Forking and pull requests

Let’s practice forking and pull requests with an iceclassic prediction!

The Nenana Ice Classic is a long-running Alaskan prediction contest centered on the annual breakup of the Tanana River ice. Every spring, participants try to predict the exact day and time when the ice will give way and a tripwire on the river breaks, stopping the famous clock. As quoted from their website:

The Ice Classic is Alaska’s greatest guessing game!

In Nenana during 1917 a group of engineers surveying for the Alaska Railroad bet $800 putting in their guesses when the river would break up. This fun little guessing game has turned into an incredible tradition that has now continued for over 100 years!

Buy and turn in your $3.00 ticket between February 1st and April 5th to be involved in this long running Alaskan tradition.

Good luck!

In class, we use this example to connect several MUDE topics: working with time series data, exploring patterns and uncertainty, building physics-based and data-driven prediction models, and thinking about how to optimize a betting strategy. The challenge is simple to explain, but rich enough to show how modeling, statistics, and decision-making come together in a real-world problem. Today, we’ll collect your predictions in a public repository!

Task 1 Open public repo

Open the public repository MUDE-2026/iceclassic. The repo already contains a figure that displays all of the historic breakup days and times, and the predictions of MUDE students will be added on top.

Task 2 Fork repository

You cannot just edit this repository as you don’t have editing rights. However, forking is a powerful GitHub feature which allows you to make your changes and propose them back to the original repository. Click Fork on the Code page of the public repository to create a fork in your own account.

Task 3 Add prediction to fork

To add your prediction, add your student ID in the list in data/predictions.txt and enter a prediction in the format (YYYY-MM-DD HH:MM:SS) (year, month, day, hour, minute, seconds).

Keep the last line in data/predictions.txt empty. Git treats line endings carefully; if you accidentally add or remove the final empty newline, it may appear as if the previous line was deleted and re-added in your commit diff.

The repo is set up to automatically update the figure every time a commit is made. However, for your fork this is disabled for security reasons. If you want to see your prediction being added to the figure in your own repo, enable workflow under Actions.

Task 4 Open Pull request to original repository

Open a pull request to the original repository by clicking Contribute or manually starting a Pull request from Pull request in the taskbar.

Create the pull request, and verify that you compare it to the original MUDE-2026/iceclassic repository

A workflow now starts to check whether your prediction is valid. If everything is correct, it will give a green checkmark and the MUDE Team will merge your pull request into the original repository.

Task 5 Expect merge conflicts?

While waiting for the MUDE Team to approve your pull request, do you expect merge conflicts? Maybe they emerge while other pull requests are being merged. Try and fix the conflicts so that the MUDE Team doesn’t have to do it ;)

Task 6 Request changes

If you didn’t do this correctly, the MUDE Team will request changes. Fix these and push them to your fork. The pull request will automatically update.

By Tom van Woudenberg and Robert Lanzafame, Delft University of Technology. CC BY 4.0, more info on the Credits page of Workbook.