Both NYU and CMU offer courses on Statistical Arbitrage and they seem to cover actual trading strategies. Does anyone have any experience with those courses? Do they really teach strategies that are similar to what the quant hedge funds use? I am curious to know what strategies they cover and how they teach backtesting.

I got curious when I read your post. From the CMU website:------------46-936 Statistical ArbitrageThis course intends to provide students with concepts and techniques for statistically and econometrically based trading. The course begins with the general principles of arbitrage pricing theory and the statistical nature of the price and volatility fluctuations in financial markets. It introduces the ideas of market neutral strategies, and provides the statistical techniques required for identifying and exploiting pricing inefficiencies. Various statistical strategies will be covered, including pairs trading, cointegration-based trading, data mining, as well as strategies using the information from derivatives markets. We will demonstrate how to search for arbitrage strategies based on intra-day patterns, long-term patterns, multi-equity relationships. At the end we stress that statistical arbitrage is not riskless, and we discuss how to assess the risk, arising from model misspecification and inappropriate estimation. The topics covered are particularly relevant for proprietary trading, such as in the context of hedge funds. ------------and my guess is that they will cover maths that they feel might be useful for stat arb. they don't know real stat arb, else they won't have mentioned cointegration (which is so passe).

School never teaches you how to make money...

hahaha, thats kinda ture GoldenArcher =)but by all means, I think theres a course coming to london, would love to go but cost too much $$$ for me right now =/http://www.technicalanalyst.co.uk/train ... rading.htm

the guy running the course - Ernie Chan - has a book on amazon and blogs regularly so might be worth having a look at his book before paying up for his course which will probably be a re-hash of his book in any case

His book ('quantitative trading - how to build your own algo trading business') is quite good if what you want is a book that1) encourages you to trade from home2) teaches basic matlab3) explains in- and out-of-sample testing4) explains survivorship bias5) lists some common data sources and APIs6) lists a few basic strategies, mainly mean reverting but also calendar effectsHe doesn't go into strategies in any huge depth.

Is there much in his book on writing algos in excel?

QuoteOriginally posted by: pb273I got curious when I read your post. From the CMU website:and my guess is that they will cover maths that they feel might be useful for stat arb. they don't know real stat arb, else they won't have mentioned cointegration (which is so passe).Why do you think cointegration is out?

QuoteOriginally posted by: momentumpartnersQuoteOriginally posted by: pb273I got curious when I read your post. From the CMU website:and my guess is that they will cover maths that they feel might be useful for stat arb. they don't know real stat arb, else they won't have mentioned cointegration (which is so passe).Why do you think cointegration is out?Sorry for renew this thread. I, too, got confused why did you say cointegration is out?Thanks!