Saturday, November 13, 2010

Obscenely Comprehensive NBA Schedule Analysis

(***WARNING!*** FLASH AND RAM INTENSIVE CONTENT BELOW!)

Greetings again Basketbawful readers. I thought that my first take on the 2010-11 NBA season schedule would suffice, but since then apparently more questions on the total effect of scheduling were raised from it. And by that, I mean literally one reader had more in-depth questions for me. So since I couldn't really let it go, and I wanted to see just how good the Suns (or how bad the Lakers) actually had it in terms of NBA scheduling, I went ahead and did the analysis. Introducing AnacondaHL's Obscenely Comprehensive NBA Schedule Analysis.

Updates:
  • *With completely random aesthetic consideration to Google Motion Charts, I have defined the "margin" formulas to Team games minus Opponent games*. This means negative numbers are margins in that team's favor and positive numbers means the team is playing more tired games than facing tired opponents.
  • All raw data spreadsheets are now sorted by ascending margin. Visually I saw enough correlation between margin and number of games against Opponents Only on their SEGABABA, so I felt seeing the margin difference across the league is more meaningful to spot the scheduling injustices.
  • Included win/loss records for all teams in their B2Bs and 4in5s. As far as I can tell, this matches up with (and goes above and beyond) Mark Cuban's data dump back in 2005.
  • The Orlando at New York November 2nd game has been removed.
  • All NBA seasons from 1999-00 to 2010-11 have been analyzed, minus the win/loss records of the current season.
  • No advanced rules were made in consideration of the neutral site games, Lakers/Clippers games, or the Hornets New Orleans/Oklahoma City games.


Raw Data:
I'm giving you the option of clicking on these spreadsheets now to view my raw data, but it may be better to skip to the next section if you hate wading through all the numbers.
1999-04
2004-09
2009-11

Motion Charts:
Inspired by HoopData's use of Google Motion Charts, and understanding that I had three-dimensional data just perfect for the tool, and being quite fond of pretty blobs of gliding color, I put all of my results into a couple of different forms, so if you've got a lot of spare RAM handy follow along:

Lines!

The default of this line chart shows every team since 1999-00 until this year. The y-axis default is Opponents Only B2B and the color is B2B margin, so blue is good, green is average, and red is bad. Overall, the average remains fairly constant over the years, just like the average number of SEGABABAs per team does, but you can see the range varies wildly, unlike the former where there are considerations to limit teams to 23 total SEGABABAs. You can get a better look at a few teams compared to the league average by clicking their checkboxes at the right.

Onto the blobs of color!

Blobs!

Here the default is setup to try and visually see the margin of teams compared to the average. Hit the play button for the blob-tastic motion. You can change the color to unique to make it a little easier on the eyes to track all the teams at once, or select a few teams and follow them. Or change the size of the blobs to Margin, so all the 0's get small and the extremes get big. Again, I will note that I added win/loss records for B2Bs and 4in5s, so you can put that parameter on the chart if you wish.

(If the above charts don't work, or you want a fullscreen view, try opening the raw spreadsheet here).

Conclusion:
By looking at B2Bmargin and overall OppOnlyB2B, I've shown the place where the NBA schedule is still lacking in fairness. Not in the number of SEGABABAs a team plays, but in the number of games played against opponents on their SEGABABA. Over the years in the dataset, teams have had a minimum of 14 and a maximum of 24 SEGABABAs in a year. In contrast, the B2Bmargin ranges from -11 to 12, a significant advantage for generally the Mountain/Central teams.

But in the end does this fatigue really matter? Sure on a league wide average, teams win 8.7 and lose 11.4 SEGABABAs a year, but eyeballing the championship teams, they've typically had schedules with unfavorable margins. OTOH, they've also had schedules with less SEGABABAs played overall, so maybe we've looped back to that original simple analysis: which is as long as a team is talented and rested, they've got the strongest chance at winning the rings. (This shallow analysis does not factor being able to pull off ridiculous star trades, or ability of refs to hand you the championship.)


-AnacondaHL

No comments:

Post a Comment