Einstein and MacGyver Fix the NBA Tanking Problem

Back in September 2011, I published two articles aimed at solving the NBA’s contract dispute with the NBPA (the players’ union). I called them “Einstein and MacGyver Fix the NBA Lockout” (parts I and II).  Einstein’s the genius who said (roughly) “the thinking that got us to where we are today is not the thinking that will get us to where we want to be tomorrow.”   MacGyver is well-known to millions of TV viewers as the secret agent who could create solutions to seemingly insurmountable problems using duct tape, a Swiss Army knife, and whatever else happened to be lying around.  They had some pretty good takes on how to end the lockout, some of which would still make basketball better today.  If it’s possible to fix the NBA’s tanking problem (and it is), these two are the ones to do it.

What is Tanking?

This is when a team feels they need to rebuild using the draft and then build their team in such a way as to ensure losing a lot of games. They do this because the best way to get a great pick in the NBA Draft is to have a lot of ping pong balls in the annual NBA Draft Lottery.  And the only way to get more ping pong balls is to lose more games than other teams.  The team that loses the most games has the greatest chance of winning the lottery, although, oddly enough, they rarely do.  No one thinks players are missing shots intentionally in order to lose games.  But if the team has average players and no stars, the chances that they’re going to win are minimized.  A “draft lottery reform” proposal set to be voted on by the NBA Board of Governors late this week is supposed to make it a little less attractive to be terrible but it’s certainly not enough to dissuade teams from doing their worst when they think it’s in their best interest.

Do Teams Really Tank?

Not according to the NBA League Office as late as a few years ago. But after Sam Hinkie’s “the Process” in Philadelphia, even NBA executives now admit what everyone else has been saying for a long time: teams intentionally lose to improve their odds in the lottery.

Problems vs. Symptoms

Tanking is not just a problem. It’s a symptom of a bigger problem.  We often focus on fixing symptoms because those are what gets our attention.  A trip to the cold and flu section of your local pharmacy will provide proof of that.  But fixing the symptom doesn’t fix the underlying problem.  And the proposed lottery reforms that will be voted on later this week don’t fix the underlying problem, either.

What is the bigger underlying problem that tanking is a symptom of? Easy.  That actual problem is that the only way to get more ping pong balls in the NBA lottery is to lose more games.  The worse you are, the greater your chances of winning.  And that, in today’s NBA, is wrong, wrong, wrong.  It teaches players to lose, causes financial damage throughout the league, and provides a bad example to young fans.

If a restaurant isn’t making enough money, would you recommend that they start selling dog food instead of high-quality steaks for a few years until they decide to sell high-quality steaks again and regain their competitiveness and their customers? Of course not.  But that, in effect, is what getting more ping pong balls if the quality of the team is lower promotes.

If a student gets a lot of Fs on tests but gets a couple of As at the end of the year, would you reward them for the Fs and then punish them for the As? Of course not.  But that, in effect, is what happened to Derek Fisher at the end of the 2014-2015 season when winning a couple of games cost the Knicks last place as the Timberwolves ended up with the worst record and the most ping pong balls.  Coincidentally, Jeff Hornacek’s 2016-2017 Knicks won their last game, sending them into a tie with (you guessed it) the Timberwolves for the 6th worst record.  As was inevitable for the Knicks, the T-Wolves won the coin toss and “earned” the better pick in the draft than the Knicks received.

Should We Do Away With the Lottery?

No. The lottery concept, which came about after a lot of experimenting, is fine.  As Winston Churchill is reported to have said in 1947 “Democracy is the worst form of government, except for all the others.”  Likewise, the lottery is better than what came before it and better than any of the alternatives that I’ve seen promoted to date.

How Do We Fix the Problem?

Here’s where Einstein and MacGyver are at their very best. If, as Einstein said, the thinking that got us to where we are today isn’t the thinking that will get us to where we want to be tomorrow, can new thinking help us change the lottery in such a way to make it work better for everyone and eliminate teams losing – and losing big – to win more ping pong balls?  The answer is “yes”.

How do we fix it? Here’s where MacGyver comes in.  Unlike when the lottery was established over 25 years ago, we now have Advanced Analytics lying around.  MacGyver loves to use what’s lying around to solve problems, so here goes:

The way to fix the lottery process and stop teams from tanking is to give out ping pong balls for playing well, not for playing badly.

How Would That Work?

Using a number of advanced analytics formulas, it would be possible to determine how one team’s players should do against another team’s players in a particular game. Let’s say that the “tanking” team (I’m using the term here to identify who needs the lottery pick) is expected to be a 15-point underdog to one of the Conference leaders.  But the final score either shows the tanking team winning, losing by 5, losing by 10, losing by 15, or losing by 20 or more.  What would happen?

