With the NBA Draft a day away, new research from BYU stats professor Jared Fisher offers front offices a smarter way to predict how it’ll all play out.
With the NBA Draft just a day away, all eyes are on where BYU stars AJ Dybantsa and Richie Saunders will hear their names called. Most mock drafts project Dybantsa as the No. 1 overall pick, while Saunders is widely expected to come off the board early in the second round.
But how much confidence should fans or NBA front offices place in those predictions? New research from BYU statistics professor Jared Fisher suggests there may be a smarter way to forecast how draft night will unfold.
Fisher, working alongside former Sacramento Kings analyst, BYU alum and Sage Sports Group co-founder Colin Montague, took a hard look at the current state of NBA mock drafts. Their research, recently published in the Journal of Quantitative Analysis in Sports, laid out a new system for measuring both the accuracy of mock drafts and combining them in a way that gives teams a more reliable forecast.
“Teams that can better forecast the draft order gain a competitive advantage,” Fisher said. “Mock drafts are a valuable but underutilized resource. We wanted to find a way to harness that expert knowledge more effectively.”
Montague and the Kings built what they believe is the largest database of NBA mock drafts, analyzing over 1,700 mocks from more than 100 different authors, between 2009 and 2021. Their goal wasn’t to say who’s right or wrong, but to figure out how to harness the collective wisdom of draft experts more effectively.
Most aggregations of mock drafts rely on something called the Borda count method: awarding players points based on where they show up in each mock draft, and the totals are used to rank them. It’s straightforward, but not perfect. If one mock is way off, it can throw the whole average.
Fisher and Montague proposed something different. A method based on ranked-choice voting—like what some cities now use in local elections. Their approach, which they call Ranked-Choice Aggregation (RCA), simulates a series of votes to determine the most agreed-upon pick at each draft spot. The process filters out outliers and zeroes in on true consensus.
And the results were impressive.
In head-to-head comparisons, RCA consistently beat the Borda method, especially in the final stretch before the draft, when accuracy is crucial. The study found that RCA-generated mocks landed in the 75th percentile for accuracy across all mock drafts on average, compared to 67th for Borda.
The study also proposed a better way to measure mock draft accuracy. Coming from the computer science literature on ranked lists, Rank-Biased Distance (RBD), is a metric that accounts for the reality that missing on the No. 1 pick by five spots is a bigger deal than being off by five spots at No. 55. It also handles incomplete mocks (not every list goes to 60) and can deal with players who don’t show up on every mock.
Using RBD, the researchers found that RCA-based mocks routinely matched or beat even the top analysts, including well-known names like ESPN’s Jonathan Givony and Bleacher Report’s Jonathan Wasserman.
And while no model can predict a bombshell pick like Anthony Bennett at No. 1 in 2013—or the surprise rise of Patrick Williams in 2020—RCA helps smooth out the chaos that comes with draft night.
"With RCA, every mock draft still gets a vote," Fisher explained. "But you're not giving full points to an outlier just because one person ranked them first. It’s like if you're on a budget buying groceries and one kid really wants bacon; they might love it, but it doesn’t mean it should be high on the shopping list."
The idea for the research took shape when Fisher reconnected with Montague, a fellow BYU alum who was then a member of the Sacramento Kings’ analytics team and now serves as co-founder of Sage Sports Group, a sports ownership and analytics venture. Montague was looking to explore how teams might better use external rankings and mocks. "We were interested in how we can take mock drafts and big boards and aggregate them in a way that’s more informative," Fisher said.
What started as a conversation turned into multiple research projects, involving two BYU undergraduate students along the way so far. "That’s one of the things I love about BYU," Fisher said. "Getting students involved in work that has real-world applications." Brady Heinig (BS 2024) and Brian Taylor (current student) have been gathering new mock draft data beyond the original study and are working with Fisher to build a new method that not only aggregates mock draft information like RCA, but also accounts for team preferences and estimates probabilities of each player and draft position instead of just a predicted order.
Beyond basketball, the model has wide potential. Ranked-choice aggregation and rank-biased distance could be useful in any scenario where experts offer ranked opinions—things like job candidate shortlists, music or movie rankings, or product reviews.
“As the world gets more input-driven and decisions involve more voices, you’re going to see more uses for this kind of ranked-choice system,” Fisher says. “Especially where you’re filling ranked positions, not just a winner-takes-all.”
But for now, the focus is on hoops. With this year’s draft just around the corner, Fisher is eager to see how RCA performs in real time. As of June 15, both RCA and Borda count, as well as most mock drafts, point to BYU’s AJ Dybantsa being the #1 pick in the draft. BYU’s Richie Saunders looks to be drafted around 39th or 40th according to Borda and RCA respectively, with most mock drafts placing him early in the second round.
If it proves to be as accurate as the research suggests, RCA could become a more utilized tool for teams trying to gain an edge on one of the biggest nights of the basketball calendar. And in a competitive league like the NBA, where one draft pick can change a teams’ trajectory, accurate forecasting might just be the difference between a playoff push and a full-blown rebuild.
