If Elton Brand bounces back...If Evan Turner can turn into a legit number one...If Jrue Holiday makes a leap...If Spencer Hawes suddenly isn't a complete waste of space...If Andres Nocioni returns to his form of six seasons ago...If Doug Collins can get the front court to suddenly change their stripes and learn to defend. Pretty much every optimistic outlook on this team for 2010-2011 starts with one of those ifs.Throughout the next couple of posts, we're going to take a look at what to expect if a miracle doesn't happen.
Joe, a regular reader and frequent contributor in the comments, got the ball rolling for this series and the research behind it. It's a fresh way to look at things, and I'd like to give Joe a big thanks for the contribution. All math in the charts to follow is strictly my work, however, so if there's an error, don't yell at Joe.
The overall goal here is to use advanced statistics to set a baseline for how many games the Sixers should win this season. I'm going to use three posts to tell the story, mainly because I have a lot to say, and probably a lot of convincing to do as well.
The first step in the process was to choose the metrics. Two stats fit the mold for me. Both aim to put a ton of available "basic" stats into a big pot and reduce until you're left with a manageable number, wins. Here's a basic primer on the methodology behind Wins Produced and Win Shares. Each of these stats values different parts of the game, but the end result is the same. How many wins did a player contribute to his team's total.
The first step is to give each stat a basic smell test. If each claims to have a formula which condenses basic stats into wins, then each team's total should equal the sum of its parts. The chart below is a look at the 2009-2010 season, broken down by team. We're looking at actual wins, the sum of WP and WS for all the players on the team (only when they actually played for the team), then the differences. If a number is in red, the team's Win Shares or Wins Produced were lower than their actual wins. If it's blue, the opposite occurred.
As you can see, league-wide WP was under by only an average of .04 wins, while Win Shares was within 1.32 games, on the over-estimating side. The Sixers were overestimated by both metrics, which could have something to do with horrible coaching, or any other number of factors. It probably just means they weren't as bad as their record, which I think we already knew. Overall, I'd say both stats are somewhat reliable, worst-case was 7 games off, and 2/3 of teams were +/- 3 wins in either direction, using both metrics, which passes the smell test for me. Obviously, we're only looking at one season here, but I just wanted to get a baseline. There is a variance, and we'll definitely keep that in mind as we continue down this road.
Now that we've got the basic introduction for each stat down, as well as a quick look at their performance, it's time to use derivatives to see how predictive each stat has been, specifically, how predictive each has been for the Sixers.
Let me know if you're with me so far in the comments. Part two is coming either later today or early tomorrow.