5 Top Trends from the Statsketball Tournament Entries
March 17, 2017
While many may choose their NCAA Men’s Basketball Tournament predictions on a classic gut-check (or maybe by mascot), statistics buffs used a more systematic approach to enter This is Statistics’ Statsketball Tournament.
The competition drew submissions from undergraduate and high school students across the country for its two challenges, the “Pick ‘Em” Upset Challenge and the “Build Your Own Bracket” Draft Challenge.
As judges took an initial review of the submissions, these five trends stood out:
1. Students used sophisticated techniques to arrive at their predictions, including machine learning, simulations, and cross validation. Standout methods in the contest may be standard for professionals, but are noticeably advanced for students.
2. The most-predicted upset is for Wichita State (seed 10) to defeat Dayton (seed 7), with 77 percent of entries making this prediction
3. Other popular upset picks include 10-seed Marquette over 7-seed South Carolina (62 percent of entries) and 12-seed Middle Tennessee over 5-seed Minnesota (60 percent of entries)
4. Students approached the Draft Challenge as strategic problem to solve with statistical thinking, rather than intuition. Their selections of teams shied away from 1 seed teams who had a low risk, high cost.
5. UCLA (3 seed), was the most selected draft team. Students often selected middle seeded teams for their brackets due to their bargain draft point cost and chances of advancement.
ASA will announce the winners of these contests shortly after the Championship game on April 3. Keep an eye on the Statsketball page for updates and winner announcements.
Get Your Class Involved Offer the challenge as extra credit or as a class assignment Volunteer to help students form teams, and/or to sponsor teams Offer your classroom as a meeting and work space for student teams Assess your students’ knowledge and interest in statistics Motivate them about the future benefits of learning statistical skills …
Demand for employees with statistical skills continues to soar. According to Glassdoor, the top job of 2017 is data scientist. In addition to this, three more of the top 10 professions rely on data analysis skills, meaning that at least four of the top 10 jobs of 2017 would require statistical skills to be hired. Following…