How Could the NBA Be Even Better?

Is the NBA in need of saving? Is it the organizational equivalent of a damsel trapped in a tower? Am I the one man on Earth who can save them?

I hope not, that’s a lot of pressure.

The NBA is doing just swell. They’re a global cultural institution and still growing. I just have some ideas on how to make the league even more entertaining.

Expansion

The league is the most deep with talent its ever been, yet it’s been eighteen years since its last expansion draft. Bringing the league up to thirty-two teams is a financial no-brainer and makes for a clean eight divisions of four.

I’m not going to sit here and make the case that St. Louis or Pittsburgh or anywhere else deserves an NBA franchise. The clear top two candidates for expansion franchises are Seattle (SEA) and Las Vegas (LVG), so I’m going to use them.

Division Realignment

With an increased number of smaller divisions, some shuffling is required. These are the teams that will play each other most often, so my primary objective is to minimize distance traveled. This has the added bonus of increasing the likelihood of bordering fandoms, and nothing stokes a rivalry quite like having to see the other fandom.

I took the geographic coordinates of each team’s home stadium and calculated the distance to each other stadium. (For Seattle and Las Vegas I used their hockey stadiums.) I didn’t want to figure out the four-hundred-something different distances one at a time, so I calculated between the points using the Pythagorean theorem. This ignores the curvature of the Earth entirely and I believe is equivalent to using the equirectangular projection, which isn’t a good one but it sure does make the math easy.

For each division, I took the distance between each of the stadiums and summed all the divisions. Then I shuffled them around until I found a configuration I could not improve upon, with an average divisional distance under four hundred miles.

The Atlantic, Southeast, and Southwest divisions stay pretty much intact, just booting a member each. With two new teams out west, the Timberwolves got to join their Midwest brethren that they play in every other sport.

A couple teams ended up out of luck on being in the same division as the team closest to them. The Wizards are on the outside looking in of the Northeast Megalopolis, but there had to be an odd team out. The Jazz would’ve found the new Vegas team as the new closest team to them, but it wasn’t meant to be since Vegas is closer to LA and Phoenix.

The biggest surprise to me was the Mavericks, who ended up departing their Texan buddies. They are actually closer to Oklahoma City than either Houston or San Antonio, and the alternative was swapping it with Memphis, which would’ve been good for the Texas teams, but it would be a hardship for the teams left out, so it couldn’t be done.

Scheduling

First, let’s look at the current NBA schedule. Teams play teams in their division four times, teams in their conference three or four times, and teams from the other conference five times for a total of eighty-two games.

Shortening the season is an idea that’s gaining a lot of traction. A lot of teams are choosing to rest their players due to the sheer number of games and how that makes each individual game less consequential.

The realistic lower bound is playing each team twice, which after our expansion to thirty-two teams would be sixty-two games (since you don’t play yourself). The only thing I would add to that is divisional games, and I want six of them.

Why more divisional games?

Playing non-divisional teams twice a season is more than enough, with other sports not even playing every team every year. But upping the number of divisional games promotes something the NBA has always lacked: plentiful, durable rivalries.

The NFL plays 35% of its games against division rivals. For the MLB, it’s 47%. College sports, it’s as much as 75%. This environment creates rivalries like Yankees-Red Sox, Bears-Packers, or Michigan-Ohio State. NBA rivalries, on the other hand, are the product of circumstance, and more often than not, transient.

The biggest NBA rivalries are around teams that happened to face each other in the playoffs several times in a short amount of time. If you do that often enough, you get the Lakers and Celtics, which is very much the exception.

In recent NBA history, the biggest rivalry material was the Warriors eliminating the Rockets in the playoffs four out of five years and meeting the Cavaliers in the Finals four straight years. I can’t deny that at the time, there was a considerable amount of drama and bad blood between the teams, and it made for very compelling series. That being said, do I honestly think either of these will really be considered a rivalry in ten years’ time when all of the players have been replaced? Sadly, no.

We’re fixing it.

This brings us down to seventy-four games. In addition to doing everything possible to eliminate back-to-backs and other load management prompters, lengthening the season slightly if need be, I’d like to use that energy for something else.

Mid-Season Tournament

Adam Silver has been floating the idea of a mid-season tournament for a while now, and with good reason. It’s sick.

Having teams play a single-elimination tournament in December would bring excitement to a part of the season that is sorely lacking in it.

