THE BOOK--Playing The Percentages In Baseball

Interesting piece to start by Vollman.  Comments are not yet enabled, so maybe he’ll reply here.

I’m wondering how he adjusts for different offensive levels.  MC79 had a recent piece about how adjusting by straight GPG will overcorrect for offensive performance in high offensive eras.

Iain’s piece on PP uses straight GPG adjustments for offensive era, so I guess Vollman’s does too.

Concerns are correct, but conclusion is not good:
http://www.puckprospectus.com/article.php?articleid=20

There are ways to adjust for plus/minus.  I have my own way.  Hawerchuk has his own way at BehindTheNet.ca .  82games.com has their way for basketball.  There are ways.  Many ways. 

To liken plus/minus to RBIs is ridiculous.

All data is about understanding the context.  Plus/minus, at its basis, is about as perfect a stat as you’d like.  Now, all you have to do is understand the context so you can adjust for it.

Iain has a good article, and you can probably expect more at that level:
http://www.puckprospectus.com/article.php?articleid=19

Here’s his archives:
http://hockeythink.com/research.html

You’ll note that they link to “Puckerings.com”.  That was his old site before he sold it.  You’ll have to manually change the links.

Here for example is Iain and Rob’s article on Plus/Minus, which makes good sense:
http://hockeythink.com/research/plusmin.html

Hopefully, they will craft their response to their colleague’s work appropriately, and not be all rah-rah about it.

Seppa misses a big concern: he doesn’t mention the effect on +/- of starting a shift with offensive vs defensive zone face-offs.

+/- actually has a relatively high year-to-year correlation (P/TOI is a bit better.) A lot of that’s usage, so I’m interested to see if he has a system to suss out the neutral +/-.

Given the lack of discussion of adjustments, and the comparison to RBIs, I would expect nothing more here.  Iain and Rob, whose article from several years ago I linked to in #4, are on the right path.

Any contact info for these guys?  PP’s contact form isn’t working right now, and they don’t have comments enabled.  I noticed a not-really-error in Tom Awad’s first column that he’d probably like to have knowledge of.

Join the Yahoo group HAG_List .  We’re all on that one.

I really do need to finish migrating the site over to hockeythink.com.

For the record, I’m using a straight GPG adjustment for offensive level because precision is not the main concern at this point. Rough is good enough for now.

Thanks for stopping by Iain, and congratulations on your efforts.

Readers will soon learn what a pleasure it will be to read your work.

As a relatively new fan to hockey (third year closely following the Blackhawks), I’m hoping Puck Prospectus can help me figure out who is actually good. Sure, you can use the eye test, but I always feel like I’m missing something. I just don’t have the grasp that I have on baseball. I’ll be following regularly.

Here’s Alan Ryder’s site:
http://hockeyanalytics.com/

He’s got links to lots of great site as well, including Hawerchuk’s.  You’ll like that one, as well as hockeynumber, mc79hockey and a few others.

Hi, and thanks for the interest in my article. 

There ARE many problems with plus/minus as I demonstrated with the article.  I’m surprised you don’t see that.  I’m not saying there isn’t any correlation to skill, just like I’m not saying that there isn’t any correlation between RBI and skill.  But there have to be better measures.  As of now, you have to look at plus/minus in context.  Ideally, with the right metric, you wouldn’t have to.

Give me a few articles and see where it goes!

I would love to comment at Puck Prospectus.  A little birdie told me that comments should be active by Monday.  Here’s to hoping!

Alan Ryder’s work, btw, is great!

Oh, and regarding where your shift starts - I don’t doubt that is true if you are looking at a single shift.  But over the course of a year this is something that must certainly even out for all skaters.

Iain - thanks for the clarification.  Looking forward to seeing what you guys come up with.

But there have to be better measures.  As of now, you have to look at plus/minus in context.  Ideally, with the right metric, you wouldn’t have to.

There isn’t a single metric around, anywhere, that meets your definition of ideal.  EVERY metric requires an understanding of the context.

I understand completely the biases in plus/minuses. But accounting for those biases is not that hard to do.  I cited several people who attempt to do so, including two of your colleagues.

