Unformatted text preview:

Statistics 434: Final ProjectsProject Overview• The nominal task is either to design and test a “trading system,” or todevelop and test a empirical thesis about return processes.• Your real goal is to show that you have achieved a high level of skilland knowledge in the practical investigation of financial time series. Youshould demonstrate both the mastery of statistical technique and thesoundness of your personal judgement.• Your report can be written individually or by a two-person team. Indi-vidual reports are not to exceed 12 pages, and team reports are not toexceed 20 pages, inclusive of well-designed tables, well-chosen plots, andwell-selected co de fragments.• It is expected that your report will be a synthesis — not a blog. You willwant to begin with an executive summary of one page . You canthen develop your individual themes in several (almost) self-standing sec-tions. The heart of each of these theme sections is likely to be a discussionof one or more individually designed tables of results. At each level, beginyour story with the punch line and then pile up the evidence.Be ruthlessly clear and never overstep the bounds of your evidence.• There is no need for a “conclusion” at the back. For an in-house report,the conclusion should be up-front. Put all of the main conclusions onpage-one. More minor conclusions should be displayed on the first part ofthe corresponding section.• Learn to write like this, and your ideas will get noticed. This is not poetry;it’s business.• Be very attentive to the principles of intellectual integrity. If youuse an idea from another source, even a non-public source, be very carefulto acknowledge the source. If you hear a snappy phrase and you want touse it, put it in quotes even if you can only attribute it to “anonymous”(or worse yet, Cramer). Please do not put me or you in the position ofworrying about integrity issues. It costs nothing to give proper attribution,and the failure to provide such attribution can cost a career.Project Process• Discussion of your project ideas will be an important part of our classduring much of the last two weeks of class. Talking about your ideasin class is one of the surest ways to make them better. The goal is foreveryone to provide a first-rate report.• We will creep up on the Final Project in three steps. The time-line willbe posted on our class blog, but here are the basic steps:1. A written project proposal (one to three pages, no longer, — prettyinformal, not graded — but on time)2. A twelve-minute project proposal presentation. Power Point andpolished — but clear and honest. It counts toward your grade but youwill be judged on organization and thoughtfulness, not your speechmaking ability.3. Your final project — the bottom line (specifications to follow)• The brief written “Project Proposal” is mainly to give you space to thinktrough what you intend to do. I will read the written reports before youroral report. This will help me to get up-to-speed with your plans. It willalso help you to make your ideas more concrete.This proposal is not expected to have a punch line, but it should stateas completely as possible the themes that you plan to explore and theprogress that you have made. Useful progress at this level would be thecollection and organization of the relevant data, and perhaps some EDAto confirm that your data really is in working order. If you have a seriousdata problem, it this is the time to discover it — not after 80% of yourproject has been written up.• You should worry about being “stuck” with what you propose. What youpropose and what you finally do, can be different. In science, one almostalways proposes one thing and then does another. The miracle is thatproposals are still useful.• You will give a 12 minute presentation of your proposal in class. We’llhandle the logistics of this as the end of the term draws near.– The scheduling is always tighter than one would like, so we cannot af-ford to be sloppy about the 12-minute time allocation. A 12-minutespresentation really is brutally brief, so you want to jump right intothe issues. We don’t need any of this “since the advent of moderncomputers” baloney.– Attendance of everyone is critical on the presentation days. Youreally do need to be there to see what your colleagues are doing.Everyone’s effort pulls up the effort of everyone else.– One can say a lot in 12 minutes if one is well prepared and wellfocused. The goal of this proposals is to stimulate thoughts that willadd depth and polish to everyone’s final project.– At the time of the proposal you don’t need to press too hard for adeeply original theme. Originality will emerge naturally after youhave begun to dig honestly into your data. What is needed at pro-posal time is a plan that is complete enough and clear enough so thatboth further opportunities and potential problems will be evident.– There may be a proposal or two that I feel has “gone off the rails.”If this happens to you, it is not a happy event, but neither is it anabsolute disaster. There is time to get back on track. Our collectivepromise is not to be “mean” to a team whose proposal bombs. Wecan all learn from such a disaster, but I’d still prefer that we not haveany.Data Collection and OrganizationYou should let your project proposal shape your data collection, but, as agauge of scope, most people will need data on 8 (or so) stock returns (or otherfinancial series). Most of our class discussions have focused on daily returns,but you may find that weekly returns are more informative. Weekly returnsare guaranteed to have a better signal to noise ratio, though this comes at theprice of increased likelihood of non-stationarity. Also, trading strategies that useweekly returns are less sensitive to your assumptions about transaction costs.We will discuss some designs in class for testing trading strategies. Therereally is quite a bit of flexibility, but you want to make sure that you haveanticipated the natural criticisms of your design.Finally, the choice of your assets and the choices of your time periods willdepend on the themes that you intend to explore. Certainly fixed income ex-plorations are possible, but they are tricky and we should discuss your ideas.The same is true with options or futures. You can use these, but you’ll need toexert some care to be sure that you are using them correctly. Also,


View Full Document

Penn STAT 434 - STAT 434 Final Projects

Download STAT 434 Final Projects
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view STAT 434 Final Projects and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view STAT 434 Final Projects 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?