The Costs, Benefits, and Purpose of Expanding the NBA Draft

Published: September 11, 2011

I initially dismissed a claim in an article that the NBA owners wanted a third round to the NBA draft. I didn’t dismiss the reporter’s claim; I dismissed the owners’ claim that they wanted it. It seemed to be merely a proposal intended to waste the other side’s time by refuting it. This, I’m told, is a tactic employed in some legal battles.

After thinking, it occurred to me that they could be serious.

After thinking more, I know they are serious.

The idea that they want the draft expanded to provide access to talent is insane. For one, they don’t need a draft to give them the ability to have the best undrafted players on their team . . . they can just sign the guys outright. Also, while there will certainly be relative gems in a third round of the draft, the overall quality of the players will not exceed that of the second round, just as the quality of the second does not exceed that of the first overall. There are notable individual exceptions to this, and the New Orleans Hornets have been beneficiaries of such an effect recently, yeah? If the talent in the third round were so, that would imply the scouts and GM are somehow bungling enough to miss on the first swing after the first round, but would be suddenly, as a whole, brighter given an extra hour or two on that night and that night only. I’m not buying that.

To wrap up the talent portion of the article, let’s take a look at the latest of Ryan’s Value of a Draft Pick series of articles. Given the dropoff in talent across the first round and the relatively flat performance across a large swath of the second round, it’s safe . . . maybe a little generous . . . to assume a third round of the draft will be essentially the same as the last portion of the second round in terms of quality. Thus, about 12 of the players in the third round would never see a minute of NBA play, while 4 of them would be a 6th man type or better. The other 14 players would be bench players of one sort or another.

But, again, they don’t need the draft to get these players. In fact, any owner who thought they could get said gems would be foolish to ask for the third round, as other owners may lock up their targeted players by blind luck. On the flip side, any owner who thought there was nothing in the third round wouldn’t ask for a third round. So if no one owner would want a third round, why would all of them want to expand it, or a least the owners as a whole?

Oh, the beauty. It’s a simply magnificent little puzzle.

Effects of the Draft

The draft’s main purpose is to distribute new talent throughout the NBA. One effect of the draft is that, and it seems to be an effective tool to achieve that goal.

Another effect, however, is to create 60 new contracts each year. This is not strictly true, of course . . . see Ricky Rubio. Likewise, contracts are generated for not-currently-in-the-NBA-players each year by means other than the draft. Closing the loop, if all 60 picks sign this year, Rubio finally joining the NBA means 61 were generated this year. So given all the ebb and flow, let’s just assume that 60 new contracts each year are generated by the draft, then check later to see how important it is for the number to be 60 on the nose.

Adding a third round to the draft does further the main goal: adding 4 good players and 14 players that sit squarely at some point on the bench. It will also burden the NBA with 12 players that will never see a minute of NBA playing time. The draft being extended is still more effective than it is ineffective . . . 60% effective, 40% ineffective . . . or 50% more effective than ineffective. For comparison, the second round is about 66% effective and 34% ineffective.

Looking at these numbers, even if we were being generous above in our estimate of the quality of the third round, I’d assume they would go with it until it was 50% effective, 50% ineffective. That may really be the case in a third round, but I can’t see the next best 30 players being significantly worse than the 23 players picked before them. The NBA picks up players from places other than the draft at about that clip each year. Also, basketball leagues all over world pick up these guys . . . they aren’t slouches.

I, think, however, it is the second purpose that may be the real prize for the owners. Bear with me . . .


We have 30 NBA teams currently, and let’s assume a team, on average, fields 14 players. The max is 15, so we leave some breathing room. Let us also assume, on average, that each team has to replace one player’s talent each season with a player not currently in the NBA. Factoring in the draft, we have about 90 new players each year out of about 420. This corresponds to a turnover rate of 90/420 ~ 0.21. This translates into an average career length of 420/90 ~ 4.7 seasons, which is very close to the typical numbers seen for average career length each season.

This proximity is strong evidence, though not proof, that this method is sound. The proof . . . is hard. It involves approximations, some basic assumptions, and mathematical objects called statistical processes. In the end, this works and it’s basically a `structural’ kind of number. The overall distribution of the career lengths depends on more than we have assumed . . . whether it’s tight, is heavy-tailed . . . whatever . . . but this number is important and will likely not vary until the factors described above have changed.

It should be noted that the computation does not actually depend on the number of teams . . . we could have used (2+1)/14 ~ 0.21. We only used the bulk numbers to set it up more clearly rather than using the less obvious, more efficient, formula involving ‘rates’.