Yes, of course stats require context.  That’s semantics.  What I want is a stat that includes as much detail as is necessary - PP, ES, SH contributions, both offense and defense, drawing and taking penalties - and not any more than is necessary.  Adjusts for TOI.  Adjusts for opposition and linemates on a per event basis.  Does not have the inaccuracies that some of the attempts out there have.

A stat that would do all of that would be very comprehensive.  Would it, like any other stat, require context?  Yes, but that’s semantics.  Would it require as much context as conventional plus/minus or even the very good attempts on improvement that are out there?  Hopefully, not - As it would encapsulate a lot of that context in itself.

Can it hurt to try to improve what’s out there?

Hey Tango,

I’m glad to see you like the work we’re doing at Puck Prospectus! We look forward to hearing from you in the future!

Andrew

Timo: when I read your article, I read it as if you were throwing your hands up in the air on the matter of plus/minus.  Can you read Iain/Rob’s article on the subject (see link 4) and tell me if you think that is a step in the right direction?

Andrew: be careful what you wish for.  Some people like what I do, and some don’t.

Tango,

What do you mean by that? Are you saying you’ve received threats because of applying statistics to sports? People need to get over themselves. What you’re doing for ALL sports is great! We also have Leslie as to provide scouting (formerly a Rangers scout for 6 years) to balance the team. All of us in this forum are not trying to start with the ‘old school’ culture. We’re just trying to provide an alternative analysis.

Andrew

No, not threats, not at all.

I’m saying that everything is lovey-dovey, until the recipient can’t take my (above the belt) punches.  Some people like a good scrap, and others don’t.

Tango,

I love a good scrap. Criticism is always warranted. If you ever see flaws in what I write, let me know. I know Gilbert of BehindTheNet didn’t think I sufficiently explained everything to bring in readers who only have a basic knowledge of hockey statistics. I will explain things more in depth next week. I’m more of a Joe Sheehan than a Nate Silver, Keith Woonlner or Dan Fox.

Andrew

Timo/13: “Oh, and regarding where your shift starts - I don’t doubt that is true if you are looking at a single shift.  But over the course of a year this is something that must certainly even out for all skaters.”

That does not happen.  It’s all about usage, and coaches put different players in for offensive and defensive zone draws.  See: http://vhockey.blogspot.com/2007/12/best-in-game.html

"Gilbert “ should be Gabriel, who is Hawerchuk in this blog.

In addition to Gabriel and Alan Ryder, I love hockeynumbers.com, Vic Ferrari, and mc79hockey.com

Sorry. I apologize to Gabe for that one. Those are all solid guys. I like the Falconer’s blog and Martin Brodeur is a Fraud as well.

Nice piece by Iain today on intangibles re: Alexandre Daigle.  I talked to a company last year that does personality testing of junior and college players for various NHL teams.  It sounds like they have a pretty big market - I just wonder if there are still teams that don’t pay attention to these things…

@ Hawerchuk: I did a table of “expected OZ draws compared to actual OZ draws” for the Thrashers and Kovalchuk’s number was closer to the team average than I would have expected. I have NOT looked at this data league wide, but I’m not sure that start-of-shift locations would have a huge bias effect on plus/minus--but it is worth some investigation.

@ Rothstein: I tired to send you an email about draft stuff but I’m not sure it went through on P.P. yesterday.

Matt,

The P.P. email addresses are finally working, so feel free to email me.

Andrew

Falconer/27: I’m going off what I get from MC79 and Vic Ferrari.  Face-offs are significant, but not as much as who you play with and against.

PPro guys, can you explain what Leslie is saying here:
http://chlnews.blogspot.com/2009_02_22_archive.html#6818895479905673246

Just for the next month, I will be focusing on my work for Puck Prospectus and blogging there (which, due to copyright law, I can’t duplicate here).

Copyright law protects the author of the original work.  Therefore, Leslie owns the copyright, and she can publish her work anywhere she wants.... unless contract law supercedes it.  She would have to explicitly transfer publishing rights to her work to PPro.  (In any case, “fair use” would allow her to cite portions of her own material.)

