|
The Rotman International Trading Competition (RITC) is an annual event held in Toronto, Ontario, the heart of Canada's financial district. Hosted by the Rotman School of Management at the University of Toronto, the competition brings teams of students and faculty from universities worldwide to participate in a unique 3 day conference. RITC'07 will be held on March 8th, 9th, and 10th, 2007.
RITC utilizes simulated trading cases that closely mimic different aspects of real world markets. Having developed these cases in its state of the art Financial Research and Trading Lab, the RITC promises to present teams with challenges that are pertinent to current market participants. While past competitions have been wildly successful, we continue to improve on the old cases and introduce new ones. After adopting the Rotman Interactive Trader (RIT) platform in the 2006 competition and receiving overwhelmingly positive feedback, we are looking to further utilize RIT's newly added features to offer more variety in our trading cases and a better trading experience. Please visit RIT's official website for more information on the application and supported cases.
Due to logistical constraints, the competition has a limitation on the number of teams that can participate. We strongly encourage faculty from each school to accompany their team as either a coach or observer. If your school is interested in sending a team, please have a faculty, staff, or club head fill in a registration form here.
RITC 2007 in the Media
http://dailynews.mcmaster.ca/story.cfm?id=4604
http://www.wlu.ca/news_detail.php?grp_id=0&nws_id=3021
http://www.haskayne.ucalgary.ca/news/RITC_2007.html
http://www.ceibs.edu/today/news/archive/16929.shtml
http://news.rpi.edu/update.do?artcenterkey=1968
http://www.uq.edu.au/news/index.html?article=11640
http://www3.babson.edu/Newsroom/Releases/RITCToronto3-07.cfm
Pictures from RITC'07 HERE!
RITC 2007 Debriefing by Kevin Mak
The Competition
This year's competition greeted me with a new level of excitement. The expansion of the competition format from a 2 day event to a 3 day event meant that we could provide more time between the cases and give competitors more chances to interact with each other and our corporate partners. Also, the increased participation and visibility of our corporate partners meant that we could provide much more value to the competitors at RITC.
The first night's social Open Outcry session seemed to have been a huge success and numerous faculty, alumni, and sponsors have thanked us for giving them a chance to participate in the competition. This is definitely going to be an event that we keep in future years and I would like to encourage grads to consider coming back to participate. I'm sure I don't need to tell you how fun it is! The candidate profile booklets were well received by the sponsors. I hope that you hold onto your profile booklet for reference because I'm sure that most of you will run into each other in your future career paths. I've heard that some competitors have already been contacted for interviews by our sponsors, so there's definitely value from including yourselves in the list.
The closing ceremonies were not as well organized as we would have liked and I apologize for not having directed clearer instructions to our master of ceremonies. The intention was to have all teams come up and have their picture taken along with being congratulated by Tom and Myself. Unfortunately this didn't happen at the beginning and we only jumped in and changed the format for the top 10 teams. This is certainly something that we will adjust for future years. For those of you who have requested your photos, they are on the web (see above link)!
The Cases
Open Outcry
I think it was noticeable that the errors and general trade mechanics were smoother this year since all of the competitors had a chance to practice their outcry trading the night before. The addition of the stamping system instead of initialing tickets appeared to be clumsy, but trust me when I say it was considerably better than the previous alternative. In previous years, as many as 10-20 tickets needed to be discarded due to illegibility. Further, in previous years students would allege that other people forged their signatures on fake tickets. The stamping system removes both of these possibilities. Two outcry issues still exist: backing out of announced quotes- which is to a certain extent a self-policing system (don't trade with people who do it); and not bringing tickets up to the front (something that traders will always need to police on their own). One suggestion made on a feedback form was to have a ticket-collector in the middle or back of the pit, and this is something that we will look into changing for next year. Some students mentioned they were at a disadvantage due to the position of their analyst classroom with respect to the trading floor, or their position within the analyst room (proximity to the door). All tent cards are randomly shuffled and distributed in the rooms. I unfortunately don't see a way around this but am open to suggestions- I noticed that some teams developed strategies to have one analyst stand by the door and one to stay at the computer. On the signaling side of things, I'm happy to say that people followed the no-noise communication rules strictly and I didn't have to warn or penalize anyone. As always it was interesting to see traders consistently assuming that the market would mean-revert. Our system, for the most part, follows a random walk process. As such, sometimes this strategy ends up being profitable, and sometimes it’s unprofitable. For next year's outcry, we're going to look into adding a cash market of tradable products to create arbitrage opportunities. Given our current technology structure, there is a lot of development work that would need to go into this. However, if we manage to get this working as I envision it, the case will become more interesting and turn into a "regular case" versus a "social/fun case", particularly with respect to score weighting.
