The art of lean problem solving, part 2: Lessons from NBA coaches

Hi everyone. I'm Art Smalley, president of Art of Lean, Incorporated. Today, on behalf of the Lean Enterprise Institute, I have another little video for you. This one won't be short. It's going to be a bit longer, because we're going to go deep. We're going to step into the world of NBA basketball, talking about coaching and problem solving and studying it through a lean lens. I think the NBA could benefit from the lean way of coaching. On the other hand, I think we have a lot to learn from the NBA and from the world of sports in general in terms of coaching as well. So stick around; I think you'll enjoy it.

Practice makes perfect

In a previous video, I talked about coaching and problem solving. I'm going to do a series of videos with LEI and, at the end, I'll explain my framework again, probably modifying it based on what I've learned going through this process with you. However, I think coaching involves, at a minimum, recognizing the types of problems and situations. A problem can be of low, medium or high complexity and involve various skill-will combinations between the coach and the participants.

Of course, on top of that, there are urgency and external noise and leadership factors, but let's ignore those for now. First, let's watch a five-second NBA video clip and look at a real problem situation that was coached, practiced and put into play. I'm borrowing material from an Atlanta Hawks D-league coach and scout named Dylan Moore. You can study this in more detail if you want, but I'll show the five-second clip and then we'll talk about the problem solving and the coaching that led to it.

This is the result of good problem solving and defensive coaching to slow the game down. This is a pick and roll situation. The center comes up to set the screen, a defender shows, and they switch back; the point guard stays on the point guard, and he has to pass to the forward. No three-point shot was allowed, and no easy drive to the paint was obtained on this possession. So, technically, the defense won this little battle. OK, I know the play was quick (too quick for the many people who don't know much about basketball). I'm going to break the play down and explain the whole coaching and thinking process behind those five seconds of NBA play, but first you have to understand that problem solving and coaching are very much present in sports.

More than just passing the ball

Frankly, I wish the NBA and the teams would communicate in a language more people could understand, because basketball (or any sport, really) is a game of situations. It isn't static; to simplify, we can say there's the offense, the defense and special teams, with special situations. On offense, there's transition, the fast break (which comes from rebounds and steals), various kinds of inbound plays and the traditional half-court offense. In the half-court offense, one thing in particular you see at every level (and one of the oldest plays in basketball) is what's called the pick and roll.

If you study basketball, you'll find there are several types of pick and roll. There's a high version, a side version and combinations of the two. There are also various combinations and permutations of who runs the play: point guards, shooting guards, forwards, centers… and some teams play a style in which there are no positions as such. So you have to be prepared for all of that. Now I'm going to make up a specific problem. Again, I'll borrow the example I found on medium.com from the Atlanta Hawks D-league coach, Dylan Murphy. This is a made-up example, but it happens a lot in reality. On average, a team has a hundred possessions in an NBA game. The average points per game is currently around 112 per team. That average has been rising in recent years. The average margin of victory is about 10 points. It varies from season to season, but the average is 10 points. When you think about it, you realize that you need to have, on average, five more defensive wins than your opponent per game. If you think defensively, you'll need to keep them from scoring five more baskets per game (let's think in terms of two-point baskets for now) to wipe out the 10-point advantage. You need to score more points or they need to score fewer. Some combination like that has to happen.

The game is won by these narrow margins. Now, here's a generic situation I've seen happen many times: they do a lot of data collection, a lot of video analysis and a lot of game observation, and that's how databases get built. This is something that isn't exposed to the public (I wish it were). A typical opponent, let's say it's our next opponent, scores about 40 points per game from the pick and roll offense. Maybe 30 points come from other offensive plays, 20 points come from the transition game I mentioned and another 20 come from free throws and things like that. It's always a mix of these factors, but they know, opponent by opponent, what that proportion is. Let's say we're facing an opponent that's really good at the pick and roll and we want to hold them from their current average of one point per possession on the pick and roll to something lower than that (the top three teams in the league on the pick and roll score more than one point per possession on it. They're very good at it). How are we going to stop them?

