What if you could penalize tanking, decrease randomness, reduce perverse incentives, and give teams more control over their fate? There's one big catch: you have to ditch the draft for an auction.
I wish Nate hadn't glossed over the transition and provided some ideas for how you could get there. That will be a major, major sticking point to get enough owners on board. Teams that have tanked and stock piled picks won't like losing all those assets. Teams that didn't, wont like proposals that reward said teams, when their tanking is what necessitated the need for something like this.
Same here. I think some sort of payment for existing draft picks acquired in trade would be required to get a system like this adopted. Nate’s clearly put a lot of thought into this system, it would be great if he spent some time talking about a viable transition strategy.
Promotion/relegation is largely a non-starter in American sports because none of them have a second tier of teams that aren't already owned by the top league. So what about the idea of stop looking for a second tier, and create it out of the first tier that already exists.
Expand to 32 teams, split the league in half. 16 teams in each tier, bottom 3 from Tier A get relegated and top 3 from tier B get promoted. Have an 8 team playoff in each tier. The Tier A playoff is for the league championship trophy. The Tier B playoff determines the promoted teams (Top 2 are promoted automatically, a consolation round is played for 3rd).
Regular season schedules are kept between tiers. Maybe have a mixed tier mid-season tournament, but the idea is that if owners want the high ticket prices that come from competitive Tier A games then they need to keep their team in Tier A. You can't just be a bottom feeder that still sells out games whenever the top teams come to town.
Once the promotion/relegation system is established you can keep expanding the league to eventually get to where we all wish it was, with multiple large leagues at various tiers. Every decade add 2 or 4 more teams so you can eventually expand to Tiers of 20 teams each. Let tier B grow past 20 teams, then split it again into a tier C and repeat.
I'm not sure if that helps the actual problem at all. If the idea is to maximize the number of competitive games that are interesting to watch I don't think that does this at all. If you are in Tier 2, you would be heavily incentivized to tank so that you can get better draft capital/ARC so that this player can boost you into Tier 1. If anything this increases the number of games people aren't interested in because you are created an actual line between "good" and "bad". It's certainly an interesting idea and worth juggling around, but I don't see how this in any way make the NBA a better overall product for fans.
1. Can you go through the thought process for why to cap the max bid at ⏜100?
2. If you are only allocated ⏜100 per season and use all of that on the top player, you won't have any left over. Teams will almost certainly bid the max to have a chance of the top players so it seems extremely likely the team who "wins" the auction for that player will have no other ARC to spend. This is listed as a bonus of the system "Worst teams get less expected value" but I'd argue the goal is to create more parity in the league (because that increases the number of of good games of basketball to watch) and that this will hurt their ability to "catch up" to the better teams in the league. Thoughts on that?
It would make it too easy for a team to guarantee themselves a generational talent through trading for ARC. There's a huge difference between trading your current stars for ARC so that you can have a random shot at one of the top players, vs selling the farm so you can for sure outbid the league on the single best player.
Couldn't you apply a free market principle to that argument though? Would it be easy for a team to guarantee themselves a generational talent or would it not be easy because there will be competition for that player and selling teams "will know what they have"? I understand the point you make, that the incentive then is to try to accumulate ARC (through selling the farm, or by rolling over ARC from previous years to be ready to act) but I think ultimately the value of ARC will be some what efficiently decided by the market.
I would bring back territorial picks, which the NBA had in its early history. Unlike the original territorial picks which were based typically on where a player went to college, I would do it based on their hometown (which did happen when Wilt Chamberlain was claimed by the Philadelphia Warriors). For example, the Celtics would be allowed to pick AJ Dybantsa based on that he was from Massachusetts prior to going to prep school in Utah. The Celtics would then lose a pick from the proper draft, and also nobody could tank for him since it would be known that the Celtics would just claim him.
I can see lots of NIL money going to colleges in the districts of the NBA teams which will corrupt college sports even more. Celtics would use UMass, BU, & BC as feeder colleges. Also what do you do with Los Angeles? One of the teams gets UCLA and the other USC?
