The Art Rondeau Story: Allan Houston and Analytics

It’s MIT Sloan season again and while no Boxscore Geeks will be in attendance, a longtime friend of the show Art Rondeau, the only Zen Master that helped the Knicks reach the playoffs, will be there. To make sure Art’s work is front and center for anyone at Sloan that should care about Art and his work -- which, in our opinion, is every NBA team that sends a representative to Sloan, we’ll be reviewing it this week.

Art's Background

Art Rondeau developed a unique physical approach to fixing bad free throw shooting. Rondeau (which is pronounced just like Rajon Rondo, who could use some help with his free throws) then studied neuro-linguistic programming (NLP) in order to add a mental component to his free throw program. That mental component evolved into what Art calls his "mental zone" program. Since the mid-1990s, Art has used customized mental exercises to help athletes get out of slumps and stay in the proper mind frame when playing. He's helped players in the WTA, LPGA, World Cup Skiing, and the US National Fencing team. Of course, his claim to fame and reason he's a friend of the show is his time coaching Allan Houston with the 1998-2000 Knicks. Don't worry, we'll talk more about that soon. In "the real world" Art is a database administrator in the IT field. You can check out Art’s webpage here, we’re specifically linking to his section on Free Throw shooting. You should also follow Art on Twitter (@artrondeau)

My Motivation

For longtime readers, you're aware we've had Art Rondeau as a regular guest on the podcast, and he's someone I routinely quote. And candidly, it's because I am still baffled and amazed by the dataset Art gave me related to his “mental zone” work with Allan Houston. The NBA analytics field is a shrewd business. And yet, in a field that is willing to reward decisions for being "analytical", when, at best, they can be called educated guesses, I feel there is a huge area for more systematic improvements in coaching. When Dave Berri did his research on coaching a decade ago for the book "Stumbling on Wins", he found most coaches in the NBA were interchangeable. As I've gotten older, I find I've collected things that irritate me and I can't let go of. If you follow me on Twitter, you know I regularly get upset with the fact that Ben Wallace is not in the Hall of Fame. Two other things that keep me annoyed regularly is that NBA players haven't improved in the aggregate at free throw shooting in the past thirty years and that Art Rondeau's time with Allan Houston is buried in NBA history. So, at minimum, I'll keep talking about it, at least until the NBA fixes all three of said travesties.

The Knicks Story

As a sports coach, Art's normal "job" was as a fixer. If someone was having a slump, Art could come in and help get them back on track. Even before the analytics movement took off, Art recognized this was not stable enough to be "analytically sound." As luck would have it, an opportunity presented itself. Art was working with another Knicks player and met Allan after a game against the Lakers during the 1998-1999 NBA season. A few weeks later, he spoke with Allan after a Knicks practice and found out that Allan was in an 8-game shooting slump. This led to him helping Allan for one game. On 4/23/1999, Art coached Allan before the Knicks played the Charlotte Hornets. Allan played well above his usual level (a +6.5 Points over Par game, which was a Game Changer, we'll talk those more soon). The Knicks eked out a win against the Hornets, which was huge. The Knicks made the playoffs as the 8th seed, beating the Hornets out by one game. And as you may recall, the Knicks became the first 8th seed to make the NBA Finals. Lose this game to the Hornets and the Knicks would have missed the playoffs altogether.

As a result of this success, Allan Houston hired Art Rondeau for the 1999-2000 NBA season, with the express goal of making him an All-Star. They succeeded. Of course, as nothing with the Knicks can ever be simple, working with Allan had some challenges. For reasons beyond Art’s control, Art worked with Allan before just 30 regular season games during the 1999-2000 season, despite Allan playing all 82 games. The games weren't a continuous block, in fact, they wound up being a fairly decent sample of game types. While this was frustrating for Art, to put it mildly, it made a perfect data set -- check out the Allan Houston case study if you prefer numbers to words! Comparing Allan's performance in the mere 30 regular season games Art coached him to his performance in the other 52 is remarkable. At the risk of offending Knicks fans, let me sum up Allan Houston's career from an analytics perspective. Allan Houston was a below average NBA player with the exception of the 1999-2000 NBA season. And Allan Houston was a below average NBA player in the 1999-2000 NBA season if you remove the 30 games Art coached him. Allan Houston was tasked with being a scorer. And while he obviously had the talent to shoot, his shooting efficiency numbers were well below league average. Art was hired specifically to help Allan with scoring, and in their time together, Allan’s efficiency numbers were the best they’d ever been and far above league average. Despite the short amount of time with Allan Houston, there were some funny "consequences" of Art coaching Allan: 