I know a few years ago, when Nate approached me for BPro, I said “no thanks” when I read the terms for my transferring IP to BPro.  There was a provision that after 6 months, I would be allowed to republish my own work, but that they maintained publishing rights for my work in perpetuity, and that I could not receive those rights back.

Is this the same kind of situation you have with the authors?

And, what is the relationship between PPro and PEV?  Is PPro a subsidiary?  Do they give an equity stake to PEV?  And what is the relationship of the authors to PPro?

Just to clear up a few things. While I do own the copyright to work as soon as it is created, I do not own it when it is a work-for-hire, which is what my articles for Puck Prospectus are. Whether the Unfiltered entry is a work-for-hire is not clear, but the site is copyrighted and the work that is published there cannot just be reproduced in its entirety elsewhere (without permission). Trust me on this, I have been a lawyer for more than 23 years. Fair use is for small portions of an article. You can’t just republish someone else’s article (even under the author’s own name in its entirety) unless you want to risk being sued.

The second thing is, despite what Andrew says above, I was not a Rangers’ scout for 6 years. I have covered the Rangers for six years and have been a scout for an independent agency for half that time, but never for the Rangers. I also have written about prospects for many outlets. But I wanted to make sure that you all knew that the Rangers were never my employer.

Thirdly, teams have for years done psych exams on players and have spent huge amounts of money trying to ascertain a player’s personality make-up and stability. There are psychologists in the interviews, and coaches, teachers, and family members are interviewed. To suggest that Daigle was not thoroughly vetted, either by the Sens or by any of the other teams that he skated for after that is just untrue. Obviously, there are a few folks who are just able to beat the testing and the psych evals, but the teams spend a fortune trying to make sure that this does not happen.

Finally, about plus/minus. I have done a lot of work that indicates that with some adjustments and other considerations added on, it is a very good stat. It just needs to be taken in context. In my opinion, the work in the field should really be devoted to determining what adjustments need to be made in what context. Timo knows that I have strong disagreements with his article, and we went rounds in emails over it. I believe that we should be working outward from the stat we have, not blow it up.

I am glad to find you all here and hope that we can continue to have good discussions on this.

Leslie

One more thing about the work-for-hire issue. I consider the works I have produced for Puck Prospectus a work-for-hire because of an email exchange regarding the articles and payment. I do not want to wind up in court in a test whether that is sufficient to create a work-for-hire or not. I believe it does, and lets leave it at that.

Leslie, thanks for stopping by.  You are correct that as a work-for-hire you relinquish all copyright (similar to an employee of a company not having ownership rights… just ask Stan Lee). 

I had presumed your agreement would have been one of author/publisher relationship. Unless the agreement you have specifically notes that it is a work-for-hire, or that you are explicitly transferring ownership, I don’t see how you are relinquishing copyright.  I would think you are assigning them an exclusive, non-perpetual licence, as I seem to remember an old agreement I saw of theirs.  But, I’m not a lawyer and only play one on the internets.  In any case, we don’t need to elaborate.

You are also quite correct that plus/minus is a worthy stat with adjustments.

An error in Timo’s post today:

http://www.puckprospectus.com/unfiltered/?p=8

“Avery has been among the more efficient “pests” in the NHL at this skill, drawing 8 more penalties than taken in 2007-2008”

That number seem really low.  If you exclude coincidental penalties and majors, Avery took 15 penalties at 5-on-5 and drew 31.  That’s worth about half a win above average.

Just a quick heads up that I have left Puck Prospectus and will be publishing my stats work on another site. (I will come back next week and let you guys know where). I hope that you will critique it in the same manner that you have been looking at the PP work.

As for why I left, we disagreed on how the site was being run and on some (certainly not all) of the work product. When I resigned, I found out that the site was uncomfortable with me too, so we have parted ways and wished each other well.

My intention in telling you this is not to create controversy or to denigrate the site in any way. I hope you all will support PP and that it becomes a leader in the hockey stats field. I just wanted to let you know that I was no longer associated with the site, and that I hope you will look at my work with the same critical eye when its published elsewhere.

I know I appreciate your forthrightness, and only hope that more people were as open as you are. 

You can count on us using our critical eye, wherever you are.