This is a course of webcasts (about 15-30 minutes apiece), designed to introduce you to the basics of statitiscs, primarily as practiced in finance and investing. As with my accounting class, I will start with the open disclosure that my knowledge in statistics is limited to what I use on a regular basis, and that I have no interest (or expertise) in delving into the depths of statistical theory. The class webcasts are right below, followed by links to the statistical tools that I find useful, and readings on each topic. With each session, I also have a post-class test and solution, some more involved than others, testing the grasp of the material in the session.
Session Webcast |
Short Description |
Supplementary Material |
|
Preview |
Provides an introduction to (my) version of a statistics class, including a lead in as to why I think it matters in finance. |
||
1 |
Lay the groundwork for what statistics covers, and why the need for understanding it is greater than ever before. |
||
2 |
Examine why we use samples instead of looking at populations, and the perils of sampling bias and error. |
||
2A |
Sampling: Applications | Stock indices as samples, albeit non-random, and use of sampling to test investment strategies and effects of events on stock prices. |
|
3 |
Describe the metrics that we use to describe data, as well as how and why we use them. |
||
3A |
Data Descriptives: Applications | Compute data descriptives for a variety of data in finance, from asset returns over time to PE ratios, costs of capital and dividend yields across companies. |
|
4 |
Examine how we pick distributions to describe data, and the advantages of doing so. |
||
4A |
Data Distributions: Applications | Use stock price and return data to test for normality in distributions. |
|
5 |
Describe how to measure relationships between two or more variables, and how those measurements can be used in prediction/analysis. |
||
5A |
Data Relationships: Applications | Look at both micro and macro examples of regression uses in finance and investing. |
|
5B |
Data Relationships: More Applications | Describe the process of building a multiple regresssion, checking for significance and non-linearities with finance examples. |
|
6 |
Lay out the rules on estimating probabilities, and describe probabilistic techniques that can be used in analysis (decision trees, probit/logit, scenario analysis) |
||
6A |
Probability Tools: Applications | Provide examples of decision trees, simulation & probability estimation in finance. |
|
6B |
Simulations: Applications | Look at the use of Monte Carlo simulation in the valuation of a company. |
Tool | My comments | How to get |
Microsoft Excel | As an Apple loyalist, even I have to admit the world runs on Microsoft Office, and I am no exception. Microsoft Excel is my workhorse for collecting data, and its data analysis toolpack is surprisingly versatile. I can do most of what I want with its pre-set functions, and If I were truly an Excel ninja (which I not) almost everything. | Subscribe to Office 365 |
StatPlus | There are many statistical add ons, to Excel, but my favorite (perhaps because it is one of the few that has expended resources to build and maintain a Mac version) is StatPlus. If you are working with a limited budget, it should do the trick for you. | StatPlus Home |
Wizard Pro | I find this data analysis program delightfully intuitive, to create pivot tables and to make sense of very large datasets. It is often my first stop, before I go on to StatPlus and SPSS to do statistical analysis. | Wizard |
SPSS | There are many stand alone pure statistics packages, and many predate personal computers. I used SPSS on mainframe computers when I first started looking at data, and it is that loyalty, and the fact that SPSS has a Mac version that keep me in its corner. That said, it is overkill for almost everything I do, containing powers that I not only have never used, but would not know how to use. | IBM SPSS site |
Crystal Ball | This is my go-to program for simulations. As an add-on to excel, the learning curve is not steep, and it comes with an impressive array of choices for distributions. I have also heard good things about @Risk, and having seen output from it, it seems to do be very similar to Crystal Ball. A significant caveat is that neither program works on a Mac, and I have to do unpleasant (for me) end runs to get around that limitations, including turning my Mac into a PC (a skin-crawling exercise) using Parallels Desktop. | Oracle Crystal Ball |
(While all these readings were accessible at the time that I put together this list, a few came with restrictions (Scientific American allows you a maximum of five free articles, and I have four on my list). It is also a reality that more and more of these readings will end up behind paywalls. The readings still remain worthwile, but not if you have to pay significant amounts to subscribe to a publication for a whole year to read them.