I like this a lot, and have one additional suggestion. Once a team is eliminated from playoff contention, losses should count as wins and wins as losses for ARC-allocation purposes. This not only discourages tanking, it incentivizes bad teams to win. It also discourages bad teams from tanking to be eliminated from the playoffs, because once that happens it becomes harder for a sub-0.500 team to gain in the ARC tables.
However, I don't like the cliff version of that system, nor the fact that it only bottom teams (although that is where most of the tanking takes place). Far too geeky for the NBA, but fun to think about, is using "playoff equity added" for a continuous system.
Start by assigning each playoff seed an equity value--perhaps based on estimated probability of winning a championship from that seed. 9 and 10 are much less valuable than 7 and 8, and 7 and 8 less than 1-6. In most years there's not much difference between 3-6--the teams are pretty even and W/L record doesn't say much about relative quality. Moreover there are lots of reasons--fan appeal, travel, match-ups--a particular team in a particular year might prefer, say, a 5 seed to a 3 seed.
Anyway, once you have playoff equity, for each game you can estimate a playoff-equity added. If a game adds zero playoff equity, as for a team already eliminated, losses count as wins and wins losses. For small playoff equity added a loss might count as 0.8 wins/0.2 loss, and a win 0.2 win/0.8 loss, and so on up to normal playoff equity added games and above where wins are wins and losses are losses.
As a long time 538 reader, this reminded me of a very old article posted about various anti-tanking ideas suggested by readers. What you are doing is like the "Tombestone" method here.
Thanks for the link. Yes, it's a similar idea with a cooler name. In either system you have the potential for making some late season games among cellar-dwellers into exciting matches for the #1 draft pick (or the top ARC award) with the winner, not the loser, winning.
I realize that it would be a complete non-starter in any US sports league (and would require a Level II system that doesn’t really exist) but I do wonder if a relegation (and promotion) structure would at all improve the issues around perpetually bad teams. It seems the average sports fan cares more about winning than some owners who have seen the value of their franchises increase in a completely uncorrelated relationship to actual success. It seems that a lot of value here is being created by scarcity with success just being an additional premium for a few “trophy franchises.” While the NFL has media rights sharing, MLB does not, but both seem to enjoy substantial TV/streaming/media revenues across all quality of teams.
Jesus I love this post. It's like the good old days of 538. What if we replaced the entire incentive structure of the NBA with a sophisticated, EV balanced, game theory optimized meta structure? Wouldn't that be great?
I also recognize that the NBA is very unlikely to take a swing this big. Instead, I propose we all invest in a new organization called the NsBA (pronounced "Ens Bee Ay"), aka Nate's Basketball Association. It will be designed from first principles to be mathematically unexploitable. The first three teams will be The Seattle Nash Equilibria, The Vegas Kolmogorovs, and The Boston Celtics.
The Wizards will still suck under this propsal.😉 One thing that occurs to me is that there will now be more of a premium on good scouting. A team might hoard ARC to pick up 2-3 players rather than some deemed to be superstar. Lots of example of failed first round draft picks over the years representing wasted ARC under this system.
Eh, why have a max cap on ARC spending for top picks at all? Just make the teams bigger (cap of 30 players) and see what happens. Teams that are good at developing, eg OKC, will find a way to create great players by paying low ARC and “selling high”. There will be plenty of diamonds in the rough; we just don’t see that now in the NBA because the current system does not incentivize it. This would have the added benefit of helping with load management since most players would stay home for any given game.
In other words, a totally “free market” approach would yield a lot more interesting possible strategies for team development than your capped proposal. Some will go bet the farm on just a few players while others would focus on player development.
Here's a simpler solution: You lose your first-round draft choice if you don't make the playoffs. Bad teams would be motivated to recruit solid non star players in free agency to make the playoffs. Playoff teams would balance future with win now planning given the incentive that if they come up short in the playoffs, they get a really good draft pick.
I wish Nate hadn't glossed over the transition and provided some ideas for how you could get there. That will be a major, major sticking point to get enough owners on board. Teams that have tanked and stock piled picks won't like losing all those assets. Teams that didn't, wont like proposals that reward said teams, when their tanking is what necessitated the need for something like this.