If the tanking team was projected to lose by 15 and then loses by 15(ish), they get no more ping pong balls but don’t lose any, either. If they actually win the game, maybe they get 5 ping pong balls.  Losing by 5 instead of 15 might get them 2 ping pong balls and losing by 10 instead of 15 gets 1 them ping pong ball.  Maybe we take a ping pong ball away from them if they lose by 20 or more in this example.

We therefore reward play that is better than expected and penalize play that is worse than expected. There is no incentive to play badly.  In fact, there is a deterrent.  But play better than expected and the team will be rewarded for it.  It’s not just about winning and losing (and losing big) anymore.  “Good job, good effort” will actually win ping pong balls.  Ping pong balls would be awarded on a game-by-game basis and could be awarded to all teams throughout the season, not just those who aren’t in the top half of the league.  Non-tanking teams that aren’t eligible for the lottery might use those earned ping pong balls to help determine their seeding for the NBA Draft, as well.

Could a Tanking Team Make the Playoffs?

You bet. They’d probably pick up a whole bunch of extra ping pong balls along the way.  And if they’d been identified as a “tanking” team at the beginning of the season, they could not only be in the playoffs but be in the draft lottery as well.

Would There Be Any Checks and Balances?

Absolutely. One thought that Einstein passed along was to create a game simulation program and feed in all the stats based on how the players have played so far that season and then re-run the simulation multiple times after the game using actual time played and who different personnel played against.  This would not only provide more accuracy but would be interesting for the fans, as well.

Running the simulation multiple times (and averaging out the results) while reflecting exactly what happened in the game (players, opponents, and minutes) would also eliminate the possibility that a star player is listed as unavailable before the game so the expectation is that the team will lose big and then ends up playing and earning the team more ping pong balls by beating the line.

Can We Start Awarding Ping Pong Balls for Good Play This Season?

Probably not. We need to define what constitutes a “tanking” team (as well as coming up with a better name for them than “the tanking teams”).  We need to decide if tanking teams are just those who announce they are tanking or if being rated below a certain metric automatically makes a team a tanking team.   The NBA needs to keep the level of play up across the league, even when a team doesn’t realize they’re not looking so good.

We might be able to use what Las Vegas uses to determine point spread but we still need to identify which of the advanced analytics measurements make sense to utilize for this process and maybe develop some new analytics measurements, as well. We need to write and test the game simulation modules.  We need to agree on what different levels of good play will earn for a team and what different levels of bad play will cost them.  We need to find – and plug – the loopholes in this system so that teams have no incentives to keep their better players off the court (whether to earn more ping pong balls or to help the other team earn more ping pong balls) and have no incentives to play their best players too many minutes.

We want teams to play as well as they can for each of their 82 regular season games and still give them a chance to rebuild when they need to.  This is the way to do it.

Getting Started

The best way to get started on this is to create a working group to develop the solution. I’m guessing that NBA Commissioner Adam Silver would head it up.  We’d want input from NBPA Executive Director Michele Roberts as well as from NBA Deputy Commissioner Mark Tatum, Executive VP of Basketball Operations Kiki VanDeWeghe, and Senior VP of Basketball Strategy and Analytics Evan Wasch.

We’d also want a couple of analytics-heavy GMs like Daryl Morey and John Hollinger to weigh in.  We’d certainly want Dean Oliver to participate and I’d suggest Andre Alvarez, as well (Dre did an analysis of my NBA work and I know him to be someone who is great at finding holes in someone’s arguments and proposing ways to fill in those holes).  Maybe Daryl can recruit a team from Sloan to help with the modeling and simulation.

I know the NBA is sponsoring a Hack-a-Thon and maybe some of the participants will have some ideas.

From the media, ESPN’s Kevin Pelton has a great grasp of analytics and already uses them to predict wins and losses over a season.  FiveThirtyEight’s Chris Herring would be a good addition to the team, too.  Of course, there are other analytics experts, as well as top thinkers in other categories, that we might want to add to the working group.

Einstein and MacGyver are, unfortunately, unavailable but have suggested that I add my name to the list of those who should be in the working group.  I’m a long time IT guy who specializes in problem-solving, has developed two unique programs that have succeeded at the NBA level, and have written game simulations.  On a group of intellectual heavies like the ones I’m proposing, I can at least be the guy who goes out for coffee.

Once the team is assembled, we’d start by defining the objective and the known issues that have to be addressed for this solution to be a success. Our goal would be to have a working model ready early enough so it can be presented to the NBA’s Competition Committee in enough time to be implemented for the 2018-2019 season.

What Are Your Thoughts?

This is a great blog for you to throw in your two cents in the Comments section. Let’s hear the ideas before we attempt to implement the solution, not afterward when people normally ask “why didn’t you think of this?”.  Eliminating tanking is important to the NBA, the players, the coaches, teams, and fans.  Let’s reward winning and effort and stop rewarding losing and pathetic play.  Let’s do this right!


To Einstein and MacGyver for once again being so inspirational. And to you for taking the time to read this.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: