PDA

View Full Version : Great Analysis of 2010 Duke's Path to the Title



tbyers11
04-07-2010, 05:13 PM
I don't think this has been posted. Neil Paine at the relatively new College Basketball at Sports-Reference (http://www.sports-reference.com/cbb/) site takes a statistical look at Duke's path to the title (http://www.sports-reference.com/cbb/blog/?p=115). It is pretty long and uses some real "dork poll analysis", but I think as long as you have the info that (SRS) stands for Standard Rating System and is a basic metric that uses point differential and schedule strength to rank teams you should be set.

He looks at the last 30 national champs and compares things in a couple of different ways, but the take home message is that Duke 2010's path was neither easy nor hard but just about in the middle of the pack (roughly 13th). An interesting side note is that no matter which of the three ways that he looks at it 2009 UNC's path was easier than 2010 Duke. :)

I know that rational things like "math" don't combat most of the irrational Duke hate out there, but I like numbers and found a quantitative look at Duke's path to the 2010 title quite interesting.

dball
04-07-2010, 06:53 PM
Interesting article and analysis by Ga Tech grad (who admits he dislikes UNC and Duke). Looking at figures in general for last 30 years, 2006 and 2007 FL teams had easiest paths but 82 UNC shows up at 5 and 09 UNC at 8. 2010 Duke is 13, pretty close to middle.

In another scenario, 2010 makes the top 10 but is preceded again by 2006 FL at number one and three! UNC squads (2005, 1993, 2009).

In all of the number crunches, 2009 UNC always shows as easier path to title than 2010 Duke.

MChambers
04-07-2010, 08:19 PM
I don't think this has been posted. Neil Paine at the relatively new College Basketball at Sports-Reference (http://www.sports-reference.com/cbb/) site takes a statistical look at Duke's path to the title (http://www.sports-reference.com/cbb/blog/?p=115). It is pretty long and uses some real "dork poll analysis", but I think as long as you have the info that (SRS) stands for Standard Rating System and is a basic metric that uses point differential and schedule strength to rank teams you should be set.

He looks at the last 30 national champs and compares things in a couple of different ways, but the take home message is that Duke 2010's path was neither easy nor hard but just about in the middle of the pack (roughly 13th). An interesting side note is that no matter which of the three ways that he looks at it 2009 UNC's path was easier than 2010 Duke. :)

I know that rational things like "math" don't combat most of the irrational Duke hate out there, but I like numbers and found a quantitative look at Duke's path to the 2010 title quite interesting.

It's an interesting article. But I'm deeply suspicious of any article that doesn't rate 1992 Duke among the best of all time. Ranked #1 every day.

David
04-07-2010, 08:43 PM
It's an interesting article. But I'm deeply suspicious of any article that doesn't rate 1992 Duke among the best of all time. Ranked #1 every day.

As I understand the SRS, you shouldn't use it to directly compare the quality of the 1992 Duke team relative to other champion teams. That is, it doesn't tell us who would win if the 1992 Duke team played the 2010 Duke team. Importantly, as tbyers11 noted above, it accounts for point differential and strength of schedule within seasons but NOT changes across seasons in the relative strength of college basketball.

You can use it to compare 1992 Duke team versus its 1992 opponents relative to 2010 Duke versus its 2010 opponents. For example, the author of the blog uses the SRS to note that 2010 Duke team looks very good relative to 2010 teams when compared to previous champions (see his first table). This analysis suggests the 2010 Duke team is a worthy champion. In subsequent tables, the blog author uses the SRS to compare the quality of our tourney opponents relative to the previous 29 champions. Once again, Duke looks somewhat middle-of-the-road here (and better than the 2009 heels).

Kfanarmy
04-08-2010, 01:15 AM
... it accounts for point differential and strength of schedule within seasons but NOT changes across seasons in the relative strength of college basketball.

I actually don't think you can get anywhere near determining "relative strength of college basketball" on the whole between specific years. There are simply too many variables and no matter what conclusion you draw it cannot be tested...though it might be fun to argue about.

MChambers
04-08-2010, 08:12 AM
As I understand the SRS, you shouldn't use it to directly compare the quality of the 1992 Duke team relative to other champion teams. That is, it doesn't tell us who would win if the 1992 Duke team played the 2010 Duke team. Importantly, as tbyers11 noted above, it accounts for point differential and strength of schedule within seasons but NOT changes across seasons in the relative strength of college basketball.

You can use it to compare 1992 Duke team versus its 1992 opponents relative to 2010 Duke versus its 2010 opponents. For example, the author of the blog uses the SRS to note that 2010 Duke team looks very good relative to 2010 teams when compared to previous champions (see his first table). This analysis suggests the 2010 Duke team is a worthy champion. In subsequent tables, the blog author uses the SRS to compare the quality of our tourney opponents relative to the previous 29 champions. Once again, Duke looks somewhat middle-of-the-road here (and better than the 2009 heels).
But I really think the 1992 team, relative to the rest of college basketball that year, was much more dominant than the 2010 team relative to the rest of college basketball in 2010. (I absolutely agree that college basketball was stronger in 1992, but that was not the point I was making.)

The 1992 team was #1 all season long, lost only two games, both in the final seconds (while Hurley was hurt, but the computers don't see that). It won the NCAA finals by 20 points.

1992 team had Grant Hill, Laettner, Hurley, and Parks, all of whom played quite a bit in the NBA. Davis, Lang, and Thomas Hill had cups of coffee.

I respect statistical analyses, but I guess on this one I'll go with the eye test.

Underdog5
04-08-2010, 08:58 AM
But I really think the 1992 team, relative to the rest of college basketball that year, was much more dominant than the 2010 team relative to the rest of college basketball in 2010. (I absolutely agree that college basketball was stronger in 1992, but that was not the point I was making.)

The 1992 team was #1 all season long, lost only two games, both in the final seconds (while Hurley was hurt, but the computers don't see that). It won the NCAA finals by 20 points.

1992 team had Grant Hill, Laettner, Hurley, and Parks, all of whom played quite a bit in the NBA. Davis, Lang, and Thomas Hill had cups of coffee.

I respect statistical analyses, but I guess on this one I'll go with the eye test.

In addition to being #1 from preaseason ranking to final buzzer, they won ACC regular season and tourny. And the talent level that year was exceptionally high. If my memory serves me correctly, there were only about 3 teams Duke played that did not have at least 1 NBA prospect (Russian national team in exhibition, Harvard and Campbell in the 1st round of the tourny). They won away games at Michigan (with the Fab 5), UCLA and LSU (with Shaq).

dukeblue225
04-08-2010, 11:17 AM
could somebody please help describe and define what SRS is. I can't find a clear and concise definition.

superdave
04-09-2010, 08:58 AM
Even the 1999, 2001, 2002 Duke teams would have gotten waxed by that 1992 team. The only teams that could have hung with them in my view would be the following:
1982 UNC
1984 Georgetown
1990 UNLV
1996 Kentucky
2001 Duke

1992 Duke was simply the best college team I've ever seen and I dont think any formula can either confirm or deny that because of the changes in the game, the talent levels, the conferences etc.