The NBA playoff structure, due to its seven-game series and high number of possessions per game, mean that much more often than in other sports, the best team wins. In the NHL, your goalie getting hot at the right time can make an eight-seed title contenders. Since the NBA has had a seven-game first round, just three eight-seeds have upset the one-seed. All were eliminated in the next round.

I wouldn’t call this a problem with the NBA playoffs, but it does have a very serious symptom: very few teams have a realistic shot at the title each year. And with the focus in basketball discussions being so strongly focused on rings or bust, the best the rest of the team can hope for is their promising young star takes the next step but still wants to stay. If you don’t have a promising young star, all that’s left is to hope to draft one.

The mid-season tournament would introduce a heretofore unseen level of variability in the NBA, more akin to March Madness. In the 2020 bubble, the Suns won eight straight games against playoff caliber teams and still ended up five games below .500. Even if your team can’t win a championship, you could pull off five games at just the right time.

You would probably see some true contenders taking it easy in this tournament or resting their stars since the extra games played might not be worth it to them. This would make way for more Cinderellas.

If you wanted to spice things up a little more (while also preventing the games from being too tiring for in-season competition), you could take a page from the All-Star Game and use a points-based ending. You could do away with a game clock entirely, just say first to one hundred wins. There’s no lead too large as long as you keep making defensive stops, and the end of the game is a laser-focused possession-valued slugfest instead of a free-throw frenzy. (Not that there isn’t a place for that. I could never say goodbye to buzzer beaters forever.) This would make the games a little shorter as-is and prevent triple overtime games hurting teams’ regular season.

As for motivation, I think the winner should be guaranteed a spot in the playoffs. With the way we’ve seen teams fight to make the play-in, I think you’d see middle-of-the-pack teams really make a push to win the tournament and punch their ticket.

Postseason

The last three games of the regular season should be against each of the three teams in your division. To make the division even more meaningful, the division champion gets an automatic playoff berth. And the division champion is decided not by overall record, but by division record.

In the 2016 NFL season, the Cardinals finished 7-8-1, but went 4-1-1 against the NFC West, best in their division. Under this proposal, they would clinch the division and the Seahawks would become a Wild Card team (who would have still made the playoffs).

This gives another path to fringe teams to focus extra hard on certain games, which is the name of the game. We want as many consequential games before the playoffs as possible.

After the regular season, the eight division winners and the top four wild card teams are locked in. Conferences have no bearing on who makes it at all. The next eight teams enter the play-in, same as the current format, with the first four needing to win once and the other four needing to win twice. The remaining twelve teams enter the lottery.

Immediately following the play-in, the NBA will hold a must-see TV event, containing the following.

Year-End Awards

Year-end awards are currently given out during or after the playoffs, at the time when people care about them the least. Distributing year-end awards before the playoffs allows the winners to have their moment to shine while fans haven’t 100% moved on to the playoffs yet. This should also include All-NBA, All-Defensive, and All-Rookie teams. Basically, anything from the regular season gets taken care of in one night.

Draft Lottery

The teams that do not make the play-in will enter the draft lottery. To be clear, if you make the play-in but miss out on the playoffs, you do not make the lottery.

I’m also revamping lottery odds, in a somewhat scientific way. No lottery results in an average expected pick decrease of one moving one spot in the standings. A perfectly flat lottery results in an average expected pick decrease of zero moving one spot in the standings, as long as you stay in the lottery. I’m going to split the difference and say that the expected pick change should be one half between standings positions.

There are still a lot of ways to make that lottery with expected pick increments of one half between teams, so the other guiding factor we will use is that we want the highest picks to have the flattest odds possible. This disincentivizes teams from having a worse record for better odds at the top pick.

In the table below, you can see that the top three picks all have perfectly flat odds for all teams A-L, and the odds on the other picks (in percent) is what establishes the half-pick increment. In the last column, you see the expected pick for each team given their odds.

The algorithm going on under the hood says team K should have a 9/224 chance at the fourth pick and that team H should have a 69/140 chance at the ninth pick. It’s a bit complicated, but logically sound given our priorities of pick increment and flatter odds on lower picks.

The current procedure of choosing the first couple picks and letting the rest fall where they may doesn’t work anymore. For instance, one of the last five teams needs to jump up into the top four because there’s only four other picks for the group of five to fall back on. Teams E and F have identical odds, except E has a better chance at the seventh pick.

