at all.

Jorge de la Rosa is another guy having an awful time with men on base.  Very few PA’s, but it gets worse the more men that are on (tOPS is two digits with a single man on, but with 2-3, it shoots up to 150-250).  So it’s not the stretch per se. It seems like he’s had this pattern both years with the Rockies, but not so much in previous years. 

I wonder if it is random variation, or are the Rockies missing something pretty important. At any rate, it kind of stinks that a guy striking out 9.4/9 with a decent HR rate in Coors has an ERA of over 5.

Tango, mgl, or some other smart math person:

Did David Cameron regress the defensive stats for Morgan enough in his analysis? I think he regressed his numbers for maybe a rest-of-season analysis but not necessarily a true talent analysis (i.e. added 2/3 of average to Morgan’s 1/3 SS of PT in center whereas I think it should’ve been like 1/6 PT v. 5/6 of average because I like to regress heavier [rightly or wrongly] when the sample size is puny.

I don’t know entirely if I make sense and there’s probably an issue with the weighting (I have no idea what defensive weights should be in a Marcel-like projection) I put up in the previous paragraph, but some feedback would be awesome.

I add 400 BIP of league average.  That would mean 100 games in CF or 133 in the corners.  For Morgan, with 97 games in the corners and 53 in CF, that means adding 121 games.

Because of the mismatch in CF and corners, the “league average” is not 0.  It’ll be -10 times 97 plus 0 times 53 for an average of -6.5 runs (in CF).  (If you were evaluating him in the corners, you’d be adding +3.5 runs.)

He has a total UZR of 31 in a total of 150 games.  You add -6.5 and 121 respectively to get 24.5 runs in 271 games, or a regressed rate of +13.6 runs per 150 G.

So, Morgan counts as +14 in CF.

Oops… we are adding -6.5 per 150.  But, since we are adding 121 games, we need to add -5.2. 

Still a +14 overall.

Ok, I think I am all caught-up, linked-up, read-up, commented-up and inboxed-up.

If there is anything outstanding (I didn’t reply to someone, or didn’t link to something I could have), let me know.

Thanks Tango. I’m not sure what I didn’t with my regression (attempt) using innings played instead of BIP, but I think I might’ve regressed too much and/or didn’t include the data from the corner outfield.