The NBA draft involves franchises projecting the probabilities behind basketball prospects. How efficient are the experts? When do they struggle? What can bettors learn about predicting outcomes against those with superior access to information? Read on to find out.
On October 16, 2020, the general managers of the 30 biggest teams in basketball will come together to place bets.
These are bets on the future career arcs of basketball players, as opposed to traditional sports bets on the 1x2 or Handicap markets. However, in reality the processes involved are not hugely different.
Like in sports betting, selecting a player in the NBA draft comes down to probabilities. With millions of dollars on the line, assessing those probabilities correctly can be the difference between drafting Jonny Flynn or Steph Curry.
This is perhaps what makes the NBA draft such a tantalising prospect, but what can bettors learn from the event?
High stakes and expert insight
NBA franchises have teams of analysts dedicated to assessing the ability and potential of the young players available for selection. They are making decisions worth multiple millions of dollars in relatively short time spans.
To do this effectively, they have access to a wide array of information and expertise which well surpasses that of the average basketball fan. Does this sound familiar?
In some ways, the franchises are in the same position a bookmaker is in. Would the average basketball fan be able to predict the trajectory of draft picks better than a team of professional analysts? What about a budding basketball expert armed with a spreadsheet?
In all likelihood, no. Just as the odds are stacked against bettors in the betting market, even the most ardent basketball fan should be beaten out by professionals.
How good are the experts?
The data tells us that the position a player is taken in the draft is a reasonable predictor of his value as a player (calculated by overall wins added):
The correlation between draft position and wins added is noisy, partly because of the small sample caused by limited drafts (Marc Gasol is solely responsible for the 48th pick spike). But overall you can see that teams predominantly rank the players correctly.
However, this correlation is a far cry from the efficiency of the betting market. Sample size will be a big factor behind this, and we would expect the correlation to become clearer over many drafts, but it is doubtful that the selection process could ever become as efficient as betting odds. Why is this?
What makes drafting so difficult?
NBA franchises invest millions in the draft process with scouts and analysts tasked with selecting the optimal player. What makes drafting such a difficult prospect?
A large number of intangible variables
The variables at play when drafting an NBA prospect are vast and often difficult to quantify. What is a player’s mentality like? Is his body type prone to injury? How does he adapt to a new city and lifestyle? That is all before you even look at his basketball ability.
The more variables there are to assess and the more abstract they get, the harder it is to accurately boil them down to a probability.
Long time period
As shown in the book Superforecasting, it becomes more difficult to project a probability accurately for an event the further it is taking place in the future. For NBA draftees, their first contract is two to three years in length and then they would hope to have future years in the league (for many this is not the case though).
Predicting how a player progresses over multiple seasons is more difficult than calculating the probability behind a team winning a one-off game. Not only is the number of variables greater, but they are also harder to quantify over a longer time period.
Incentives and bias
Leslie Alexander of the Houston Rockets stated that the decision making of his franchise prior to hiring Daryl Morey “wasn’t that good”. Leslie said that the Rockets had access to all this data and technology, but were not using it in a progressive way.
Morey himself speaks of the bias in the process at that time with far too much hinging upon face-to-face interviews and gut feeling, as opposed to objective data points.
The stakeholders within NBA franchises are incentivised to maintain the status quo because it helps keep their jobs. People are resistant to change.
This is in sharp contrast to the betting world, where refusing to adapt to better processes will lead to heavy financial losses for a bettor or bookmaker. This probably accounts for a fair amount of the difference in efficiency between the two areas.
Predicting the unpredictable: LaMelo and Zion
Among this year’s top draft prospects is a player who demonstrates why the analysts of the NBA often have their work cut out.
LaMelo Ball has had a very untraditional path to the draft. He is exceptionally talented as a passer and has the potential to control games at a high level, but he is seen by some as a defensive liability with poor shot mechanics.
There is such limited data on LaMelo to compare with his peers that it is very difficult for analysts to come to an agreement over his probable success as an NBA player. One scout described the potential top three pick as “a very dangerous choice”, whereas others feel he has by far the biggest upside of any potential pick.
Therefore, how LaMelo should be valued by a team is highly subjective. This is in contrast to last year’s number one pick Zion Williamson.
Zion had a much more conventional path into the draft with metrics that could be held up to previous draftees and used to see what kind of pathway he could take in the league.
Pinnacle do not limit winning bettorsLatest odds
The only players other than Zion to collect 500 points, 50 steals and 50 blocks in their freshman year were Kevin Durant and Anthony Davis, who have 17 All-star game appearances between them.
Being in that company narrows down Zion’s range of career outcomes significantly. Analysts generally agreed that his likely floor level was that of an All-star reserve and his ceiling that of an All-star starter.
Compare this to Ball, who could be projected as anywhere between a role player and a genuine MVP candidate, and you can see why franchises love the prospect of an easily assessable star like Zion.
LaMelo brings with him flawed mechanics, no college stats and an attention-loving father, but also a productive spell playing for the worst team in the NBL and an arguably top level basketball IQ. This is a much more difficult equation to solve than college superstar Zion presented.
Bookmakers and NBA analysts
Zion has a small range of likely outcomes and it’s easier to put a probability on his success in the league.
The same reason NBA franchises loved him is the reason why bookmakers prefer to take bets on a more Zion type probability range. When a Premier League game is played they know what the true probabilities making it an easier task to set a margin and profit.
On the other hand, LaMelo style bets are a bit trickier as the possible outcomes are so difficult to define. This may be part of the reason why we see the existence of a favourite-longshot bias.
A larger margin on a more unpredictable market or selection could go some way towards protecting the bookmaker from a miscalculation. This is a luxury NBA franchises do not have with LaMelo Ball.
What can bettors learn from this?
If bookmakers were drawing up odds on NBA players’ careers before a draft, what would they like to see?
While bookmakers would probably prefer the more defined outcomes offered by someone like Zion Williamson, bettors who are in competition with those bookmakers may want things to be a little more difficult to predict.
The wider the range of variables and the larger their effect on the realistic outcomes, the better it is for the bettor, as there is more opportunity to find an overlooked edge.
The more variable thrown into the mix the more likely the betting market, and by extension the bookmaker, is to miscalculate a probability.
However, bookmakers can attempt to combat this with larger margins on such markets.
Lessons from the NBA Draft
When approaching a bet it’s good to think about what new information you have, and how this helps to define probabilities better than the experts. What do you know that nobody else does?
If we can learn anything from the NBA draft, it’s that even the experts struggle with making predictions a distance into the future. Perhaps this suggests looking for longer term bets (such as outright bets) could be an avenue to explore.
Equally, experts seem to struggle when lots of difficult to quantify variables are introduced into a scenario. Is there an edge on mentality, desire or team cohesion you may have over the market? For most that is unlikely but it’s something worth thinking about.