Setting goals based on previous experience

Just as in problem solving, we can achieve that goal. Some coaches do. And they do it by setting goals for the game. These goals aren't generic objectives like the ones all of us fans set (we're going to win the game. We're not going to get hurt. We're not going to go to overtime; we're going to win in regulation). They set more specific goals. They might say: “we're going to defend the pick and roll effectively in this game”. More specifically, they say: “we're going to specifically defend the side pick and roll between the point guard and the center”, if that's their strongest point. So we'll try to bring their average down from 1 point per play to 0.8 (there's no way to get to 0, but if you can stop them even a little, over a series of possessions, it makes a difference and you win the game).

So that could be a goal for the next game. To do that, of course, we have to do an analysis, and that's what the announcers and TV crews don't show (I wish they would). A pick and roll analysis shows how the opponent scores points. Of those 40 points I mentioned, they run the high pick and roll 20% of the time and the side pick and roll 80%. The real numbers are different; I'm just creating an 80-20 rule for this example. Looking at the side pick and roll, the point guard and center combination is used 80% of the time, shooting guard and center 15%, and point guard and forward 5%. Every team is different; you need to know the data, the details of who you're going to face and how you need to defend. That's what Dylan Murphy explained in this article. I encourage anyone who's a basketball fan to read it.

Now, just as you have analysis and problem solving in basketball, you have to come up with countermeasures to stop the offense. There are five generic ways to stop the pick and roll (of course there are more than that, you need to dig deeper, but let's focus on the five most generic ones): 1) Over the screen; 2) Under the screen; 3) Double-team and trap the ball; 4) Switch; 5) What we call the hedge: threaten the switch and then recover to the original assignment.

They all have pros and cons. That's where you, as a coach, need to know which one will work best for the next game. Good coaches and good scouts need to prepare for all of these scenarios, not just one. It's also no good just asking open questions like “what do you think the problem is? What's the obstacle? Let me read my card and ask a few questions”. At the highest levels, that doesn't work, not in the NBA, nor at Toyota or in good lean companies.

Breaking down the play

Now, for anyone curious about the play, I'll break it down step by step. First, the point guard dribbled to the side (you could call this a high side pick and roll). Then the center came up to set a screen (a kind of block in which the center, in this case, puts himself between the defender and the player with the ball, the point guard in this case, to open up space for him). This creates confusion for the defense and frees up space for the point guard to dribble and get open. In the third part, the defensive center steps out to guard the point guard (a big player guarding a smaller one), threatening the switch. They make the offense think they're going to switch, but in reality they go back, because you want the smaller player guarding the smaller one and the bigger one guarding the bigger one.

It doesn't always work like that, because it depends on the game plan and the analysis done, but they determined, in this example, that they didn't want to use the other types of defense. They wanted to hedge and then switch back, because they concluded that this would give them a bigger advantage in this game. They came to that conclusion by studying previous games. So, in this case, they threatened to switch (and temporarily it looks like they're out of position), but in reality, when they go back, the player recovers and shuts down the pick and roll threat. The point guard doesn't try to shoot the three, and the center trying to drive into the paint is well defended. So the main pick and roll options are eliminated right there. The point guard had no space at all and couldn't get the easy three-point shot. Nor was there space for the center to get an easy give-and-go when driving into the paint. You can conclude that the offense was disrupted, because the point guard had to pass the ball to the forward. The defense won technically. They won that little battle and stopped the offense.

In a basketball game, you have 24 seconds to complete each possession. You have to get these defensive wins successfully many times if you want to win the game. Doing it once isn't enough; you have to do it repeatedly. At the end of games, we all look at the scoreboard and ask: “Who won?”. Let's say the team that defended the pick and roll in our example won 110 to 102. The real question they take home to study with the analytics department (the video people sometimes stay up all night studying) is “How many points per possession did they get on the pick and roll? How well did we really defend the pick and roll? That was our goal”. Maybe our goal was to give up only 0.8 points per possession, but they got 0.85. So we did well at it (better than in the past), but we didn't hit our goal.

Even so, we won, because the analysis doesn't have just one variable. There are several things going on. Maybe we limited our turnovers and got more fast-break points or something like that. It's always a combination of things that results in a win or a loss. But there's another game in 48 hours. The NBA has a frenetic pace. You may have another game in 24, 48 or 72 hours. You have another game and you'll probably face another team's pick and roll and have to decide how to defend it all over again. So they have to create action items every day after a game. They have to review video of previous games with the players in a meeting. They need to go over the hedge technique again and practice it with coach Dylan. They have to go over the finer points of how to hedge and what the offense will do to counter it.

That's mainly what Dylan's article is about: the hedge, a defensive technique, and the six techniques for countering it, which I'll list here: 1) the counter slip, 2) the counter split, 3) the hedge and the mark, 4) the hedge and the empty corner, 5) the hedge against the pick and pop and 6) small hedges to avoid the switch. These are the things they'll try to do to confuse you next time. They have strategies against your strategy. So you have to prepare video to study all of these things. You have to practice, practice, practice and train the skill over and over again.

Basketball coaching applied to lean

I know the NBA is very different from the lean world. They have big contracts, a lot of money and a lot of people to study problems, but I still think we can learn a few things from this example. Number one: the season has 82 games, with many workdays throughout the year, just like most of our jobs. There are many types of coaching, many types of problems and many types of situations and practices they do, not just one. Second: they make very good use of video and analytics to figure out what to coach. We use analysis and lean to frame the problem definition and analysis in our best examples, but we don't usually do a good job of using data and doing video analysis for study. We just have a few ready-made questions, show up at the meeting and do our best.

Planning what to coach specifically for each game is important in the NBA, not just general coaching. The third lesson is that, in lean, we have to be more specific about what we plan for our coaching and how to do it. Number four: practice planning with specific situations in mind; plan for each person. The NBA doesn't coach one player; they don't coach only the point guard. You have to coach all five positions on the court and the bench. In lean, I wish we did this better too. Let's not pretend there's only one person in coaching, only one style. What you want the operations manager to do is different from what you want the lean specialist to do. All of that is different from what you want the quality manager and the engineering manager to do. Of course, they all generally do the same thing, but just as in the NBA, you also have specific roles that I think we need to improve in problem solving.

Number five: There are highly developed patterns of situational awareness in basketball because they record and study a lot. They study five different ways to defend the pick and roll. I'd like to have that in the lean system. Five ways to frame the problem, five ways to break down the problem definition, five ways to think about your goal statement, five ways to think about your root cause analysis, five ways to think about countermeasures. Variation still exists and needs to be thought through, but I'd like us to reach that level in our coaching too. Numbers six, seven and eight are similar, but I think the important thing is that, in that five-second clip, you didn't notice any coaching When you succeed, you don't see the coaching directly on the court. The coaching happens before the game. During timeouts and halftime there are only adjustments; what you can do is quite limited at those moments. There's a lot of coaching again after the game. I wish we thought more like this in the lean world: before the problem-solving meeting or before the problem-definition activity, we should do a lot of coaching beforehand. There's some refinement during the presentation or execution and then a lot more coaching afterward to keep heading in the right direction.

Number nine: Another thing worth reflecting on about the NBA, the military and Toyota is that coaches are developed internally. That doesn't mean you can't hire someone from outside, but extreme emphasis is placed on developing and growing people and on learning coaching as part of their job throughout their career. Number ten: there are different types of coaches besides the head coach in the NBA: head coach, assistant coach, strength and conditioning coach, offensive coach, defensive coach, analytics specialist, scouts, coaches who deal only with dribbling, coaches who deal only with shooting…

There's a lot of room for improvement in lean coaching; we have this generic coaching, and I think we should be more specific about the type of coaching we're trying to do and the types of coaching we can employ to do it. I fully understand we don't have the NBA's big budget to do this, but I want all of us to think about it, brainstorm and look at what we can do better by using ideas from the NBA. I hope you enjoyed this video. We'll do another one soon, but I'm going to leave the world of sports and step into martial arts. It may be different from what you're imagining.