I think the only solution is to brute force the different valid draft order permutations, weight them so that the odds are what is shown in the table, have a lottery backstage with team representatives that selects which draft order is used, then announce the draft order to the public. This is the method used currently, albeit with only one pick being chosen at a time.

The issue is that due to the number of permutations and the weights given to them, exhaustively listing the draft order arrangements with each arrangement appearing a proportional number of times would result in a list at least 157,840,048,128,000 long. Assuming each row has twelve 16-bit data points, that would result in a file size of about 3.8 petabytes. The largest database in the world is 6 petabytes. An exhaustive list is a futile approach.

Listing a proportionally weighted subset of the total possible permutations could be done in as few as 3,360 choosable arrangements. Just twenty-one of these would have team D receive the eighth pick, so there would not be a full complement of draft arrangements with that given. Conditional probability breaks down, but the overall probability can hold. This tangent has gotten way out of control, back to the NBA.

Playoff Draft

This is where things get really spicy. After the play-in games, our field of sixteen is set. They are ranked by record with no preference given to division winners. Then descending by record, each team will choose who they will face in the playoffs.

First the overall #1 will choose any of #9-#16 to face in the first round. After their selection, #2 can choose from any of the remaining seven teams. Then #3 chooses. This continues until #8 is assigned the #9-#16 team that went unpicked.

Then the #1 will choose which pair of #5-#8 and their opponent they would like to face in the second round. Then #2 will choose from the remaining three pairs, then #3 chooses, and #4 gets who’s left.

Finally, #1 chooses if it would like to face #3’s or #4’s group of four. The unchosen group is put on #2’s side of the bracket.

This prevents the scenario of teams trying to lose at the end of the season to secure certain playoff matchups and makes the race for seeds 1-8 much more consequential. In the current system, if a contender has some injuries and ends up as the #5 seed in its conference, but their team is healthy for the playoffs, the #1 would maybe rather be the #2.

Not only does this give more importance to regular season games that decide seeding, it introduces a significant level of strategy. For the first round, do you pick your division rival who you beat over and over in the regular season or the young team without any playoff experience? Do you try to influence how tough of a matchup other teams have? Do you get revenge on the team that beat you last year?

The consequence of this setup is that every underdog has the best bulletin board material in the world: the other team chose to play you because they thought you’d be easy. How much more dramatic is David vs. Goliath if Goliath chooses David and David beats him anyway?

There are a couple drawbacks to this approach. First of all, by getting rid of conference segregation, you have dramatically increased travel distances. The old home-away format of 2-3-2 helped mitigate this extra distance, but was eventually replaced with 2-2-1-1-1. Part of this was due to the fact that if the series ends in five, the favored team had fewer home games. A new system of (0)-2-3-1-1 would help both issues. The first two games are played in the underdog’s city, then the next three are played at the favorites, before alternating the last two. Assuming 50/50 odds, the old system has an average of 1.6 trips while the current system has 2.8. My proposed system has an average of 1.9 trips, but ensures that the better seed plays an equal or greater-by-one number of home games at each point the series could end. The weirdness of starting the series in the underdog’s stadium would wear off, and the momentum generated by a couple home wins could make for some very interesting series.

Another drawback is that there is virtually no difference between seeds 9-12. If #12 is perceived as stronger than #10, it is not penalized in any way for ending the season with a worse record. A counterpoint to this is that this is often a section of the standings that are extremely close, especially with no conference separation. For the 2018 playoffs, the overall #8, with home court advantage in our first round, had a record of 48-34. The #13, which would be in our play-in, had a record of 47-35. These teams can’t afford to take it easy because one game could send you tumbling down the standings. Combining the conferences means there’s double the teams to worry about taking your spot.

All in all, I’m nearly always in favor of proposals that increase the amount of strategy teams need to use. The added bonus of there being so much more intrigue and regular season games having more importance is too exciting to ignore.

Between the awards, the lottery, and the playoff draft, this would be an unmissable event for a fan of any team.

Is this all necessary?

No. Full stop.

But not all changes should be made out of necessity. Sometimes things can be done just because they’re better. I’m not saying these ideas are impeccable, but I don’t think they’re change for the sake of change either. I think concepts like “make things other than the playoffs matter” are worthy of trying to find a solution. I can only hope these ideas get the wheels turning in someone important’s head to make something cool happen.