Market Indices, Market Close as of November 13, 2009
Index	Open	Close	% Chng
DOW INDU	10226.9	10270.5	0.43
S&P 500	1093.08	1093.48	0.04
NASDAQ	2154.06	2167.88	0.64
S&P / TSX	11486.9	11407.7	-0.69
BOLSA	30646.2	31002.1	1.16
FTSE 100	5235.18	5296.38	1.17
DAX	5619.72	5686.83	1.19
NIKKEI 225	9808.99	9770.31	-0.39
HANG SENG	22207.6	22553.6	1.56
SHANGHAI COMP	3175.585	3187.647	0.38

Economic Calendar November 16 - November 20
Date	Release	Consensus	Prior
Mon	CA Manufacturing Sales MoM (%) US Advance Retail Sales (%)	1.7 0.9	-2.1 -1.5
Tues	US PPI YoY (%) Industrial Production (%)	-1.8 0.4	-4.8 0.7
Wed	CA CPI YoY (%) US Housing Starts (K)	0.1 600	-0.9 590
Thurs	CA Wholesale Sales MoM (%) US Initial Jobless Claims (K)	1.0 505	-1.4 502
Fri	MX GDP Constant $ YoY (%) MX Global Economic Indicator IGAE (%)	-6.8 -6.0	-10.3 -6.9

Market Summary, November 9 - November 13

The week of November 9 continued the recent horizontal movement of global equity indices, with the week's close only marginally above open across the board. North American markets performed only slightly worse than the global average. In Canada, mixed statistics showed that while the country's economy was indeed returning to growth, recovery would continue to be rocky. Housing starts, while having improved from the previous month, failed to meet economists expectations, while both international merchandise trade, and new motor vehicle sales beat expectations. The picture was similar in the United States where a greater than expected trade deficit disappointed investors, while both initial jobless claims, and continuing claims were less than had been expected.

Market Outlook, November 16 - November 20

Watch out for a number of important economic releases this week. In Canada, economists are expecting manufacturing sales (MoM) to have increased by 1.7%, a dramatic turnaround from the last reported figures, which showed a loss of 2.1%. In America, Advance Retail Sales and business inventories tell a similar story of revival in domestic consumption. In a positive sign of economic recovery, numbers for both Canadian and American Consumer Price Indices (CPI) are thought to have either stabilized or begun to increase slightly.

Earnings reporting for Q3 2009 continues this week, with a number of companies whose earnings are thought to be a bellweather for economic activity. Notable companies reportig this week include American big box home improvement retailers Lowe's and Home Depot, luxury retailer Saks Inc., and department stores Target Corp., and Sears Holdings Corp.

Story of Interest, November 16 - November 20

Applying Theory in Quantitative Finance: Data Mining (Part 2 of 2)

By Sean Jewell

What if we could develop a system that could (to a certain degree of accuracy) predict where the price of a financial security tomorrow would be, with respect to today’s? Data mining is one attempt to do just this.

An exciting and increasingly popular field at the junction of computer science and statistics, data mining seeks to recognize patterns in data. While the specifics are complex, conceptually the process of using historical data from a financial security to infer future prices is relatively simple.

The first step in designing a data mining system would be to collect historical data for the securities in question. Step two would involve choosing a set of ‘predictors’. These can be anything from a collection of technical indicators (ie: its simple moving average), to fundamental analysis (ie: accounting ratios such as the EPS) or econometric models. The role of our system would involve determining the optimal weightings for each of these predictors in our prediction. The third step would be a ‘training period’ of, say, five years within which we could calibrate our modeling system.

As an example, we could examine Apple Inc.’s stock (AAPL.O) over the last 5 years based on the stock’s relative strength index (RSI index), a moving average (MA), and the company’s earnings per share (EPS). If we were to record on a daily basis whether or not each of the RSI, MA, and EPS correctly predicted the next days move we could assign different scores to our ‘predictors’, based on their successes during the calibration period. Almost magically then, our model would be able to read today’s closing price and determine whether or we should take either a long or short position on the stock for the next trading day.

Traders using these models often find success in the marketplace, with some of the more accurate models accurately predicting the next day’s movements approximately 60% of the time. That is why today quantitative models such as data mining systems make up an increasing proportion of financial institutions intra-day trading. For the financial mathematician then, the future is bright.

GTF Announcements, November 16 - November 20

Please join us as Mr. Arjun Kumar speaks during Trader-in-Residence on Tuesday at 5:30pm in the DSB Atrium.

Mr. Kumar is an Analyst at Sprucegrove Investment Managemnt Ltd. and a returning speaker.

Join us afterwards for FREE pizza and pop in the Commerce Lounge! If you have attended over 80% of TIR events throughout the year then you are eligible to receive a TIR certificate!


To subscribe to the GTF letter and be included in the distribution list please, please request via e-mail to gtf@mcmaster.ca

Disclaimer: The GTF MarketWatch is prepared primarily by students and is viewed as an educational tool.  The GTF is not responsible for consequences of actions taken based on the GTF MarketWatch.