MATH3901: Higher Probability and Stochastic Processes

- 2 x 7.5% class tests (question bank given, randomly question selected to complete in TEN minutes)
- 25% midsession test (one-sided A4 cheat sheet allowed)
- 60% final exam (two-sided A4 cheat sheet allowed)

For the higher version,
- MATH2901 or DN in MATH2801
- MATH2501 or MATH2601
- MATH2011 or MATH2111
(Essentially all the core level 2's. No linear models is required. You should definitely know your MATH2901 before coming here - it is retaught but crash coursed and pretty much assumed all the way. I think calculus is more useful than linear algebra for this course though.)

This is one of the Level 3 core courses to students pursuing a statistics major.

Among the three cores, I personally believe this is the hardest of the lot. MATH3821 gets its difficulty from having too much content. This course is about probability theory, and that's not something you can simply overcome with memorisation.

This course is pretty popular among mathematics students as a whole due to the fact that MATH2931 is not a prerequisite and that it's really just a probability course, not statistics. Part of this is influenced by the fact that many people genuinely enjoy the first half of MATH2901, and drop off in the second. Mastery of this course can be extremely useful in the short term, because the strengthened intuition in probability can help useful in trading interviews, where probability based questions resurface more than you might expect.

Stochastic processes are just useful in general for people who want to work with the market. A simple summary: a stochastic process is just a sequence of random variables measured over time (can be both discrete time and continuous time).

Assessment tasks are definitely more challenging in this course, compared to its sister course MATH3911. Although most questions are doable, every past paper I've seen seems to introduce a completely unrelated question to that of the previous paper. They all use the same techniques, but the question style appears foreign.

For the class tests, the questions are the same. The difference is that he changes the letters in the actual test. (Or for MATH3801, he might change a number instead.) Figure out a way to write it as quickly as you can, but not too quickly that you'd struggle with different letters/numbers. Some explanation skipping is okay, just not too much.

Also, after studying for the finals, I would advise being careful with Brownian motion. It's also a really interesting concept to learn about, but it can take a while to understand because it's so far to the end and is fairly different to the stochastic processes you see before it.

(The difficulty would've been 4/5, but it jumped to 4.5/5 after I saw the challenge of this year's midsession. It was a pretty spiked paper to offset that we had all the lecture notes available, and not just the cheat sheet.) Edit: I'm convinced 4.5/5 was the right choice now after studying for finals.

2 x 2hr lectures, 1hr tutorial

4.5/5

Yes, however prior to COVID-19 he also used the whiteboard, which is obviously not recorded.

Dr. Gery Geenens.

Full lecture slides given at the start of the lecture. Pretty much all you need to learn the course. An abundance of midsession and final past papers are provided but with NO ANSWERS.

4.5/5

"An Introduction to Probability Models" by Sheldon Ross (8ED, 9ED and 10ED all work). The course was written off this textbook so it should be useful, but I haven't had need for it.

20T1

SY (will never find out what I truly got )

3 * 5% class tests, 25% mid-semester exam and a final exam. (yep, no assignments)

Theres a lot haha

Prerequisite: MATH2901 or MATH2801(DN) and MATH2501 or MATH2601 and MATH2011 or MATH2111 or MATH2510 or MATH2610.

Level 3 statistic courses usually have a lot of contents (700+ slides iirc), you will learn a lot of interesting stuff like Markov chains, Queueing theory, Branching process etc. It is fun and challenging, you would learn a lot of new ways to solve probability related questions. If you are doing ACTL/Adv Sci (Math), it is pretty much a revision of ACTL2102 without time series. But be sure to be familiar with brownian motions, stochastic differential equations (SDEs) and martingales, coz ACTL3182 pretty much assumes them and straight away jump into the derivation of Black Scholes models and other stochastic models.

3 Hours Lecture and 1 Hour tutorial

2/5

Yes

Dr Gery Geenens

N/A

5/5

Introduction to Probability Models by Ross

2018/1 (i know, my memory sucks)

HD