The Rookie Year Week 1: Accounting for the perils of small sample sizes

Posted on 11/04/2010 by


“The Road goes ever on and on
Down from the door where it began.
Now far ahead the Road has gone,
And I must follow, if I can,
Pursuing it with eager feet,
Until it joins some larger way
Where many paths and errands meet.
And whither then? I cannot say.” –Bilbo Baggins ,The Lord of the Rings J.R.R. Tolkien

One of the favorite parts of my favorite books, the walking song of Tolkien’s hobbits represents the promise and wonder of the future and of things to come. In every life, journey and endeavour, there is a beginning, a middle and an ending.

For the rookies coming into the NBA this year, this week represented a beginning full of promise and expectation. For some it was a tantalizing vision of greatness to come or the beginning of a long career, and for others a foreshadowing of mediocrity or abject failure. For the media and those who love the game, it was a retelling of one of our favorite stories, a reboot if you will with a brand new and exciting cast that even though it retreads old familiar ground will provide endless entertainment and wonder for years to come.

What make;s this compelling viewing is that we see ourselves reflected in their journey and their struggles. The battle to stop overthinking, stressing and stumbling and the beauty of finding that transcendent moment of unexpected greatness or tragedy is a universally relatable experience. And then there’s that moment when someone like Griffin or Wall just are for the first time on an NBA court

Just do it (Image courtesy of

For me personally as well, this is a week of beginnings and endings. I’m in the process of changing jobs this week. Much like an incoming rookie, I’m moving from a safe job to an opportunity with  more challenges and more potential rewards.  For this blog as well, it’s a beginning. In this my first full season of coverage, I am building the structure, traditions and series which will  allow me to keep you amused for years to come.

Ok What?

Ok, ok, i’ll stop with the flowery sportswriter prose and pathos before someone calls down Fire Joe Morgan on me. Let’s get to the business at hand.You came here for silly stats, funny pics and my lame attempts at wit. Let’s talk about rookies.

Griffin and Favors?

Before we start sling tables and numbers, if you’re a rookie here yourself go read the Basics. All the math,numbers, tables, and conclusions are Powered by Nerd Numbers.

With this series I will rank and review rookie performances during this NBA season. This will be one of the lynchpins of my blog for this season (and however long we continue to inhabit this space). As longtime readers of this blog know, nothing inspires me like the NBA draft and rookies. To illustrate and provide some background, here are a few of my pieces.

The Draft:

Part 1: Finding Elite Rookies in the NBA Draft or How the NBA Draft is a Lottery

Part1a: The Top 33 Rookies in the Past 33 Years

The WSJ Piece: Arturo Galletti Evaluates 30 Years of the NBA Draft for the Wall Street Journal

Part 2: Ranking 30 Years of Draft Picks

Part 3 :A Sunday Kind of Piece: Return of The Draft

Where I ask How good are GMs at finding talent?

(The answer: Not Very)

Part 3a: The Draft,The Draft,The Draft………

Where I concluded that:

  • Talent is always available in late in the first round.
  • The trick is finding it with some accuracy (which I postulate we can do).
  • Given that the identification is risky a later,cheaper pick is better.

This  lead to a lengthy draft strategy segment in my guide to running an NBA franchise (Build me a winner rev.2). Which you can go read for detail (it’s really good I swear 🙂 )

Based on my findings I wanted to see if I could build a model to predict rookie performance. So I did.

The Rookie Models:
I built two models (go here for the model build parts 1 & part 2 ). I called them : Yogi and Booboo.

This is a totally legit picture

To give you an idea of the value of the models I decided to look at:

  • The probability of landing a better than average player (>.090 WP48) for his first four seasons
  • The probability of landing a good player (>.150 WP48) for his first four seasons

I also decided to show this for:

  • Any qualifying pick (>400 MP in his rookie Year)
  • Any Top 5 pick
  • Any Top 10 pick
  • Any 1st Round Pick
  • And Both models.

And this was done for 1995 to 2009. The table is here:

The best performing scenario is both models calling for you to draft the player, followed by Yogi then Boo Boo then having the Top five picks. Yogi is more picky, Boo Boo casts a broader net and is more accurate.

