A couple of updates to report. First, I've left-justified the website. Previously StatSheet has been centered on the page. I think the site looks better centered, but that only works well for fix-width sites. Sites that are only ever, say 900 pixels wide, work well being centered. For StatSheet, there are some pages that go beyond 900 pixels and javascript table transformations push the width of the page even more. To make the really wide pages more readable, I've decided to left justify the site. Not a huge deal, but thought I'd share ;-)

A good example of one of the wide pages I mentioned above is also the subject of my next update. Jeff at wsuhoops.wordpress.com requested to have the tempo-neutral stats for the teams in a conference (in Jeff's case the Pac 10) on a single page. I updated the Team Stats section for each conference to show a comprehensive list of stats (including tempo-neutral) for all the teams in the conference. The table columns are sortable and if you hover over a heading, you should see a pop-up describing the stat abbreviation. Thanks for the request Jeff!

As always, keep the suggestions coming.

Posted by Robbie | Permalink | Comments

The Referee pages on StatSheet are some of the most popular. Many fans have a gripe with how the refs call a game at least once during a season. If you are like me, it is more like every other game.

So I always thought that providing ref stats would be interesting. Unfortunately we don't have much to go on. An official boxscore only provides the names of the three refs that called the game. The only stat that shows up in a boxscore you can really attribute to refs are fouls (personal and technical). Many turnovers are subject to a ref's call (did he step on the line or didn't he?), but it is probably a small percentage that are debatable. Fouls, however, are purely on the ref. So that's what I show on the Referee pages. I show on average how many fouls a ref calls, how many days in between games, how many technicals, etc. etc.

There is one problem. The people that fill out the boxscore do a REALLY poor job of spelling the names of the officials. That gives my stats some trouble because the name is all I have to go by. If the person filling out the boxscore spells Curtis Shaw's name Curt Shaw they show up as two different refs on StatSheet.com. But I don't mind the name variations that much. Those aren't really mistakes per se. Michael for Mike, Ricky for Rick, and Joseph for Joe are common variations I've seen.

The much bigger problem are downright misspellings. Here is an example:
Rick Shoulder
Rick Shoulders
Ricky Shouldrs

Here I have to assume there is really only one ref out there named Rick Shoulders, but sloppy boxscore scribes fill it out in different ways.

In order to get halfway clean Ref stats, I have to go through and manually correct these. I'm only half way through and so far have found 300 misspellings! There have been around 3500 completed games so far this season, which means I'm looking at a 8.5% error rate on ref names and I'm only half way through the alphabet! It wouldn't surprise me to find at least 15% of all ref names in boxscores are either misspelled or some variation of the person's name.

In some cases I can't assume that a variation of a name is the same person. I was going to combine Jefferey Smith, Jeffery Smith and Jeffrey Smith all into the same ref named Jeff Smith, but I started doing a little digging and found out that Jeffery Smith called a game in a different part of the country on the same day as a Jeff Smith. There are many situations where I suspect that two names are for the same person, but I can't verify.

Oh, and one last thing. Whoever is filling in ref names using the first initial of the first name followed by the last name (e.g., "A. Salaam")--PLEASE STOP. These are fairly easy to correct, but why be different from everyone else? This just adds to the workload of trying to use a consistent name.

Ok, this rant is over (for now :)

Posted by Robbie | Permalink | Comments

I've touched up the home page a bit so that two stories are highlighted under the "News and Photos" section. Not sure I like it, but we'll see. Some things like that require I stare at it a few times before I can decide if it works or not.

One of the early StatSheet.com users asked for tools to help an annual Mock Selection Committee for the NCAA Tourney that he runs. I developed a Proximity Report several weeks ago to show how far the top 105 teams are from the various opening round arenas. Now I've got a first stab at a top 105 comparison tool.

Posted by Robbie | Permalink | Comments

When I was at MIT, I got my first taste of the business of sports when a MIT alum named Daryl Morey gave a lunch time talk to the Sloanies (I blogged my notes here). At the time Daryl was an SVP with the Boston Celtics. Now he's the GM of the Houston Rockets. Daryl takes a very statistical approach to managing an NBA team, much as Billy Beane has done in Baseball.

Last year Daryl helped organize the first Sports Business Conference at MIT and this year they are doing it again. And since I'm now in the business of sports with StatSheet.com, how could I miss it?

This year the conference is Feb 9th at the Stata Center. I plan on taking copious notes and will post them here. I'm especially interested in the Basketball Analytics session. Stay tuned for updates.

Posted by Robbie | Permalink | Comments

About StatSheet.com

Have a suggestion?
Let me know