NBA Draft Lottery Odds and How the Drawing Works
The first thing I did after installing last year's 2K was start a MyNBA with the worst roster in the league and simulate to May, because I wanted to see the lottery cutscene. It took nine seconds.

The first thing I did after installing last year's 2K was start a MyNBA with the worst roster in the league and simulate to May, because I wanted to see the lottery cutscene. It took nine seconds. Logos slid into a board, my team landed fourth, and the game moved on to the combine. I have done that maybe forty times across various franchise modes and it has never once felt like the thing it is modelling.
The real version takes about ten minutes, happens in a sealed room with no cameras, and nobody in the building watching the broadcast sees a single ball drawn. What they see is a man in a suit turning over envelopes in reverse order, which is theatre built on top of a drawing that already finished. The mechanics underneath are genuinely strange, and most fans who argue about tanking every February have never looked at them.
How the drawing actually works
Fourteen ping pong balls, numbered 1 to 14, in a drum. Four balls get drawn, and the order they come out does not matter, so a draw of 3, 7, 11, 2 is the same result as 2, 3, 7, 11. The number of unique four ball combinations from fourteen balls is 1,001.
One thousand of those combinations are assigned to the fourteen non playoff teams in proportion to their odds. The one left over, 11 through 14, is the redraw: if it comes up, the balls go back and the drawing happens again. The drum gets mixed for 20 seconds before the first ball and 10 seconds between each one after that, and the whole thing is witnessed by team representatives and an accounting firm in a room where phones are collected.
That drawing is repeated four times, once for each of the top four picks, with the combinations belonging to an already drawn team set aside. Picks 5 through 14 are then filled in by inverse record, with coin flips having already settled any ties in the standings. So the lottery decides four slots, and the other ten were decided in April.
The reveal order on television is deliberately backwards. The commissioner's deputy starts at pick 14 and works down, which is why the tension in the room belongs to teams who already know they did not win anything.
The odds table, and the one line that matters
Since 2019 the three worst records share identical odds. Before that, the worst team had 25 percent at the top pick and the second worst had 19.9. Here is what the board looks like now.
| Lottery seed | Chance at pick 1 | Chance at a top four pick | Worst case |
|---|---|---|---|
| 1st, 2nd, 3rd worst | 14.0 percent | 52.1 percent | Pick 5 |
| 4th | 12.5 percent | 48.1 percent | Pick 8 |
| 5th | 10.5 percent | 42.1 percent | Pick 9 |
| 6th | 9.0 percent | 37.2 percent | Pick 10 |
| 7th | 7.5 percent | 31.9 percent | Pick 11 |
| 8th | 6.0 percent | 26.3 percent | Pick 12 |
| 9th | 4.5 percent | 20.3 percent | Pick 13 |
| 10th | 3.0 percent | 13.9 percent | Pick 14 |
| 11th | 2.0 percent | 9.4 percent | Pick 15 |
| 12th | 1.5 percent | 7.1 percent | Pick 16 |
| 13th | 1.0 percent | 4.8 percent | Pick 17 |
| 14th | 0.5 percent | 2.4 percent | Pick 18 |
Look at the last column rather than the first. A team with the worst record in basketball has a 48 percent chance of leaving the lottery with the fifth pick, and the fifth pick is where a rebuild goes to get complicated. Under the old system that same team could fall no further than fourth. The flattening lowered the ceiling and also lowered the floor, which is the part nobody mentions when they describe the change as anti tanking.
Why the flattened odds did not kill tanking
The standard line, repeated every spring since the rule passed in 2017, is that making the bottom three identical removed the incentive to lose. I think that gets the mechanism backwards, and two seasons of standings since then have not changed my mind.
If three teams share the same odds, then once you are in that group there is no benefit to being worse, which is true. The problem is getting into the group. A team sitting eleventh worst has 2 percent at the top pick and 9.4 at a top four selection. Falling three spots in the standings takes that to 7.5 and 31.9. That is a tripling of top four probability available to any front office willing to shut a veteran down in March, and the cost of acquiring it is a few wins nobody was paying to see anyway.
