My overall feelings about the unit is that ETC2420 is a comprehensive unit that shifts the focus from traditional formula-memorisation to the logic of making decisions under uncertainty. It is a rewarding but challenging to engage with that provides students an understanding about basic probability concepts and how they elevate to advanced computer-simulated models.

The lectures essentially went over the lecture slides provided in the Moodle. Because of this, you could still do well in the unit without attending or watching the lectures as this was what I did. The lectures were largely theory-based, but would go into detail with the coding or math behind concepts if the lecturer was on pace with his lecture as they are fairly long slideshows (60-100 slides long). Lectures were recorded and put on the unit information tab of the Moodle, so you can also check up on your understanding of the content if one part of the slideshow gives you trouble.

The tutorials typically started with the tutor introducing the exercises and dataset we'll be working on, then we would work through the chosen exercises either in pairs or individually, and then once people had gotten up to a certain exercises our tutor would then go over solutions/how they would've done it. It was not necessary to attend, and I do feel you could still do well in the unit by just doing the exercises at home as the hour length feels very short as we frequently didn't finish the exercises. The best preparation for the tutorials was to be up to date with the content covered in last week's lecture, as well as downloading the dataset that is used for the exercises of that week. The tutorials covered lots of the coding aspects of the unit in R, so it would be useful to attend if coding in R is not a strength of yourself.

The quizzes were quite easy although questions can be worded weirdly/ambiguously. There was no preparation needed for the quizzes and marking was mostly automated since it was largely multiple-choice questions, and the few short answer questions there are were marked very leniently.

The first assignment was very easy and is really just to make sure you know how to present a report in RStudio.

The second assignment was a group project which took significant time from all members, around 8 or so hours each if you're aiming for a high result. The marking for the second assignment was quite fair, and feedback is detailed enough to know where you went wrong.

The assignments are not very similar to the exam due to the nature of the exam being a short-answer/MCQ test, it would be closer to the tutorials. Practice materials were not provided for the quizzes and the assignments, but was provided for the exams.

The exam was based on all the weeks of content, where Part A covered the first 5 weeks, Part B covered the middle 3 weeks, and Part C covered the last 3 weeks, where some of Part B and C also assessed the first few weeks too. It was a closed book exam with no calculator as no calculations are done, and there was some time-difficulty due to the amount of writing expected of you, but if you know the content well enough to where you can apply it quickly, then you will have no issues time-wise. There were three practice exams given, and these were very similar in format and question type to the actual exam so make sure to complete these and know the types of questions and the process of the answer very well.

When answering questions in the unit, it should be answered realistically rather than strictly theoretically. For example, if you got a question that asked if the R-Squared value will go up when adding another coefficient to the model, it should be answered true even though they're is a small chance that the R-Squared stays the same which is very unlikely to happen in practice.