Accuracy of NBA Mock Drafts in 2012, Best, Worst

NBA Draft 2012 is in the books so it’s time for a lot of grading… no, I don’t mean judging teams or drafted players because it’s too early but there’s one draft-related topic which doesn’t require waiting few years:
Who Had The Best and Worst 2012 Mock Drafts on the Internet?

After humble beginnings last season I expanded my efforts to 122 NBA Mock Drafts this year.
Obviously, it doesn’t mean I covered every mock available on the web but I think I was close ;-)

Seriously speaking though here’s what I did:
1) after the draft lottery I saved on my hard drive pages from as many 1st round mocks as I could quickly gather. Mostly from widely known authors/sources but I also added many googled ones.
I repeated this exercise weekly up to the day before actual NBA Draft when I had a final push for the data.

There are two important notes here:
a) because I tried to capture content which was quickly changing some of the links provided won’t work. The most extreme example was probably with Chad Ford from ESPN who published 10 mock drafts and the last 3 or 4 had multiple “updates” which by the way block the entry to original versions.
I could just upload those saved files but that probably wouldn’t be legal so I just posted a PDF file with every pick of every mock I’ve processed.
b) I included 3 mocks from the day of the draft from [my guess at] 3 most popular sources but I don’t feel great about it. On a draft day rumors from teams, possible leaks and updates are so rampant that I question the value of those mocks for 10 minutes… especially when at this point you can just wait for the real event.

2) I graded every mock by an average absolute difference between the mocked and actual picks
For example Bradley Beal was drafted third so if you had him at #2 you would get a score of “1″ but if you had him at #5 you would get a score of “2″. Got it? If a player went undrafted I counted that as a “30″ because of the entire second round. I did the aforementioned thing for every pick in every mock and calculate an average for the entire first round so the lower the number the more accurate mock was.

OK, let the mock draft challenge begin! Here are the results…

 

Dunk Numbers for all Players in the NBA in 2007-08

Slam dunks are one of the most attention-grabbing plays in basketball, as proved recently by a meeting between Blake Griffin and Kendrick Perkins, and they become a hot topic quickly but there is a surprising scarcity of cold facts about them. CBS Sports has a nice leader-board… but they don’t include attempts and links to previous years don’t work anymore. Regarding dunks in the last 6 regular seasons in the NBA you can find this breakdown for 2010-11 by EvanZ but it includes only 30 top players and some non-dunkers, this copy of leader-board for 2009-10 and… flashback to days of activity on 82games.com which doesn’t even cover entire 2005-06 season. That’s just wrong, isn’t it? So I’d like to change that sad state.

UPDATE: Just one week after this post (!) invaluable basketball-reference.com introduced Play Index+ which allows us to find those numbers quickly. So the good news is fantastic new data became available! The bad news is it made this post somewhat obsolete ;-)

I started this project as usual with ESPN’s play-by-play data… but after I’ve finished and compared to aforementioned sources I discovered a shocking fact: ESPN’s version has a nasty habit of changing dunks into 1-foot shots! And it’s not a small issue either – we are talking about difference of around 3000 plays per year! So those files landed in trash along mine time spent on it and I had to start from scratch.

For my second attempt I used play-by-play data from basketballvalue.com… and even though their numbers looked fine it wasn’t a smooth ride either because of another nasty habit – they identify players only by surnames… but in some games they added a first letter to make a distinction between players on different teams. Such situation creates a litany of possible identification conflicts, for example just note how many players in the NBA are named Williams or Johnson, but through battery of tests and corrections by hand I think I sorted it all out successfully.
With that said keep in mind that there could be a couple of mistakes here, especially with players who played little or those who had the same name on the same team, but overall I think it should be accurate.

So here is a full version of dunks leader-board in the NBA in 2007-08
[it's sorted by made dunks then by attempts and then by minutes played so you can also easily find at the bottom players who didn't dunk. All non-dunk data is from basketball-reference.com]
You can also check other seasons: 2005-06, 2006-07, 2008-09, 2009-10 and 2010-11.

 

NBA Dunks Data 2006-07 Best, Worst, Attempts

Slam dunks are one of the most attention-grabbing plays in basketball, as proved recently by a meeting between Blake Griffin and Kendrick Perkins, and they become a hot topic quickly but there is a surprising scarcity of cold facts about them. CBS Sports has a nice leader-board… but they don’t include attempts and links to previous years don’t work anymore. Regarding dunks in the last 6 regular seasons in the NBA you can find this breakdown for 2010-11 by EvanZ but it includes only 30 top players and some non-dunkers, this copy of leader-board for 2009-10 and… flashback to days of activity on 82games.com which doesn’t even cover entire 2005-06 season. That’s just wrong, isn’t it? So I’d like to change that sad state.