What actually changed behaviour was a different rule entirely. The play-in gave teams in the 7 to 10 range a live reason to finish the season trying, which is a much larger group than the three teams at the bottom, and the effect on late March rosters was visible within a year. We went through that in the piece on how the play-in changed tanking, and the short version is that the carrot moved more teams than the stick did.
The other thing front offices do now, which the odds change arguably encouraged, is collect other teams' unprotected picks instead of damaging their own season. One bad team is 14 percent. Three separate lottery picks from three separate teams is a portfolio, and portfolios do not require you to lose 60 games in front of your own fans.
What the games get wrong
I have run the lottery in MyNBA, in franchise modes going back to the 2K14 era, and in a couple of basketball management sims, and the odds themselves are usually right. 2K uses the current table and the editor even lets you change it, which I have done out of curiosity more than once. What the games miss is everything around the drawing.
No game I have played models the ten week stretch before it. In the real league, April is a negotiation between a coach who wants to win games, a medical staff producing load management decisions, and a front office reading the standings. In a franchise mode you set a tanking slider and the AI handles it invisibly. There is also no consequence for being obvious, where the real league has fined teams for resting healthy players and has a 65 game threshold that pulls in the other direction, which we covered in the piece on the NBA 65 game rule.
The second miss is the trade market around the pick. In a game, pick 5 and pick 1 have values from a static table. In the real league the value of a pick depends on who is at the top of that specific draft class, and a fifth pick in a class with two obvious stars is worth less than a seventh in a class with eight. General managers spend February asking scouts how deep the class is, and the answer changes what they are willing to do with the rest of the roster.
One thing to watch at the next lottery
When the broadcast starts, ignore the top of the board and watch which team is standing in the fifth slot. A franchise that finished with the worst or second worst record and lands at five has just lost the player it spent a season acquiring, and you can read the next eighteen months of that organisation off that one envelope: a veteran gets traded, a coach gets a one year contract, and the payroll decisions start running into the restrictions described in our breakdown of the NBA second apron.
And if you want to feel the odds rather than read them, open a franchise mode, take the worst team in the league, and simulate the lottery twenty times in a row. You will land outside the top four about half the time. Twenty seconds of that does more for your intuition than any table, including mine.
Frequently asked questions
Can the team with the worst record pick later than fifth?
No. Because only four slots are drawn, the three worst records are guaranteed a selection no later than pick 5. The rest of the order follows inverse record, so the fourth worst can fall as far as eight and the fourteenth seed as far as eighteen.
Why are there 1,001 combinations and not 1,000?
Because 1,001 is how many distinct four ball groups come out of fourteen numbered balls. The league assigns 1,000 of them to teams and treats the remaining group, 11 through 14, as a redraw so that every assigned combination stays exactly one thousandth of the pool.
Does the lottery happen live on television?
The drawing does not. It takes place beforehand in a separate room with team representatives, league officials and auditors present, and the televised portion is the reveal of a result that already exists. The envelopes on stage are sealed before the cameras turn on.
Do traded picks keep the original team's odds?
Yes. The odds belong to the record, so a pick acquired from a bad team carries that team's lottery chances, and the acquiring club sends a representative on its behalf. Protections written into the trade then decide whether the pick actually conveys, which is where most draft night confusion comes from.
More in NBA
NBA
NBA Pick and Roll Defense Explained Without the Jargon
Watch any fourth quarter and you will see the same thing happen nine or ten times.
NBA
NBA 65 Game Rule Explained and How It Changed Awards
In March 2024 I watched a broadcast spend four minutes of a blowout explaining that the reigning MVP could not win the award...
NBA
NBA Restricted Free Agency and How Offer Sheets Work
Every July, a young player signs a big offer sheet with another team, the fan base celebrates, and two days later he is back...