Same here. I think some sort of payment for existing draft picks acquired in trade would be required to get a system like this adopted. Nate’s clearly put a lot of thought into this system, it would be great if he spent some time talking about a viable transition strategy.
I know its not realistic, but a promotion/relegation system prevents all tanking and makes games at the bottom the most competitive of all.
Promotion/relegation is largely a non-starter in American sports because none of them have a second tier of teams that aren't already owned by the top league. So what about the idea of stop looking for a second tier, and create it out of the first tier that already exists.
Expand to 32 teams, split the league in half. 16 teams in each tier, bottom 3 from Tier A get relegated and top 3 from tier B get promoted. Have an 8 team playoff in each tier. The Tier A playoff is for the league championship trophy. The Tier B playoff determines the promoted teams (Top 2 are promoted automatically, a consolation round is played for 3rd).
Regular season schedules are kept between tiers. Maybe have a mixed tier mid-season tournament, but the idea is that if owners want the high ticket prices that come from competitive Tier A games then they need to keep their team in Tier A. You can't just be a bottom feeder that still sells out games whenever the top teams come to town.
Once the promotion/relegation system is established you can keep expanding the league to eventually get to where we all wish it was, with multiple large leagues at various tiers. Every decade add 2 or 4 more teams so you can eventually expand to Tiers of 20 teams each. Let tier B grow past 20 teams, then split it again into a tier C and repeat.
I'm not sure if that helps the actual problem at all. If the idea is to maximize the number of competitive games that are interesting to watch I don't think that does this at all. If you are in Tier 2, you would be heavily incentivized to tank so that you can get better draft capital/ARC so that this player can boost you into Tier 1. If anything this increases the number of games people aren't interested in because you are created an actual line between "good" and "bad". It's certainly an interesting idea and worth juggling around, but I don't see how this in any way make the NBA a better overall product for fans.
A few questions and comments
1. Can you go through the thought process for why to cap the max bid at ⏜100?
2. If you are only allocated ⏜100 per season and use all of that on the top player, you won't have any left over. Teams will almost certainly bid the max to have a chance of the top players so it seems extremely likely the team who "wins" the auction for that player will have no other ARC to spend. This is listed as a bonus of the system "Worst teams get less expected value" but I'd argue the goal is to create more parity in the league (because that increases the number of of good games of basketball to watch) and that this will hurt their ability to "catch up" to the better teams in the league. Thoughts on that?
It would make it too easy for a team to guarantee themselves a generational talent through trading for ARC. There's a huge difference between trading your current stars for ARC so that you can have a random shot at one of the top players, vs selling the farm so you can for sure outbid the league on the single best player.
Couldn't you apply a free market principle to that argument though? Would it be easy for a team to guarantee themselves a generational talent or would it not be easy because there will be competition for that player and selling teams "will know what they have"? I understand the point you make, that the incentive then is to try to accumulate ARC (through selling the farm, or by rolling over ARC from previous years to be ready to act) but I think ultimately the value of ARC will be some what efficiently decided by the market.
Not sure an "auction" would go over very well in a mostly black league with mostly white owners and execs...
I would bring back territorial picks, which the NBA had in its early history. Unlike the original territorial picks which were based typically on where a player went to college, I would do it based on their hometown (which did happen when Wilt Chamberlain was claimed by the Philadelphia Warriors). For example, the Celtics would be allowed to pick AJ Dybantsa based on that he was from Massachusetts prior to going to prep school in Utah. The Celtics would then lose a pick from the proper draft, and also nobody could tank for him since it would be known that the Celtics would just claim him.
I can see lots of NIL money going to colleges in the districts of the NBA teams which will corrupt college sports even more. Celtics would use UMass, BU, & BC as feeder colleges. Also what do you do with Los Angeles? One of the teams gets UCLA and the other USC?
I like this a lot, and have one additional suggestion. Once a team is eliminated from playoff contention, losses should count as wins and wins as losses for ARC-allocation purposes. This not only discourages tanking, it incentivizes bad teams to win. It also discourages bad teams from tanking to be eliminated from the playoffs, because once that happens it becomes harder for a sub-0.500 team to gain in the ARC tables.
