Baseball for the Thinking Fan

Login | Register | Feedback

btf_logo
You are here > Home > Baseball Newsstand > Baseball Primer Newsblog > Discussion
Baseball Primer Newsblog
— The Best News Links from the Baseball Newsstand

Thursday, January 09, 2014

Baseball Prospectus: How Secret Ballots Skewed the Hall of Fame Election Results

The writer sees this as a case in favor of transparency…. interesting info either way.

Extrapolating the results for anonymous voters from the full results and the Ballot Collecting Gizmo, here’s how the two groups voted on the serious candidates:

 

The District Attorney Posted: January 09, 2014 at 07:58 PM | 17 comment(s) Login to Bookmark
  Tags: baseball prospectus, bbwaa, hall of fame, media

Reader Comments and Retorts

Go to end of page

Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.

   1. Ray K Posted: January 10, 2014 at 12:42 AM (#4635420)
Article should be titled, "How a Biased Sampling of Ballots Skewed Our Expectations of the Election Results"

I mean, c'mon. This is Statistics 101.
   2. Buck Coats Posted: January 10, 2014 at 03:21 AM (#4635454)
Surprising to see Larry Walker in the "got a boost" group with terrible picks Lee Smith and Don Mattingly.

Also surprising to see McGwire with a small gain when Clemens and Bonds had two of the largest drops between groups. But maybe the 10-vote limit had something to do with that
   3. Fancy Pants Handles lap changes with class Posted: January 10, 2014 at 04:24 AM (#4635463)
Article should be titled, "How a Biased Sampling of Ballots Skewed Our Expectations of the Election Results"

I mean, c'mon. This is Statistics 101.

For next year, we should contact the unskewed polls guy, and see if he is interested in a job...
   4. Cooper Nielson Posted: January 10, 2014 at 05:39 AM (#4635468)
I like the key:

* — significant at 10% level
* — significant at 5% level
* — significant at 1% level


Very helpful!
   5. jmurph Posted: January 10, 2014 at 10:25 AM (#4635547)
I have to say, I was all ready to get my pitchfork out but instead people keep writing over the top stuff like this that leaves me with the conclusion that the BBWAA is actually doing a fine job.

EDITed because I can't right good.
   6. Austin Kearns: The Spy Who Shagged Flies Posted: January 10, 2014 at 11:16 AM (#4635603)
There are a lot of voters in both categories, so relatively small differences will be statistically significant. So the most surprising thing to me isn't who did have a statistically significant difference in voting rates, but who didn't - especially Jeff Bagwell, Alan Trammell and (of course) Morris.
   7. villageidiom Posted: January 10, 2014 at 11:33 AM (#4635628)
I mean, c'mon. This is Statistics 101.
Consider the source.
   8. Jose Can Still Seabiscuit Posted: January 10, 2014 at 11:41 AM (#4635644)
One of the curious things is that the people with public votes are more likely to vote for the steroid users than those not publicizing their votes. If PED use is such a terrible thing shouldn't it be the other way around?
   9. Ray K Posted: January 10, 2014 at 02:55 PM (#4635855)
Consider the source.


Is Baseball Prospectus considered a poor source for analysis?
   10. Ray K Posted: January 10, 2014 at 02:56 PM (#4635857)
One of the curious things is that the people with public votes are more likely to vote for the steroid users than those not publicizing their votes. If PED use is such a terrible thing shouldn't it be the other way around?


Could simply be that PED supporters feel the need to crusade.
   11. Davo Dozier Posted: January 10, 2014 at 02:59 PM (#4635860)
#9--Baseball Prospectus lost all their best talent to the Astros.
   12. Danny Posted: January 10, 2014 at 03:18 PM (#4635878)
One of the curious things is that the people with public votes are more likely to vote for the steroid users than those not publicizing their votes. If PED use is such a terrible thing shouldn't it be the other way around?

I'd guess that's just a symptom of (i) the private ballots had used fewer slots on average, and (ii) the old vs. young divide.
   13. Ray K Posted: January 10, 2014 at 03:57 PM (#4635905)
#9--Baseball Prospectus lost all their best talent to the Astros.


Awesome! I look forward to the team's turnaround.
   14. Ben V-L Posted: January 10, 2014 at 05:48 PM (#4635985)
I like the key:

* — significant at 10% level
* — significant at 5% level
* — significant at 1% level

Very helpful!