Quant Outcry
As all of the competitors are aware, we had a data error in the first heat of the quant outcry case. All of our casefiles are built in spreadsheets with randomly generated elements in them to create variability of results. These spreadsheets can theoretically be run thousands of times with different outcomes each time. From a development standpoint, this makes sense because we develop cases that can be reused. From a security standpoint, this makes sense because nobody knows the result of the case prior to the event. This year, right before the case, we randomized our results, saved the files and then uploaded them. However, during the upload we sent in one of the old datafiles from a previous test iteration, causing some of the data released via the news engine to be incorrect. I'm not trying to use this as an excuse, rather, an explanation of our process and why things ended up being wrong. We will be changing our case procedures in the future to remove this possibility of error.
Since much of my time was spent dealing with the quant outcry error checking and subsequent correction, I didn't get a chance to really observe how well the second heat ran.
Sales and Trader
This is one of my favorite cases. One element that I particularly enjoy with this case is that we are able to run many iterations of the case and collect a relatively large sample of data. For me, this means that this case helps distinguish skill amongst a team better than most other cases where only 1 or 2 iterations are run. There's less randomness in the results. The Sales and Trader case ran very smoothly and there were only a handful of trades that needed to be broken due to out-of-market transactions. I don't foresee any major adjustments in this case for the future as it was very successful both last year and this year.
APT Trading
Our brand new case this year, I'm still undecided on whether I liked this case or not. The feedback was very mixed, with some students claiming it was the best case and others claiming it was the worst. I'm going to take some time to explain some of the strategies that we thought of when testing out the case. First of all, the model is relatively straightforward and students were expected to be able to build a working model that could calculate the final price of the securities given all currently available information. An example was given with the case, and also three trial sessions were run with full sets of data to compare to. If you ran through an entire trading session, you should have been able to say at the end "the closeout price was $56.00 and my model said $56.00 once I typed in all the data". If not, then your model has some errors in it. Keep in mind that in this case, competitors do not need to forecast where indicators are going (they can't). All they can do is react to news shocks by buying or selling the stock based on the new calculated value of the stock. For example, if BANK is trading at $57.00, the GNP value comes out at 0.0046, your model might say BANK is now worth $65.00. As a result of this, you should be buying shares of BANK.
However, we made things more complicated than this. Those who practiced the case, or those who simulated some of the data, realized that the factor sensitivity of stocks was very high. In period 3, you may calculate the value of a stock is $27.00, but by the time you are in period 5, your new calculation shows the stock is worth $23.00. This hurts some of the "buy undervalued, sell overvalued" stock strategies (which will work in the long-run over many iterations, but has volatile results). Instead, what students could do is look at the case from a portfolio perspective. What stocks are overvalued with respect to others? Given the data, students could have simulated long series of economic indicators and returns associated with each security. They could then calculate correlations between the different securities and create efficient portfolios, etc. There are ways to significantly reduce one’s risk through diversification and you could capture profits by trading the differentials between the model prices and market prices while minimizing the risk of the portfolio to factor sensitivity shocks.
One thing I didn't understand during the APT session was when I saw two traders on the same team typing the same data into the same model (on two different computers). I agree that we are bombarding students with data, and we did that intentionally because we originally thought the case was too simple. However, traders have 3 screens between the two of them and it seemed the strategy to share an excel worksheet on the extra monitor worked best. I would suggest a good strategy would be having a dedicated market maker, and an analyst, who trades value (and updates the spreadsheets), while the market maker makes money off the bid/ask spreads with overall value inputs from the teammate. Just something to pass on to future competitors. Some mentioned that ANON was too liquid in that case. ANON is programmed to basically have a net-0 position. They buy shares and sell shares but they generally buy and sell the same amount. When I looked at the book, both the bid and ask side was heavily populated with traders that were buying and selling. I'll look at the data more closely in the future though.