However, I don't like the cliff version of that system, nor the fact that it only bottom teams (although that is where most of the tanking takes place). Far too geeky for the NBA, but fun to think about, is using "playoff equity added" for a continuous system.
Start by assigning each playoff seed an equity value--perhaps based on estimated probability of winning a championship from that seed. 9 and 10 are much less valuable than 7 and 8, and 7 and 8 less than 1-6. In most years there's not much difference between 3-6--the teams are pretty even and W/L record doesn't say much about relative quality. Moreover there are lots of reasons--fan appeal, travel, match-ups--a particular team in a particular year might prefer, say, a 5 seed to a 3 seed.
Anyway, once you have playoff equity, for each game you can estimate a playoff-equity added. If a game adds zero playoff equity, as for a team already eliminated, losses count as wins and wins losses. For small playoff equity added a loss might count as 0.8 wins/0.2 loss, and a win 0.2 win/0.8 loss, and so on up to normal playoff equity added games and above where wins are wins and losses are losses.
As a long time 538 reader, this reminded me of a very old article posted about various anti-tanking ideas suggested by readers. What you are doing is like the "Tombestone" method here.
https://fivethirtyeight.com/features/how-to-stop-nba-tanking-tie-your-fate-to-another-teams-record/
Thanks for the link. Yes, it's a similar idea with a cooler name. In either system you have the potential for making some late season games among cellar-dwellers into exciting matches for the #1 draft pick (or the top ARC award) with the winner, not the loser, winning.
I realize that it would be a complete non-starter in any US sports league (and would require a Level II system that doesn’t really exist) but I do wonder if a relegation (and promotion) structure would at all improve the issues around perpetually bad teams. It seems the average sports fan cares more about winning than some owners who have seen the value of their franchises increase in a completely uncorrelated relationship to actual success. It seems that a lot of value here is being created by scarcity with success just being an additional premium for a few “trophy franchises.” While the NFL has media rights sharing, MLB does not, but both seem to enjoy substantial TV/streaming/media revenues across all quality of teams.
Jesus I love this post. It's like the good old days of 538. What if we replaced the entire incentive structure of the NBA with a sophisticated, EV balanced, game theory optimized meta structure? Wouldn't that be great?
I also recognize that the NBA is very unlikely to take a swing this big. Instead, I propose we all invest in a new organization called the NsBA (pronounced "Ens Bee Ay"), aka Nate's Basketball Association. It will be designed from first principles to be mathematically unexploitable. The first three teams will be The Seattle Nash Equilibria, The Vegas Kolmogorovs, and The Boston Celtics.
Why not use a 2nd price auction if you're doing a secret simultaneous auction?
Or just do a live auction for each player, would definitely make the draft more exciting.
Yes, this is what I came here to say. Doing a sealed simultaneous bid auction that's not a second price auction seems perverse.
The Wizards will still suck under this propsal.😉 One thing that occurs to me is that there will now be more of a premium on good scouting. A team might hoard ARC to pick up 2-3 players rather than some deemed to be superstar. Lots of example of failed first round draft picks over the years representing wasted ARC under this system.
1 / Get rid of the draft 2 / Hard Cap 3 / Allow any player to declare professional eligibility and sign a contract based on the market
Eh, why have a max cap on ARC spending for top picks at all? Just make the teams bigger (cap of 30 players) and see what happens. Teams that are good at developing, eg OKC, will find a way to create great players by paying low ARC and “selling high”. There will be plenty of diamonds in the rough; we just don’t see that now in the NBA because the current system does not incentivize it. This would have the added benefit of helping with load management since most players would stay home for any given game.
In other words, a totally “free market” approach would yield a lot more interesting possible strategies for team development than your capped proposal. Some will go bet the farm on just a few players while others would focus on player development.
Here's a simpler solution: You lose your first-round draft choice if you don't make the playoffs. Bad teams would be motivated to recruit solid non star players in free agency to make the playoffs. Playoff teams would balance future with win now planning given the incentive that if they come up short in the playoffs, they get a really good draft pick.