It can be twisted into making sense. If the result is significant at a 10% level, it gets only the 10% star. If it's significant at the 5% level, it is necessarily also significant at the 10% level, so it gets both the 5% star and the 10% star. Etc.
   15. Athletic Supporter can feel the slow rot Posted: January 10, 2014 at 07:19 PM (#4636029)
I would have titled this piece "How Public Ballots Skewed the Hall of Fame Election Results" and otherwise left it unchanged. (Well, except for obvious selection biases on public/private.)
   16. Walt Davis Posted: January 11, 2014 at 01:39 AM (#4636132)
Surprising to see Larry Walker in the "got a boost" group with terrible picks Lee Smith and Don Mattingly.

The Gizmo has been low on Walker every year. Maybe the non-Gizmo voters are more impressed by MVPs or raw rate stats (can't be raw counting stats) or maybe they assume all Canadians are clean.

The problem with the article is that he couldn't wait a couple of days to include the ballots posted on the BBWAA website. His differences aren't public vs. non-public, they're Gizmo (published before the announcement) and non-Gizmo.

Ray, your point is rather meaningless. The size of the "bias" is exactly what he's trying to estimate which is a good thing to do. I stopped reading once I saw he didn't have the latest data but nowhere in the first few paragraphs does he suggest that this result is surprising, that the Gizmo sample should provide identical results, etc.

Just because samples come from two different sources or even if your sample is a convenience sample does not, in itself, make the sample biased. The bias derives from different outcomes not different characteristics. Usually you can't test whether or not the sample outcome differs from the population outcome -- i.e. in general, there's no point collecting the sample data if you already have the population outcome. Here you can and good for him for doing so. I just wish he'd waited to have the BBWAA data.

   17. SoSH U at work Posted: January 11, 2014 at 02:08 AM (#4636139)
I think the data to use would be Ryan's Weird-Meat site. Not only does it add in the public ballots revealed after the vote, it also separates the ballots that are known but public and those known but anonymous (a percentage of ones run through the Gizmo are given to Repoz under a cloak of darkness, probably in return for bizarre favors best not to think about).

By simply using the Gizmo, he's mixing some essentially anonymous votes in with the truly public votes.

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
Chicago Joe
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Hot Topics

NewsblogGraphic designer creates Star Wars/MLB mash-up logos for all 30 teams
(15 - 4:14am, Aug 02)
Last: Barnaby Jones

NewsblogMike Carp designated for assignment, Mookie Betts called up - Over the Monster
(31 - 4:00am, Aug 02)
Last: Walt Davis

NewsblogOT: Politics, August 2014: DNC criticizes Christie’s economic record with baseball video
(218 - 3:08am, Aug 02)
Last: David Nieporent (now, with children)

NewsblogOT:  2014 College Football pre-season thread
(10 - 2:46am, Aug 02)
Last: Every Inge Counts

NewsblogCubs deal Emilio Bonifacio, James Russell to Braves
(47 - 2:38am, Aug 02)
Last: Mess with the Meat, you get the Wad!

NewsblogOT: Monthly NBA Thread - August 2014
(43 - 2:22am, Aug 02)
Last: Norcan

NewsblogJim Bowden Caught Stealing From Fake Twitter Account, Deletes Everything
(35 - 1:33am, Aug 02)
Last: Jose Can Still Seabiscuit

NewsblogMiklasz: Mozeliak sets off a loud wake-up call
(14 - 1:13am, Aug 02)
Last: Walt Davis

NewsblogGiants Designate Dan Uggla, Tyler Colvin
(13 - 12:26am, Aug 02)
Last: Infinite Joost (Voxter)

NewsblogOMNICHATTER 8-1
(29 - 12:13am, Aug 02)
Last: Los Angeles El Hombre of Anaheim

NewsblogRosendo ‘Rusty’ Torres found guilty of 5 counts of sex abuse, acquitted of 3 other charges
(15 - 12:11am, Aug 02)
Last: Ray (RDP)

NewsblogOT: Monthly NBA Thread- July 2014
(1081 - 10:39pm, Aug 01)
Last: Quaker

NewsblogGeorge "The Animal" Steele Mangles A Baseball
(160 - 9:12pm, Aug 01)
Last: NJ in DC (Now unemployed!)

NewsblogBill James Mailbag - 7/13/14 - 7/22/14 (Subscription Required)
(11 - 8:40pm, Aug 01)
Last: robinred

NewsblogJose Bautista, Casey Janssen “frustrated” and “disappointed” by Jays’ lack of deadline activity
(6 - 8:31pm, Aug 01)
Last: Random Transaction Generator

Page rendered in 0.2775 seconds
52 querie(s) executed