Mock Drafts Database - Interesting conclusions

I was thinking this year's draft is so special and exciting so I collected the listings of 14 NBA mock drafts and put them all on a filterable, sortable, searchable table. Also I run some math to draw some conclusions. For example most will say Wiggins is ahead of Jabari Parker but their average rankings are almost identical, to be precise 2.28 vs 2.35.

Visit the site here: [URL="http://nbadraft.dahoops.com/"]http://nbadraft.dahoops.com/[/URL]

What do you see? Any kind of feedback will be great, as I put a lot of effort into this.

Great work man! I like it.. A lot better than I expected.. Are you a web designer? I'm actually in the process of looking for web designers/programmers.

Wow this is nice. BTW your twitter button is barely visible on that background. Easy fix.

Very Cool!

How did you define the variance column? I find it interesting that as the draft goes deeper, the variance gets worse, which makes sense, but then there are a couple guys tucked away up there with low variance. Tyler Ennis and PJ Hairston (Which you type-o'd as RJ Hairston by the way). Maybe they just hit a specific need for someone or a couple someone's in that range. Maybe the mark the end of a tier as generally definied by management, so they'd be hard pressed to slip much further in spite of need?

I don't know, but it's interesting.

Nice work. I posted it to FB.

Hey Thorpesaurous, thanks for the feedback.

Variance column is heavily about the standart deviation of each player's respective ranks on different mocks. This number is somewhat proportional to the average rank, i.e. a std-dev of 3 is huge for a top-10 player but not much so for a late first rounder so I adjusted those accordingly and come up with a reader friendly 'low/high variance' labelling.

If you ask me variance here is about two main things: 1. Player is a mystery of some kind, so everyone has their opinion given the little information available. The only player who does not follow this pattern is Dante Exum. He is between 4-7 among all 14 boards, which I find surprising. 2. Sometimes a player is well-known but he is so controversial by nature. Kyle Anderson is such a freak combining a unique set of attributes.

You asked why certain players have such low variance, well it's certainly open to interpretation. If you ask me this draft is so deep that it may not be about each team's needs but the talent level. Even so for a 20th pick. Actually this is the reason this collective ranking table makes sense. In another draft team-player matches would be absurd to say the least, but this year not so much. TOR for example picks Jusuf Nurkic at #20 just because he is collectively rated so and it makes sense too. If I was TOR and Nurkic was available at #20, I'd pick him regardless of how much I need him.

Very cool stuff. It'd be cool to see the actual variances although you get the picture from the website lists. I'd also be kind of interested to see a "mode" for each player to get an idea if a lot of places are picking someone to land in a certain spot or on a certain team.

This kind of stuff would be interesting to collect for future research. I'd be interesting to see how the variance of a player predicts their future performance.

[QUOTE=gutshot]Hey Thorpesaurous, thanks for the feedback.

Variance column is heavily about the standart deviation of each player's respective ranks on different mocks. This number is somewhat proportional to the average rank, i.e. a std-dev of 3 is huge for a top-10 player but not much so for a late first rounder so I adjusted those accordingly and come up with a reader friendly 'low/high variance' labelling.

If you ask me variance here is about two main things: 1. Player is a mystery of some kind, so everyone has their opinion given the little information available. The only player who does not follow this pattern is Dante Exum. He is between 4-7 among all 14 boards, which I find surprising. 2. Sometimes a player is well-known but he is so controversial by nature. Kyle Anderson is such a freak combining a unique set of attributes.

You asked why certain players have such low variance, well it's certainly open to interpretation. If you ask me this draft is so deep that it may not be about each team's needs but the talent level. Even so for a 20th pick. Actually this is the reason this collective ranking table makes sense. In another draft team-player matches would be absurd to say the least, but this year not so much. TOR for example picks Jusuf Nurkic at #20 just because he is collectively rated so and it makes sense too. If I was TOR and Nurkic was available at #20, I'd pick him regardless of how much I need him.[/QUOTE]


That's sort of what I figured. I was a math guy back in my day, and I was just wondering if went by actual standard deviation, or if it was just feel.

You can see the tiers lay out based on the ranges of the players. Obviously the top three guys form a concensus tier one, as none of them dip below those three spots.

The thing about the variance is that it makes total sense for the ranges of players to be projected to move more the further down you get. If you look at most tier schemes, the tiers get progressively bigger as you go down. It was just weird that those couple of guys are hanging out there with low variance in that fourth tier that would seem to indicate a great deal of variance. They must project as ideal fits within the tier to some couple of teams bunched together with super similar needs. Or they represent some sort of mini tier that puts them definitively between two groups of players and sort of isolates their projection.

Interesting post, thanks. Tiered analysis of draft picks will a challenge mathematically obviosly but it may make perfect sense for a usual draft but this one is crazy. Draft is so deep that talent may come before need in majority of cases. For example one of my sleeper picks, TJ Warren, has [URL="http://www.nbadraft.net/players/tj-warren"]93 points[/URL] overall and is ranked #21. #20 Tyler Ennis has 93 too just like #28 Grant. For comparision 2013's top 3 actual picks had: 95, 94, 93. Also about 10 NBA teams were tanking/rebuilding this season so they need players at pretty much every position.

Anyway the compilation is final, here are the top 10 consensus picks:
1. Wiggins 2. Parker 3. Exum 4. Embiid 5. Vonleh 6. Smart 7. Gordon 8. Randle 9. McDermott 10. Payton.. [URL="http://nbadraft.dahoops.com/"]check the full compilation[/URL]

Elfrid Payton has been moving up consistently last few weeks and he has finally made the top-10! Jusuf Nurkic is on fire as well. And as expected Dario Saric has moved down dramatically after announcing his decision to stay in Europe for another year or two.