Overview
Students are initially screened based on work samples. Instructions for completing a work sample are provided below. If you have completed relevant research previously, you can submit it as is.
Students who clear our screening process are guided through professional grade research by industry mentors. Top performers are announced at the end of the program. Authors of top data science and quant strategy research are invited to present their work at our annual Demo Day in the Spring.
Timeline
Research competitions for 2016/2017 will follow the following schedule:
| 2016 | Aug 15 | Research competitions launch |
| Oct 2 | Fall equity research submissions due | |
| Oct 9 | Data science & quant strategy submission drafts due (optional) | |
| Oct 23 | Data science & quant strategy submissions due | |
| Oct 24 | Fall equity research screening completed; mentees contacted | |
| Nov 7 |
Data science & quant strategy screening completed; mentees contacted |
|
| 2017 | Jan 29 | Spring equity research submissions due |
| Feb 13 | Spring equity research screening completed; mentees contacted | |
| Mar 6 | Announcement of data science & quant strategy projects shortlisted for Demo Day | |
| Apr 27 | Data science & quant strategy Demo Day in New York |
If you missed a deadline but would still like to get involved, please email us.
Screening
To be accepted into the competition, you are required to share a sample of your work.
Equity research
Submission guidelines
We accept any piece of equity research provided that it concerns taking a position in a single publicly traded stock on the basis of fundamental research. Option, pair, IPO, and technical plays are not accepted.
We accept both write-ups and pitch decks, both longs and shorts. There is no minimum market cap, though our hedge fund partners tend to prefer names with $15m+ in average daily trading volume. It is not a problem if your thesis is out of date and is no longer actionable.
Preparing a work sample
If you have not previously completed a stock pitch, we would suggest getting involved in your university's investing club(s). If none of your university's clubs offer a standard equity research training program, try and find more experienced peers to learn the basics from.
Keep in mind that what we are most interested in is not a mastery of DCF modeling but simply evidence of a strong interest in understanding businesses.
You can download a recommended write-up template here and an outstanding sample write-up here.
Common pitfalls
- Qualitative arguments
Building an original narrative around widely available information is highly unlikely to impress our mentors. If you find yourself writing out long qualitative arguments, think about whether or not they contain any fundamentally new information. - Absence of variant view
Warren Buffet has said that “it's far better to buy a wonderful company at a fair price than a fair company at a wonderful price.” Focus on why the stock is mispriced rather than why the company is likely to do well or poorly.
Submission process
Equity research samples are screened twice a year. Fall submissions are due by 11:59pm Pacific Time on October 2, 2016. Spring submissions are due by 11:59 Pacific Time on January 29, 2017.
Data science
Submission guidelines
We accept any work which involves using alternative data to assess company performance - for example, using scraped data to generate revenue, inventory, and / or pricing estimates.
If you are writing up your own data sourcing project, such a web scraping exercise, one page should be enough. If you are writing up your analysis of an existing dataset, we would suggest aiming for two to three pages.
Preparing a work sample
If you have not completed relevant work in the past, we would suggest completing one of the following assignments:
- Building a scraper
If you are familiar with Python and are interested in doing something creative, this assignment challenges you to find an opportunity to track a business through web scraping.
- Predicting revenues using web traffic
If you are new to data science, we would suggest completing Two Sigma's data science assignment from last year. In addition to detailed instructions, it comes with tips addressing students' most common queries.
Common pitfalls
- Differencing
It is not difficult to find meaningful correlations between stock variables. For example, if you are looking at total traffic and total revenues - so long as they are both rising or falling, you will get deceptively high correlations. Be sure to look at changes in your variables. To avoid seasonality issues, analysts typically look at Year-on-Year changes. - Relevance
When looking for new alternative datasets, it is important to think about how closely the dataset you are following tracks the variable(s) you are looking to predict. For example, web traffic may be meaningful for a web-based business but irrelevant for brick & mortar retailers or B2B service providers.
Submission process
Submissions received by 11:59pm Pacific Time on October 9, 2016 will receive feedback from us or from our industry partners. Final submissions are due by 11:59pm on October 23, 2016.
Quant strategy
Submission guidelines
We accept any work which involves building statistical models to gain insight into broad market dynamics. Single-asset, relative value, and special situation plays are generally not appropriate.
A brief write-up outlining your preliminary findings is sufficient. There is no minimum length, but it typically takes at least a couple of pages to summarise a substantive piece of work.
Preparing a work sample
If you have not completed relevant work in the past, we would suggest completing one of the following assignments:
- Supraview of Return Predictive Signals
If you have extensive applied statistics experience, this assignment should be a fun challenge. It is based on data kindly provided by Professor Jeremiah Green of Penn State University.
- FX markets analysis
If you are new to quant research, we would suggest completing Morgan Stanley's quant strategy assignment from last year. In addition to detailed instructions, it comes with tips addressing students' most common queries.
Common pitfalls
- Static models
Financial markets are highly dynamic, and your model should account for it. Even if the methodology you are using is ill-suited for modeling dynamic systems - for example, PCA factors tend to be highly unstable through time - you can generally still generate valuable insight into market dynamics - for example, by looking at variance explained through time. - Misinterpretation
It is often tempting to attach a narrative to your findings. While this is often a core goal of quant strategy, try to avoid speculating when it is not immediately clear what the data is telling you. If this means that your analysis yields no interesting conclusions - that's completely fine.
Submission process
Submissions received by 11:59pm Pacific Time on October 9, 2016 will receive feedback from us or from our industry partners. Final submissions are due by 11:59pm on October 23, 2016.
Please note that quant work submitted for screening purposes is not intended to be complete – you are encouraged to add a Next Steps section. While equity research program participants begin their work anew once they have been accepted into the mentorship program, for data science and quant strategy program participants your initial submission will serve as a starting point for work you will be doing under the guidance of your industry mentor.
Mentorship
Once you have been accepted into the competition, you will gain access to a mentorship program where industry analysts will help guide your continued work.
Logistics
Mentorship consists of regular calls between groups of 4-6 mentees and a dedicated industry mentor. The calls are generally held weekly or biweekly over the course of anywhere from two to six months. We have found that it is most helpful to have a single dedicated mentor working with the same group throughout the program.
Mentors
Starting in Fall 2016, each student accepted into the mentorship program will be asked what kind of background and skillset you would want to see in your mentor. We will then find someone who matches that profile as closely as possible.
Resources
Many data and platform providers are keen to have their products cited in our students’ research. In just the past year, we have been able to offer participants in our research program access to RS Metrics, Omega Point, Ayasdi, ThinkNum, and Alpha Sense.
How to join
To submit an existing sample of your work, please use this form if you have an Upgrade Capital account and this form otherwise.
If you would like to prepare a work sample, please register so that we know to share additional materials and general program updates as they become available:
All registrants are added to the research program mailing list. You can unsubscribe at any time.
If you attend one of our Core Universities, you will receive an invite to join the Upgrade Capital platform. In addition to gaining access to examples of exceptional work we received in the past, you will gain the ability to receive feedback on drafts of your work before you file the final submission.