Nektar Insights - Data & Sports

The worlds of data science and sports naturally converge in fascinating ways. As an enthusiast in both domains, I've spent years exploring this intersection through a personal blog dedicated to sports analytics. While that project has concluded, my passion for applying data science techniques to sports remains strong.

The following analyses serve as both creative outlets and opportunities to enhance my technical capabilities, all driven by genuine curiosity rather than professional obligation. Explore below to see how this hobby continues to yield interesting insights.

April 2025

The Lie: One Minute is Not a Drought

During NCAA Men's tournament broadcasts, I've noticed partners obsessively display "scoring drought" counters (I've tracked this since 2022). I find this emphasis misplaced because these droughts are normal occurrences in college basketball. My hypothesis is that these "scoring droughts" misrepresent normal statistical variation as extraordinary events. In other words, I would see a running timer come up at 1:00 minute of play and it just seem ridiculous.

To clarify, I am not talking at scoring runs, only droughts in terms of time (interesting story here about scoring runs).

My critique centers on two methodological flaws:

Players Ability: College basketball's skill distribution naturally produces more variable scoring patterns than the NBA.
Measurement error: Current drought calculations include opponent possession time. For example, if Team A uses their full 30-second shot clock without scoring, Team B scores in 15 seconds, and Team A misses again after 15 seconds—all 60 seconds count toward Team A's "drought," despite Team B actively possessing the ball for part of that time.

Given these flaws in how scoring droughts are measured and highlighted during broadcasts, I decided to apply a more rigorous analytical approach to determine what truly constitutes a statistically significant drought in college basketball.

To tackle this question properly, I needed to define what constitutes a genuine drought. I settled on 60 seconds as my minimum threshold, ensuring at least two possessions for a team (reasonable considering KenPom data shows average college possessions last 20-21.5 seconds).

With this more statistically sound definition established, I analyzed scoring patterns across multiple seasons (2002 to 2023) of NCAA men's basketball to reveal what the data actually tells us about scoring droughts.

The Truth: Wake Me Up at 2:00!

My analysis revealed a left-skewed distribution (see plot below); most scoring droughts last under three minutes, with an average of approximately two minutes. This suggests what broadcasters dramatically highlight as "droughts" are actually standard gameplay patterns.

Looking at this from a percentage basis (as shown in the following plot below), the data reveals that approximately 70% of all scoring droughts last two minutes or less. This reinforces our earlier point—since we established 60 seconds as the minimum threshold for what should even be considered a drought, these common occurrences under two minutes represent normal gameplay patterns, not exceptional events worthy of special broadcast attention.

Let's turn our attention now to the abnormal, the atypical, diabolical games where the hoop has a steel cage around it and the players are totally inept.

From the data, I pulled seven or so game trend plots to validate my findings (you can cycle through the below image carousel for a good chuckle). For all but two of the games, a team didn't score more than 11 points in the first half! A top 10 list of the longest scoring droughts.

So What?

This analysis on scoring droughts reveals how data analytics can transform our understanding of what's truly significant in sports. You can ignore the scoring drought until it hits 2:00 or get mad with me!

What broadcasters dramatically highlight as exceptional events often represent normal statistical variation—the natural ebb and flow of college basketball. By establishing data-driven benchmarks and examining the outliers, we gain a more nuanced perspective on gameplay patterns.

This approach demonstrates the power of analytics beyond traditional applications... cutting through narrative bias to reveal objective insights. Whether analyzing market fluctuations or basketball scoring patterns, the principle remains the same: let the data tell the story, not the good intention but misaligned timer on your tv.

Citations: ncaahoopsR, ESPN

Appendix: Side Quest

Final worm hole here... I was curious what teams were habitually bad a scoring a bucket (see the below plot for teams that average droughts greater than 60 seconds). In order for a balanced result, I set a threshold of 33 games or more in the dataset.

Not surprising to me was that I didn't recognize most of the names. My hunch is that these teams generally schedule better opponents due to regional proximity, where there may be a tradition of teams playing each other. And/or the teams get paid to get beaten up. Additionally, with several of the names of being Christian, I assumed smaller enrollment and thereby less competitive on the basketball court (6 of 10 have a clear religious affiliation).

I took a deeper look at Carver College, whose record from 2017 to 2021 tells a sobering story. Care to take a guess at their win-loss ratio during this period? It's an astonishing 3-102, a statistic that reveals much more than just basketball struggles. Carver is an independent HBCU bible college in Atlanta with just 60 students, hidden away on a tiny campus that's easy to miss even when you're looking for it. Their basketball program receives no funding from the school itself and must raise money entirely on its own; even covering their own COVID-19 testing costs. This explains their strategy of scheduling numerous games against Division I opponents, where they receive guarantees between $2,000-$2,600 per game (and often a post-game pizza meal). In the 2020-21 season alone, they played 17 Division I games in just over a month, losing by an average margin of 59 points. My goodness, that would be a tough pill to swallow. For more on Carver, give this article a read.

Page updated

Report abuse