  • As mentioned, Allan's game against the Hornets in 1999 helped get the Knicks into the playoffs, giving Jeff Van Gundy the chance to save his job (Van Gundy was on the hot seat in the 1999 season, and nothing short of a strong playoff push was saving him).
  • Allan's performance in games Art worked with him in 1999-2000 helped the Knicks earn the second best record heading into the All-Star break. As a result, Jeff Van Gundy earned the All-Star coach job by half a game (Pat Riley had been the coach the previous year, and couldn't coach back to back years)
  • Allan's performance earned him his first All-Star nod.
  • Allan Houston's performance netted him a lucrative $100+ million dollar contract in the 2001 offseason. A fun byproduct of this is in the 2005 CBA, the NBA included a provision to allow teams to waive players on overpriced contracts. Google “Allan Houston Rule”. Whoops!
  • I do feel it's worth noting that of Allan Houston's 839 regular season NBA games, 13 of his top 100 games were from his time with Art.

Again, it blows my mind how impactful a mere 30 games of one season had on the legacy of the Knicks. Of course, it's bittersweet. Allan Houston got an All-Star bid, and later a big payday. Jeff Van Gundy kept his job thanks to making the playoffs in 1999, and the All-Star coaching bid only helped his stock. Jeff stayed a coach in the NBA for a while and is still regularly calling NBA games. Luckily for us, Art was able to give us an amazing set of data, but we're hoping for more in the future. As one last thing, we want to talk about one other beneficial part of this study were two fun custom metrics.

Game Changers

Using the Wins Produced metric we can translate it into point-margin. The Points over Par metric tells us how much our player helps our team win (or lose) by relative to an average NBA player. If a player's Points over Par exceeds how much a team wins (or in the case of negative Points over Par, loses) by, then we consider the player's performance a "Game Changer." This is useful for seeing which players are impactful enough to completely change the game on their own. As we noted, Allan Houston had five game-changing performances under Art's thirty-one games of coaching. Of course, basketball is a team sport, so often it takes multiple players to win or lose a game. So it's amazing how well Allan played such that he directly changed the outcome of that many games.

I'm also excited by this idea catching on. Dean Oliver started Tweeting a metric called a "Threshold Win" (to my knowledge it first appeared this last year), which uses Net Rating versus point margin to determine the same thing. I'll say proudly too that I know Dean Oliver has read the Allan Houston Case Study multiple times.

The Rondeau Ratio

One part of Allan Houston's 1999-2000 season that stuck with me is that he was an above average player, and did earn an All-Star bid. That said, his performance was very lopsided. I've wondered about other such players, who might look decent or good on paper, but have lopsided performances. So I used the fun method. Take a players performances and sort them by Points over Par. Take their top 1/3 games and see what percent of their total wins are generated by those games. A completely consistent player (these don’t exist) would have a "Rondeau Ratio" of 0.333, but lopsided players will have a Rondeau Ratio > 1.0. In 1999-2000, Allan Houston's Rondeau Ratio was 1.02. This ranked 18th out of 82 for "above average" players. First, by the way, was future Knick Stephon Marbury at 1.28. Also, using Art's games instead of the top 1/3, Allan Houston's "ratio" is 0.81. Namely, Art's coaching games were remarkably close to cherry picking Allan Houston's best 30 games from the 1999-2000 season. Expect more posts on the Rondeau Ratio soon!

Summing Up

Again, Art will be at the Sloan MIT Analytics Conference. We think properly conveying information to players is key for success in the NBA. And we think NBA teams are still suboptimal in how well their players could shoot. So if you're at Sloan and get a chance to pick Art's brain, it's definitely something we'd suggest!

Loading...