27. A number that will stay with the Rockets franchise for a long time. They missed 27 straight threes and lost to the Warriors last season, and most of the world sees that as the main cause of them not making the playoffs.
The Houston Rockets have a different point of view. The leading pro-analytics team in the NBA did their analysis of the refereeing in their 2018 Western Conference Finals against the Warriors and concluded that referee mistakes stopped them from making the Finals. How did they reach that conclusion?
The NBA releases the last 2-minute report to transparently go over all decisions made at the end of games. They also make a record of all calls made in the entire game, and the Rockets used that report to project the net result those referee mistakes had on the score of the game. The Rockets concluded the consequence of the errors in Game 7 resulted in an 18.6 points damage to the Rockets. How did they reach this number?
The report lists 81 potential missed calls and non-calls. In a story by ESPN’s Zach Lowe and Rachel Nichols, several examples are given on the Rockets methodology:
For instance: with about 6:10 remaining in the first quarter, Stephen Curry drove on the right side of the floor. Gerald Green, the Rocket defending Curry, placed his right arm on Curry’s hip as Curry rose for a layup. There was no call. The NBA flagged it as a “potential infraction” — inconclusive, according to Houston’s analysis. The Rockets counted that as a mistake that cost the Warriors 1.8 points — a figure that appears to have been derived from Curry’s career free throw percentage.
This is what they did with every of the 81 listed potential missed calls and non-calls. I’d like to point out they did that for both teams, they also added up points the Warriors were “robbed of” and concluded the net result was negative for the Rockets.
The league released a statement that said they do not agree with the Rockets methodology of projecting player’s averages to free throws and shots that didn’t happen. The logic of assigning a probability to events that might have happened in a parallel universe is not acceptable because of a number the Rockets didn’t include in this report. 27.
If you went by previous data, you would never project the Rockets to miss 27 straight threes, yet it happened. That is why we keep coming back to watch the games. Statistical anomalies happen, and that makes it unpredictable. Daryl Morey would probably argue some of those 27 missed shots were influenced by referee mistakes, and that is true as much as it is a butterfly flying in China causes a hurricane halfway around the world, and Kevin Bacon worked with everybody in Hollywood.
The problem the Rockets are facing, and the next great leap in analytics is to accept that they are projecting human behavior – a very inexact science. The Rockets gameplan is heavily dependant on the team and James Harden, in particular, getting a lot of foul calls. When you build your strategy on something as unpredictable as a human making an instantaneous decision and blowing a whistle, that is a hazardous model to have. So how do you adapt?
You try to reduce the significance of the variable, maybe even skew it in your favor. Overflowing the media with stories on how your team’s chances are being influenced by refereeing to get a more favorable whistle seems like something an analytical person would do.