What are the most useful results that I should put on a mathematical finance cheat sheet?
Am I missing anything important:
$\begingroup$
$\endgroup$
1
-
2$\begingroup$ (1) Beware of model risk, (2) beware of model risk, (3) beware of model risk. If you always keep this in mind you will do just fine. If someone wakes you up at 3 in the morning and tells you that the trading performance deviated from expected model performance by metric x then you should immediately be able to answer the question whether the model should be retired/improved/reworked. I am telling you because it is not what might happen but what will happen. $\endgroup$– Matt WolfMay 2, 2014 at 5:51
Add a comment
|
3 Answers
$\begingroup$
$\endgroup$
2
There is a very famous math finance cheat sheet already (by Prof. Wystup), you can find the content here:
-
$\begingroup$ funny, some professors really seem to operate on stringent budgets. +vd $\endgroup$ May 2, 2014 at 8:11
-
1$\begingroup$ Mine is free and customisable though :) $\endgroup$ May 17, 2015 at 9:25
$\begingroup$
$\endgroup$
At this stage your sheet is focus on "stochastic calculus for derivative pricing". It is just a subset of math finance. You are missing:
- risk management (VaR, quantiles, etc) -- more statistics than stochastic calculus. See for instance the content of Attilio Meucci's book.
- quantitative trading (optimal trade scheduling, smart order routing, microstructure) -- more control and point processes. See for instance the content of Lehalle-Laruelle's book.
You may consider that portfolio allocation is a subset of the first topic and thus include it in it. Or have another section on its own that for.
You may thus need at least three double pages instead of one...
$\begingroup$
$\endgroup$
Depending on how long you want the cheat sheet to be, I think maybe touching on some numerical methods that are of importance would be useful. Simply because quant jobs will often require intense analysis of the error in your models and the algorithms that implement them, something along those lines I could see being useful.
Additionally, many quant interviews may consist of asking for pseudocode implementations of various ideas, so when it comes to automating the process of solving a differential equation to use in a model, numerical computation becomes important. Especially for firms where speed of deployment is a concern, knowing how to control error and create streamlined and robust numerical algorithms that run quickly will be important.