So, contraction or expansion of the NBA does not affect this average career length so long as the draft varies in step with the NBA and replacement rate and roster sizes are unchanged. The general formula is thus: (Number of Draft Rounds + Per Team Replacement Rate) / (Average Roster Size).

Adjusting the same analysis to account for a third round for a draft, the turnover rate is (3+1)/14 ~ 0.29. This translates into an average career length of 14/(3+1) ~ 3.5 years.

Messing with the nits in the draft rate discussed above mean changing the number of rounds to some decimal very near an integer will change the end numbers slightly, but the song reamins the same: The 3rd round of the draft reduces the average NBA career by about a season, or 20% – 25%.

Cost, Benefits, Purpose

Thus, by adding in a third round to the draft the following occurs:

1) 1 – 4 sixth man players or better are added to the NBA talent pool by force and with small contracts
2) 11 – 14 bench players are added to the NBA talent pool by force and with small contracts
3) 12 – 15 players who will never play a minute are added to the NBA talent pool by force and with small contracts
4) The average NBA player’s career drops by from about 4.7 years to about 3.5 years.

The ranges were generated by projecting a 50% ineffective draft in the following way: Make half the picks ineffective by pulling players from `bench level’, the fill in the `bench level’ as close as possible to its previous count with players from the `sixth man’ level.

As mentioned above, Effect 1 and Effect 2 are available to the owners already, but the decision is not forced. If the scouts are good, these players would have been drafted in the first round . . . but they weren’t, so the scouts aren’t 100% efficient in their selections, which is not surprising in the least. In this sense, maybe adding the draft round is good for the NBA in terms of Effect 1 and Effect 2, though it would hurt teams with great scouts, effective player development programs, and a clear picture of their needs. Maybe this can be viewed as the talent development analog of revenue sharing. Assuming each player costs $0.8m on average over a 2 year contract, and this may be inflated even under the old CBA, this cost is about $12m – $15m.

Effect 3 is pure waste. It’s the cost of doing business. This will cost the NBA owners about $10m – $12m if they insist on expanding the draft.

Thus, the cost of the 15 – 18 `playable’ players is $25m per year to the owners, or about $1.4m – $1.7m per player.

These 15 – 18 players are each a dagger into the hearts of the NBA’s affluent middle class, which is where the real player cost growth has been. Even after these players outgrow their rookie contracts, their contracts may out-bang-for-the-buck a significantly higher salary for a not-as-significantly better player. This will result in less salary being used for fulfilling the need the higher dollar player was being considered for filling, whether it be by the now more available young talent or the vet.

The players described are just like some already in the NBA, but there is currently an availability issue. Making these kinds of players more available at contract time induces the economic effects. The young, cheap bench player signed on another team’s roster has no effect of the competition for a spot on this team’s roster.

Player costs rising due to increases in basketball related income is not an issue. That is a good problem to have. The issue is that money spent on tier 2 and tier 3 players has risen dramatically while their talent hasn’t increased as dramatically.

Another way to state this is that cost effectiveness has decreased. Dollars spent on ‘real’ max players and high draft picks are spent most efficiently. Insane contracts (e.g. Arenas) are beyond this and any rational discussion; there’s no explaining crazy. Maximizing talent on the floor, however is not about being cost effective, and this is where the inefficient expenditures stem from. By getting more talent in the pipeline sooner, more cost-effective rosters can be fielded.

Also, by lowering the average career length, Effect 4, the owners as a whole are lowering what they pay for talent as fewer players will reach 5 years, 7 years, 10 years, etc. than now. As a result, fewer raises are given and compounded. Thus, we have cheap talent coming from the draft that can be used in lieu of some, not all, vet talent and, in the end, fewer players reach the really high pay NBA pay scales, using the old CBA as a gauge. The players that reach these career lengths would have competed for their jobs more severely and should also be more ‘deserving’ of the larger salary.

All totaled, this seems like a good idea from the owners point of view. On the player side, the superstars won’t be affected by this in the slightest, nor will the one-contract-and-done types. If the underachieving segments of the NBA’s affluent middle class . . . think Josh Smith more than Al Horford . . . think Morris Peterson and James Posey, Hornets fans . . . don’t pick up on this or have the stroke and numbers in the NBPA . . . or read this article (everyone needs a mathematician) . . . then they just might, as a group, get hammered come pay day.

This is why the owners want to expand the draft.

This seemingly modest proposal just may be what saves the season if the owners can get revenue sharing situated. It will certainly need tons of help, but it could be the extra bit the owners need. Of course, this article may have let the cat out of the bag and ruined it all.



  1. Pingback: The NBA's Labor Revolutions | New Orleans Hornets |

Leave a Reply

Your email address will not be published.