UPDATE: Just one week after this post (!) invaluable basketball-reference.com introduced Play Index+ which allows us to find those numbers quickly. So the good news is fantastic new data became available! The bad news is it made this post somewhat obsolete ;-)

I started this project as usual with ESPN’s play-by-play data… but after I’ve finished and compared to aforementioned sources I discovered a shocking fact: ESPN’s version has a nasty habit of changing dunks into 1-foot shots! And it’s not a small issue either – we are talking about difference of around 3000 plays per year! So those files landed in trash along mine time spent on it and I had to start from scratch.

For my second attempt I used play-by-play data from basketballvalue.com… and even though their numbers looked fine it wasn’t a smooth ride either because of another nasty habit – they identify players only by surnames… but in some games they added a first letter to make a distinction between players on different teams. Such situation creates a litany of possible identification conflicts, for example just note how many players in the NBA are named Williams or Johnson, but through battery of tests and corrections by hand I think I sorted it all out successfully.
With that said keep in mind that there could be a couple of mistakes here, especially with players who played little or those who had the same name on the same team, but overall I think it should be accurate.

So here is a full version of dunks leader-board in the NBA in 2006-07
[it's sorted by made dunks then by attempts and then by minutes played so you can also easily find at the bottom players who didn't dunk. All non-dunk data is from basketball-reference.com]
You can also check other seasons: 2005-06, 2007-08, 2008-09, 2009-10 and 2010-11.

 

The Best Game Stat Lines in the Last 20 NBA Seasons

After I wrote about worst single game stat lines out of curiosity and almost morally I had to check the other side: The Best Single Performance Box Score Stats in the Last 20 NBA Seasons.

Once again I used my old tool for searching through player’s game logs [over 8100 of them] from
Yahoo [Last 5 seasons] and basketballreference.com [previous 15 seasons] and two measures:
Game Score where the formula is PTS + 0.4 * FG – 0.7 * FGA – 0.4*(FTA – FT) + 0.7 * ORB + 0.3 * DRB + STL + 0.7 * AST + 0.7 * BLK – 0.4 * PF – TOV and Win Score which is calculated by the following equation: PTS + REB + STL + 1/2*BLK + 1/2*AST – FGA – 1/2*FTA – TO – 1/2*PF.

So who had The Best Stat Line in the Last 20 Regular Seasons?
Here’s a Top 25 for Single Game Stats according to Game Score:

 

Watching The Watchers – The Best NBA Mock Drafts

NBA Draft 2011 didn’t disappoint in terms of providing entertainment with interesting decisions, picks and trades but it’s over so it’s time for some conclusions… no, I don’t mean judging teams or drafted players with special grading scales [where almost all teams were above average!] but to answer a question which doesn’t require waiting few years: Who Had The Best and Worst 2011 Mock Drafts on the Internet?
(You can also check expanded version for 2012 NBA Mock Drafts!

Obviously, I couldn’t cover every mock available on the web but here’s what I did:
1) 12 hours before NBA Draft I saved on my hard drive pages from as many latest 1st round mocks as possible. Mostly from widely known authors/sources but I also added some random googled ones.
For more fun and information I also added older versions.
2) I graded every mock by an average absolute difference between the mocked and actual picks.
For example Derrick Williams was drafted at #2 so if you had him at #1 you would get a score of “1″ but if you had him at #5 you would get a score of “3″. Got it? If a player went undrafted I counted that as a “30″ because of the entire second round. I did the aforementioned thing for every pick in every mock and calculate an average for the entire first round so the lower the number the more accurate mock was.

BTW, if you are interested what’s the score for YOUR mock draft let me know in the comments.
Although there better be some kind of valid information about a date.
It won’t be fun to judge mock which has been written after real draft ;-)

OK, let the mock draft challenge begin! Here are the results… (Number before name means “earlier version”):

 

Effects of Draft Lottery Results on NBA History Part2

NBA Draft is days away so it’s time for some topics about it…

In part 1 I focused on a difference between players who were drafted at any given spot and those who would have been drafted there had lottery played out exactly as odds suggested. In part 2 I’ll take into consideration those odds. Incidentally I’ve stumbled into a solid way of quantifying how well teams have drafted!

But first, here’s a remainder of rules which were used:
- I assumed continuity of teams so “OKC Thunder” means “Seattle SuperSonics” in earlier years,
- I included new ownership of picks via trades only when they were completed before making a pick,
- I assumed that Career Win Shares in regular season adequately describe player’s career,
- I assumed that players would be drafted at the same slot no matter which team have owned given pick.

What’s more, I changed perspective to the Last 10 Years because IMHO drafts from 1995 to 2000 had too big impact on numbers because careers of those players have lasted way longer than any other. For every season in last decade I re-created probabilities of getting every pick before draft [this is an example from 2010]: