A Different Way to Generate NBA Lottery Odds

The NBA is the only major American sports league that does not have a set draft order for each upcoming draft based on the win-loss records of the prior season. In most sports leagues, the worst team (in terms of win-loss record) will automatically receive the #1 overall pick of the upcoming draft. But the NBA takes non-playoff teams and gives them odds to have certain picks in the upcoming draft based on their win-loss record. While the expected draft pick for each team is correlated with the inverse of the team’s win-loss records, teams may end up with better or worse draft picks than expected due to the lottery system.

The NBA implemented new lottery odds for the 2019 NBA Draft, and slightly changed each team’s draft odds. I say slight because the whole system of ordering non-playoff teams by records from worst to best and distributing odds based on that is still the same, albeit now the top-3 teams have worsened chances of landing a top-4 pick (although they still have the best chances) while teams 4–14 have improved chances for a top-4 pick.

But months ago I had proposed a much more radical system which I have never seen discussed ever (by proposed, I mean I actually submitted it to nba.com, but I foolishly thought a copy would be saved into my email and therefore I no longer have my original thesis): rather than distribute draft pick odds based on the team’s records alone, distribute draft pick odds based on the team’s record in relation to the records of the entire league by way of standard deviation. Standard deviation, if you remember high school math (I don’t unfortunately, so hopefully a statistician can correct me), helps identify what we expect to be ‘normal’. And while we don’t expect all teams to perform the same, there is a range of ‘normalcy’ with which we’d expect teams out to win would perform at. What I mean is that given the same resources (salary cap space, roster size, etc.), we should expect NBA teams to perform similarly. Of course, each team has different players, but that’s the point of the draft: to provide bottom teams a way to improve with better players.

But the problem with the NBA (and why the NBA implemented the draft) is that NBA teams could easily sink (or ‘tank’) themselves to the bottom to increase their chances for top picks. Of course, only the very worst teams could realistically tank and leapfrog each other to have the worst records and therefore the best chances for the highest picks.

In the 2018–19 season, these would have been the Knicks (17 wins), Cavaliers (19 wins), Suns (19 wins), and Bulls (22 wins). The 5th worst team by record (Hawks) had 29 wins are would have had far too much difficulty to try to drop into the bottom 4. (You could make the case that the Bulls at 5 wins above last place and 3 above 2nd to last were already too far ‘ahead’ to drop into the bottom 3). While I don’t believe any player or coach purposely tries to lose games, I do believe organizations do not always attempt to win. It’s the same with any organization and it starts at the top. If management at the top want a certain outcome, they are the ones to best manipulate the people below to achieve said desired outcome. And in the NBA, that means a number of things from holding players out of games (‘injury concerns’) to not signing or dumping players who would otherwise improve your team (‘cap management’) (it’s one thing to do these as a championship contender, but it’s another to do these when you’re one of the very worst teams in the league).

Look at the worst teams this season: the Knicks traded practically all their serviceable players in a midseason trade. The Cavaliers brought back their best player Kevin Love back from injury in February but still alternated between playing and resting him and even had former starter J.R. Smith removed from the team since November. And the Suns didn’t bother signing a point guard, so they were playing one of their two rookie 2nd round point guards for nearly the entire season. And the same pattern often follows the teams that gain the best lottery odds in the upcoming draft.

But not only does this have a tangible effect on team records, but I believe this also has more unseen effects on organizations, coaches, and players. For players, they are giving away one season of the few they play in the NBA for nothing. Sure they get paid, but chances are they aren’t playing as hard as they should knowing the team doesn’t care about winning. And chances are they aren’t developing as much either. The same goes for coaches. And bad coaches I find are all to often retained on bad teams in order to minimize costs as a better coach would cost more. All this I believe creates more indifference within an organization which makes creating a winning culture all the more difficult.

And yet its these bad organizations who actually receive the prize of the best draft odds for new talent, instead of bad teams that were actually playing hard and trying to win (the aforementioned 5th worst Hawks being the prime example, although most non-playoff teams outside the bottom 3 were fairly fun to watch) and therefore have slightly better records.

So back to my proposition: we take wins for each team and identify the ‘normal range’ of wins and provide the best lottery odds to teams closest to the worst records we’d expect.

With that said, I identified the standard deviation of wins across each of the past 11 seasons. From from the average number of wins per season (usually 41, but not for 2 seasons), I subtracted 1.644854 times the standard deviation and took this number to be an expected number of wins for our worst team (why 1.644854? the number comes from the central limit theorem that states we’d expect about 90% or 27 teams would be within that range. Assuming the 3 teams are a combination of both best and worse I find that to be reasonable, although I’d love to hear from a real statistician a better multiple). I then found the teams with the number of wins closest to our expected wins and gave the draft odds based how close they were. If this had done for the past 11 years, we’d have the following draft orders (I posted the new order up to the point where the order remained the same):

As you can see, it’s almost always just the team with the worst record that gets bumped (although it would be different depending on the multiple of standard deviation used). Although with the new draft lottery odds, only if a team falls outside the bottom 3 does it really impact it’s draft odds. And while I’m skeptical of 20/20 hindsight, I think most of the worst teams were not attempting to win and I think the statistics seem to pick up on that.

2015–16 Philadelphia 76ers: they were probably the straw that broke the camel’s back in creating the new lottery odds as they were terrible from the get-go (8.5 wins behind what standard deviation was expecting!) and eventually landed Ben Simmons.

2014–15 Minnesota Timberwolves: they were going to be bad after Kevin Love was traded. But they were only 3.2 wins behind our ‘target win’ count. Maybe if they had put in some more effort in their offseason (like, actually trying to find someone who would actually carry the scoring load in place of Love) they would have been a bit more competitive.

2013–14 Milwaukee Bucks: this team was 4 wins worse than a 76ers team that was still part of The Process. The leading scorer was Brandon Knight at 17.9 PPG, followed by Ramon Sessions at 15.8 PPG. (Yep…) But perhaps most telling was that no player on this team was paid more than $8M this season (and that was…O.J. Mayo). Again, I’m going to have to chalk this up to the fact that the organization wasn’t even trying from the get-go.

2011–12 Charlotte Bobcats: this was was 13 wins worse than the next worse team. It actually set the record for worst winning percentage in a season. It’s hard to say that any organization trying to win could have ended up this badly, even with a lockout shortened season. It’s quite possible this team wasn’t even bothering to win games at a certain point knowing it was looking forward to presumed #1 draft pick Anthony Davis (essentially the Zion Williamson of the 2012 draft).

2009–10 New Jersey Nets: this was that team that started the season 0–16, then fired Lawrence Frank, then fired the interim coach (Tom Barrise), and then put their general manager Kiki Vandeweghe as coach for the rest of the season. This was a disaster, and I think that was the plan. One factor? They made zero offseason free agent signings before the start of the season leading them to start a mix of Brook Lopez (fine), Devin Harris (ok), Courtney Lee (ok), Yi Jianlian (and this is where it’s clear there was zero effort put into the roster), Chris Douglas-Roberts, 33 year-old Rafer Alston, Jarvis Hayes, Trenton Hassell, and Josh Boone.

Please note that while the above teams (and the aforementioned 2018–19 Knicks) would drop a few spots, the actual drop is chances for the best pick with the new lottery odds is closer to 5% which is hardly game changing. On the flip side, more competitive teams would gain those few percentage points leading to a number of benefits.

Too complicated? Perhaps maybe, although I think the NBA’s trend toward analytics makes this much more palatable now than say 3 seasons ago. The NBA is finding out now traditional methods were maybe not the most effective methods of performing the same task. And I think the traditional method of ranking draft odds was creating unintentional consequences that are reduced with these new odds.

For one, teams will try to win more and there will be fewer ‘lost seasons’. Almost all the cases of ‘lost seasons’, where prior to the season starting everyone from the fans to the owners already have written off the season, are so extreme that the team has a win-loss record that actually causes their lottery odds to drop in our new algorithm.

This means even when teams know they have no shot at the playoffs, they will still need to compete. That means at least trying in the offseason to sign players (albeit I’m sure at reasonable or short term contracts). This also means in better coaching hires and more urgency with getting the right coach. (The coaches for the Bobcats and Bucks teams above were Paul Silas and Larry Drew respectively. Both of them lasted the entire season and neither ever coached again after those seasons. Ownership could have at least cut them during the season to test an assistant coach who would try to win.)

This means even during the last stretch of games, non-playoff teams will still play to win games at the risk of losing too much and dropping their draft odds. It means we’ll have fewer games like this.

This means more teams who are willing to sign players before and during the season, which is definitely better for the players. It means more urgency in finding better coaches, which is better for everyone from the players to the organizations to the fans. It means better games and more reason to follow their team, even if they aren’t contenders. And perhaps most importantly, it means organizations are going to start getting in the habit of doing what it takes to win, even though they aren’t going to.

It’s easy to say that you can just cobble together your championship team with some splashly free agent signings, but I don’t think that’s true. For one, look at how often a championship team at least made the playoffs in the previous year. The only one I can think of is the 2007–08 Boston Celtics over the past 40 odd years. What I think that implies is championship habits don’t occur over one season. And more so I think that organizations can’t build championship teams until they know how to build playoff teams. And they can’t build playoff teams until they start getting their teams consistently in the playoffs. And they can’t do that hoping a rookie draft pick is going to do it for them. They need to know how to operate in free agency and just reach their goal of building a playoff team that doesn’t break the bank (see 2018–19 Los Angeles Clippers and Brooklyn Nets which despite a lack of stars have managed their cap space well, have made the playoffs, and have enthusiastic fan bases now despite not being contenders).

The NBA has for the most part tried to help team owners not screw up their teams. Think about the salary cap, maximum contracts, and capped rookie deals. Or the Ted Stepien rule about trading future draft picks. Or the Allan Houston Rule on terminating contracts. I think this new draft order ranking would not only help NBA teams not screw up their organizations, but also indirectly promote the skills to create better organizations and teams by encouraging every team to consistently win.