So with a good model in hand, I got to work on evaluating the incoming rookie class.

The Rookie Projections for 2010-2011:

The full rookie projection is here

For the reader’s digest version, It took all the player data, fed it into the models and got this (ranked by  projected wins):

Only Griffin projected  out as a difference maker for his team based on the model and his college numbers.

At reader’s requests, I then added some of the undrafted players:

The final component in this recap is  rookie performance for the preseason:

Still here? Good. Now that you’re all caught up on all my silly little stats about rookies and the draft from the preseason we can talk about how the rookies have performed since the season tipped of.

The Rookie Year Rankings for Week#1

No preamble just the numbers. 46 rookies have played so far in the regular season and there are some standouts. Landry Fields is one (as noted by WOW godfather Prof. Dave Berri). Favors,  Wall , Gary Neal (seriously, pop and San Antonio strike again), Evan Turner and Blake Griffin follow. Now while someone like John Wall has exceeded my and the models expectations so far, before we get too excited and I’m asked to issue an apology to Ted Leonsis we need to be aware of a simple statistical truth: 3 to 5 games is a small sample.

The  law of large numbers (LLN) describes the result of performing the same experiment a large number of times. It’s a simple enough theorem, the average of results obtained from a large sample (or number of trials) will get closer and closer to the real value of something the larger the sample. Conversely, the error (or more accurately  the possibility of it) gets larger and larger the smaller the sample . What does this mean?

Something about popsicles and not getting ahead of ourselves

So rushing to judgement based on a small sample is premature. A larger sample size is called for before we can make any solid conclusions. We will simple need to wait a few months before the in-season numbers tell us anything concrete.

Well,  I for one say: that’s no fun, nuts to that.  Luckily enough I,  I came up with a nifty work-around. Take all the data available, mix it all up and voila!

Ok, i’ll explain. Here’s what I’ll actually do.


  • The Regular Season Wins Produced numbers and minutes played for each rookie
  • The Preseason Wins Produced numbers and minutes played for each rookie
  • The Projected Wins Produced numbers based on their college performance (for minutes here we’ll treat it as a 10 game sample times the minutes played per game in the current season) for each rookie
  • Then we take a minute weighted average of  the three and use that to rank our rookies.

The advantage to this is that it increases our sample size and allows us to project the rookies  based on a combination of  who their college numbers tell us they were, what they showed in the preseason in a small sample and what they’ve done in the pros (so far). As the season goes along our regular season minute sample will increase for players (on average) accordingly and we’ll be inch closer and closer to their true value (and the importance of the projection and the preseason will fall away). I call it the Combo-Breaker.

ēower mǣte sample micelu nāht ætstandan mec

Using this method,the Rookie rankings look like this:

Griffin regains the number one slot based on his ridiculous performance in the preseason and his projections. Landry Fields nips at his heels at number two.  The player some have called a “bust” Evan Turner for not scoring is liked by all parties and comes in at number 3. Evans, Harris and Omer Asik are also liked by all parties involved for 4 thru 6.

John Wall  has exceeded his projections both in the preseason and so far in a small sample size in the regular season to move up to number 7. Derrick Favors has destroyed his preseason numbers for number 8. Both of these guys will move up massively by next week if they keep it up. Five more players round out the group expected to contribute at least 2 wins : Paul George,Gary Forbes, Tiago Splitter (who’s riding his projection right now and may drop rapidly if he doesn’t perform), Harvard man Jeremy Lin and the starting center of my fantasy team Demarcus Cousins.

After this group, the most notable player is Jeff Adrien, who’s totally killed it when he’s gotten on the court in the regular or pre-season.

From a team perspective we see:

The Clippers and the Knicks unsurprisingly project to get the most value from the draft with the Nets and the Wizards so far exceeding projections. One final note is that right now, for the Wizards, John Wall literally and figuratively is the franchise (as in he accounts for 12.9 of their projected 13.2 wins).

Game over for today folks

Posted in: Uncategorized