— Science Olympiad — 2 min read
I supervised Data Science C at UT Invite this past weekend. It was really awesome to see the turnout and teams did pretty well! I couldn't have done it without the help and support of many other hardworking volunteers. In particular, Sid Arora, Roger Zhong, and Ryan Chhong contributed extensively to writing the exam and coding challenges.
I first ran Algorithm Design (the precursor to this event) over 1.5 years ago, at the 2019 UT Regional tournament. Around six teams participated. This past weekend, 88 teams from around the country participated remotely in Data Science, including highly competitive out-of-state schools such as Troy (CA) and Mason (OH). This is a huge win for this trial event!
Kudos to all the teams that are studying and preparing entirely from home. Being unable to talk to your team and partners in-person has got to suck. It takes a lot of grit to stick with it during these times, so congrats to every team that participated!
Being remote, this competition was markedly different from previous tournaments. The written test was administered via Scilympiad, and since there were so many teams, we made the majority of the written test multiple-choice (to hasten grading). This is no fun for test-takers, but sadly that's the reality of running a large tournament. The coding challenges were provided via a link to a Google Colab file, which contained the instructions and starter code. Competitors were instructed to clone the file into their Google Drive, write the solutions in their cloned copy, and then submit a link to this file. This worked pretty well, and I'll probably continue doing this for the rest of the season.
The score distributions are shown below. The total score is out of 100. The written exam (60 possible points) has a pretty nice bell shape, possibly due to the multiple choice. The coding challenges (40 possible points) allowed teams to exercise a bit more creativity, but understandably, it was also more difficult for teams. Some teams did not even attempt to solve them; others submitted solutions with no working code. Out of 88 teams, 34 had nonzero coding subscores. On the other hand, a few teams excelled, getting nearly perfect coding scores. The score histograms are shown below (on the coding subscore, I removed the zero-scoring teams).
The advice I would give teams is: practice your python! And practice using Google Colab. 10 points out of the 40 were doable with only a beginner's understanding of python. Those 10 points can dramatically change your placement! Also, remember that there's no limit to the amount of pre-written notes you can use, and remember to manage time carefully. As always, if you have questions, please reach out to me!