BR.com's Sports Reference Blog

Most Games Played with no World Series Appearance - Baseball-Reference.com

Rafael Palmeiro leads this new list on the site.

Recently Most Popular Pages - Baseball-Reference.com

I took a few minutes today to add a feature that shows the players getting the most interest on baseball-reference.com. It is a collection of the most popular pages for yesterday and then for the last week (pre-seeded with 7*yesterday’s traffic, but that will start updating tomorrow). Not surprisingly, Niekro and Eckstein start out numbers one and two.

ESPN.com - NFL - Chargers’ Merriman apologizes, will appeal suspension

Merriman was the NFL’s defensive rookie of the year and a starter in the Pro Bowl last year as a rookie. Where is the media frenzy? ESPN even goes so far as to print his lawyer’s comments about how this is Congress’s fault. Yeah right. Same old, same old. The NFL gets a free pass and baseball gets slammed.

I was heartened to see Joe Buck asked in the SBJ a few weeks back about the double standard with regards to PED use in the two leagues.

ACM TOIT

I have to say that I find it a little bit amusing that the ACM’s Transactions on Internet Technology can’t be ordered over the internet.

Our condolences go out the families of those killed in the crash of Lidle’s plane and also to the families of Buck O’Neil, Ivan Murrell and Eddie Pellagrini who also passed away this week.

I hope this doesn’t come across as unseemly, but I thought some fellow webmasters might be interested in how the traffic on the site changed the day of Cory Lidle’s death and some of the steps I took to handle it.

The site gets a lot of traffic, Alexa has us as the 9,500-11,500th most popular site on the internet, and we currently use two servers to handle all of the traffic and may be looking at a third pretty soon, so this isn’t a small operation. Even so, the traffic on the 11th came very close to locking up and crashing our servers.

Here are the number of times each hour on the 11th and 12th that Lidle’s page was accessed (BR is the second link on Google which explains a lot of the traffic finding the page).

Oct 11, 12-13:          4
Oct 11, 13-14:          4
Oct 11, 14-15:         12
Oct 11, 15-16:          2
Oct 11, 16-17:       6619
Oct 11, 17-18:      21223
Oct 11, 18-19:       7582
Oct 11, 19-20:       4199
Oct 11, 20-21:       3225
Oct 11, 21-22:       2764
Oct 11, 22-23:       2262
Oct 11, 23-24:       1662
Oct 12, 00-01:       1166
Oct 12, 01-02:        865
Oct 12, 02-03:        843
Oct 12, 03-04:        734
Oct 12, 04-05:        620
Oct 12, 05-06:        574
Oct 12, 06-07:        611
Oct 12, 07-08:        745
Oct 12, 08-09:       1138  <---- People arriving at work.
Oct 12, 09-10:       1248
Oct 12, 10-11:       1205
Oct 12, 11-12:       1046
Oct 12, 12-13:        985
Oct 12, 13-14:        873
Oct 12, 14-15:        886
Oct 12, 15-16:        756
Oct 12, 16-17:        702
Oct 12, 17-18:        502
Oct 12, 18-19:        473
Oct 12, 19-20:        415
Oct 12, 20-21:        336
Oct 12, 21-22:        335
Oct 12, 22-23:        297
Oct 12, 23-24:        230

In the hour after the story broke, the site was getting 60 hits/second (files served). Note that in web parlance, hits includes thinks like images, javascript files, stylesheets and the like that have to be served separately. Every access of a player page on BR causes six or so hits (though the browsers cache files and don’t re-download files unnecessarily for the second page you visit), and the site was getting very slow. Running scripts like the player search also slows things down considerably.

I took out all of the extra files from the Lidle page, no ssi’s, no images, no external javascript, no external stylesheet, therefore each page access was now just one hit. I also modified the front page to direct people directly to the Lidle page, so they wouldn’t use the search engine unnecessarily. That caused a dramatic improvement in the performance and the site was able to maintain its service to the 100,000+ that visited the site that day.

I’ve updated the postseason sections through the first round of the playoffs, so you can refer to the site if you wish to see how the Tigers beat up on the Yankees ( :-), couldn’t resist).

I also have play-by-play for the 2006 season now, so I’ve rolled out some of those features (like debuts game results for the new players).

2006 Oakland Athletics Lineups and Defense - Baseball-Reference.com

We got the 2006 gamelogs done and updated, so now you can check out the 2006 lineups and batting orders for teams. Also team schedules and splits are updated and on the site.

Players Debuting in 2006 - Baseball-Reference.com

The 2006 update is now live on the site. I’m regenerating the search engine, so it will be a few minutes before the new players are in the search engine.

Puzzles from BallPuzzle.com and Baseball-Reference.com

I’m happy to announce a partnership with Clayton DuVall of BallPuzzle.com. Clayton is allowing us to run some of his baseball crossword puzzles on the site. If you are the type of person who enjoys puzzles, this will be right up your alley. Unfortunately, the puzzle applet doesn’t appear to work with linux as it causes my browser to cross, but it appears to work all right with both IE and FireFox in Windows and the Mac. Official launch is in a day or two, so let me know if you come upon any issues.

Barring some unforeseen circumstances, the site should be updated tomorrow morning. Doing this full-time makes updating it much, much easier. I would have been done today, but I had to watch Carl for a few hours this morning. Nonetheless, it will be nice to have the update done.

Deep in the Count

The New Republic now has a post-season baseball blog. It will be interesting to see how slightly left-of-center (yes, I’m painting with a very broad brush) politics plays out in baseball analysis.