Play Index Performance Tweak
Posted by Sean Forman on December 13, 2010
Ultimate Grand Slams - Baseball-Reference.com.
We've been having some issues relating to the Play Index performance which has caused some previously working queries to now fail. We've implemented a performance tweak that will has sped things up about 30-45% in my tests. The basic search I always try is running the event finder for every plate appearance for a season. This no longer times out and returns the correct response.
Please keep in mind we are searching massive datasets and there are lots of people doing it at the same time. Sometimes on the event finder you can get to a result in a roundabout way. The link above is all walk-off grand slams with two ours and the home team down 3 runs since 1950.
The easiest way to get to this result is to search a team year for home runs and then select 2-out grand slam home runs. Next change the teams to all teams for a year and search again. Next select, all two outs and game-ending. Next select the years 1950-2010 and then select ones with a margin of 3 runs. With a few intermediate steps you can build up to the query you want.
December 13th, 2010 at 10:21 am
I always get a "No matching events for this player. Criteria may be too restrictive" comment for players (not teams) in the event finder during the non SO outs to positions.
December 13th, 2010 at 10:48 am
[...] posted here: Play Index Performance Tweak » Baseball-Reference Blog » Blog Archive ch_client = "laman6130"; ch_width = 468; ch_height = 120; ch_type = "mpu"; ch_sid = "Chitika [...]
December 13th, 2010 at 2:29 pm
"The link above is all walk-off grand slams with two ours and the home team down 3 runs since 1950"
You shouldn't say "all" because Clemente's in 1956 isn't listed.
December 13th, 2010 at 4:08 pm
Jason Giambi's slam in the rain vs. Minnesota in 2002 is also not on the list. And wasn't the Brooks Conrad homer in Atlanta this year an "ultimate" grand slam?
December 13th, 2010 at 4:11 pm
Nevermind, in both cases there was only one out in the inning.
http://www.baseball-reference.com/boxes/NYA/NYA200205170.shtml
http://www.baseball-reference.com/boxes/ATL/ATL201005200.shtml
December 14th, 2010 at 6:15 pm
Clemente's looks like it was with nobody out, since the Cubs only have 8 IP for the game. (No play by play with the box score.)
December 15th, 2010 at 9:19 am
I'm working on a project to rate players across baseball history. I have been using Play Index and WAR in my research thus far. Can anyone weigh in on WAR and how effective it is to rate players? What are the pros and cons of using WAR? Is it comparable to Win Shares?
I appreciate any feedback.
December 15th, 2010 at 10:50 am
Dan, those questions are far too general and complicated to effectively answer in a single post here. The theories behind WAR are based on research going back a long way. I would suggest you do some reading. Here are a few places you could start (not necessarily the best primers, but what I found on with a brief search):
http://www.insidethebook.com/ee/index.php/site/comments/baseball_prospectus_warp1_is_wrong/
http://www.baseballthinkfactory.org/files/hall_of_merit/discussion/chones_warp_and_the_hall_of_merit/
http://www.insidethebook.com/ee/index.php/site/comments/rallys_historical_war/
Fangraphs WAR (different than the one used on B-R): http://www.fangraphs.com/blogs/index.php/glossary/#winvalues
If you have more specific questions, I think people here will certainly try to answer them.
December 15th, 2010 at 10:59 am
Dan, those questions are far too general and complicated to effectively answer in a single post here. The theories behind WAR are based on research going back a long way. I would suggest you do some reading. Here are a few places you could start (not necessarily the best primers, but what I found on with a brief search). If you have more specific questions, I think people here will certainly try to answer them. (I am posting one link per post so this doesn't get flagged as spam)
http://www.insidethebook.com/ee/index.php/site/comments/baseball_prospectus_warp1_is_wrong/
December 15th, 2010 at 11:00 am
http://www.baseballthinkfactory.org/files/hall_of_merit/discussion/chones_warp_and_the_hall_of_merit/
December 15th, 2010 at 11:00 am
http://www.insidethebook.com/ee/index.php/site/comments/rallys_historical_war/
December 15th, 2010 at 11:01 am
Fangraphs WAR (different than the one used on B-R): http://www.fangraphs.com/blogs/index.php/glossary/#winvalues
December 15th, 2010 at 11:01 am
Fangraphs WAR (different than the one used here on B-R): http://www.fangraphs.com/blogs/index.php/glossary/#winvalues
December 15th, 2010 at 11:02 am
test?
December 15th, 2010 at 11:03 am
Link to Fangraphs WAR (different than the one used on B-R): http://www.fangraphs.com/blogs/index.php/glossary/#winvalues
December 15th, 2010 at 11:05 am
OK, I don't understand why this site is giving me so many problems in posting. The last link is for Fangraphs WAR (different than the one used on B-R). Remove the spaces in this link: fangraphs.com/ blogs/ index.php/ glossary/#winvalues
December 15th, 2010 at 4:31 pm
Haha, now I look insane. Please note that not all of the above posts went through originally, so now that they've been restored and left alongside their follow-ups, it appears that I was on a posting bender.
December 17th, 2010 at 1:22 pm
Posts with links can be flagged as spam, so we need to then go in and approve them as they come through.