Feedback
I want to thank all of you for the feedback that you have provided. I am happy to say that almost all of it is positive. The comments definitely fell on both sides of the fence, from "too much outcry" to "not enough outcry", so please understand that although we read all of the comments, we may make changes contradictory to your suggestions in order to accommodate the majority. Many people suggested splitting the cases so that there are 2 cases per day instead of 3 on Friday and 1 on Saturday. The timing was considerably easier with the new schedule and I think this suggestion will even further help our scheduling. The breaks between events and meals are always hard to gauge, but I think by large this year we kept unnecessary downtime to a minimum (keep in mind we have to move 160 people around the building, not to mention setup outcry tickets, laptops, analyst rooms, and reset case files-properly). I don't think we were ever more than 15-25 minutes behind schedule. Another common comment was getting more sponsors reps to be present at the competition. This is something that we work hard on every year. One of the best ways for us to reach out is to encourage you to come back as a sponsor rep once you move on to your position on the street!
Scoring - The scoring system will be re-evaluated this year to see if we can find a better way to assign points to returns. We used to use an ordinal ranking system (First place = 1, Second place = 2, etc.) The problem with a system like this is that it doesn't take into account the value difference between ranks. For example, it doesn't matter whether you make $1.00 or $1 million more that the next team, you still are ranked 1 higher than them. Secondly, our current system tries to prevent people from being able to outright win (or lose) the competition in a single case, thus scaling profits according to the best and worst traders. Scaling creates difficulties because cases often become games of bluffing, "how big are you willing to go." In our current system, making $10,000 profit (say on the outcry) might sound reasonable, but if the top trader made $200,000 and the worst trader lost $50,000, your score will often be a mere 25% of the top trader's (they get 10/10, you get 2.5/10, etc). As with all years, traders have asked about their scores, and my main response is that almost everyone feels that they did better than they actually did, and the usual reason is because they don't take into account how they did relative to the top and bottom scores. Also, risk adjustment calculations almost always hit people harder than they thought it would. It was designed that way to discourage taking unreasonably large positions.
Operational Risks - these will always exist, and issues will always come up. We strive to do our best to avoid them and to plan contingencies around them. One issue that came up which most people don't know about is that we were extremely close to losing the wireless network during one of the Sales & Trader sessions. The University of Toronto internet line went down for about 30 seconds. Usually if it's down for more than a minute, the wireless login system re-initializes and kicks everyone off- gladly it didn't. Suffice to say, that is a risk that we have to deal with and sometimes we have to hope that the IT gremlins don't come knocking on our door too often. We also try to give students a chance to test the software in advance as much as possible. Unfortunately, it's impossible to guarantee the stability of our 40 workstations, never mind the 40 laptops coming in the door for the weekend. If people would like to suggest potential remedies to deal with broken excel links, windows crashing, crashing RIT clients, etc., the advisory board and I would be happy to make addendums to next year's rules for score adjustments to accommodate these issues. I'm already thinking about drafting specific rules such as "If more than x% of the competitors get disconnected, the case is thrown out and re-done.", "If a competitor is disconnected for more than y minutes, their score automatically becomes weighted to the remaining partner's score, etc." The more explicit we can be, the less chances there are of surprises coming from on-the-spot decisions that are forced to be made. However, we will never be able to plan for everything.
Feedback Case Scores - With a best score of 1.0, and worst of 4.0, the average ranks that were given to the cases were as follows: S&T 2.13, Open Outcry 2.43, Quant Outcry 2.65, APT 2.69.
Lastly, some people have requested a live scoreboard. Although this is possible with the RIT interface, we are hesitant to implement it due to the lack of anonymity brought on by this. Multiple schools compete with the understanding that their score, unless they place, will be anonymous. I'd like to think that the level of competitiveness will eventually be such that anonymous scores won't be needed, but I think it would be premature to say that we could do that this year.
I once again want to thank you for sharing your interest in the trading competition with me by coming to participate, I hope that you will return in future years.
Kevin
|
