Statistical Arbitrage resources
- Thread starter hirtop
- Start date
rmorse
Sponsor
After you do the typical google searches, try search statistical arbitrage and add the name of some top math Universities. You might find white papers. I know this professor at NYU so I added him to the search and got this.
https://www.math.nyu.edu/faculty/avellane/AvellanedaLeeStatArb071108.pdf
If you search enough you will find something at your level and work your way up.
https://www.math.nyu.edu/faculty/avellane/AvellanedaLeeStatArb071108.pdf
If you search enough you will find something at your level and work your way up.
What exactly are you trying to learn about. The word can mean many things. If i new nothing about it but was simply interested in getting started thinking about statistically-based trading, I would suggest
http://edwardothorp.com/
One of the true innovators in the field. Start by reading all of his material.
http://edwardothorp.com/
One of the true innovators in the field. Start by reading all of his material.
Thank you all for the replies...much appreciated!
I do have a market neutral strategy that is working pretty good with decent returns although smaller returns than my other directional strategies (fewer trading opportunities). If you seek to diversify, hedging might be a good avenue.
What I am doing is pretty simple and it is based on empirical observation ...i.e. on a divergence from the mean on two correlated instruments, buy the oversold instrument, short the overbought one, take out a profit when both instruments are returning to the mean.
I am looking to automate this strategy but I need to dig deeper...since I am not a quant nor aiming to become one, I hope to find some clearer info about this matter...the available info on the web is either very superficial or it requires a deep knowledge of statistics and calculus.
After all, we know that all hedge funds are employing statistical arbitrage along with risk arbitrage strategies for their trading portfolios...LTCM was a huge success, but they have miserably failed due to poor risk management.
I do have a market neutral strategy that is working pretty good with decent returns although smaller returns than my other directional strategies (fewer trading opportunities). If you seek to diversify, hedging might be a good avenue.
What I am doing is pretty simple and it is based on empirical observation ...i.e. on a divergence from the mean on two correlated instruments, buy the oversold instrument, short the overbought one, take out a profit when both instruments are returning to the mean.
I am looking to automate this strategy but I need to dig deeper...since I am not a quant nor aiming to become one, I hope to find some clearer info about this matter...the available info on the web is either very superficial or it requires a deep knowledge of statistics and calculus.
After all, we know that all hedge funds are employing statistical arbitrage along with risk arbitrage strategies for their trading portfolios...LTCM was a huge success, but they have miserably failed due to poor risk management.
Find an example of two securities that you think have a relationship, get data, and start playing around looking for relationships or working rules. Most of the interesting things you will learn will be by your own effort, and once you get started you will have a better idea of what outside material will be helpful or not helpful. I think Ernest Chan's blog has some good info on it, as do his books, if i recall.
I agree, you should get out there and search the charts for answers. I've been looking for a good FX PDF on it but have found nothing but info on swap arbitrage and spreading from the CME, but I'm not looking to do that. I've had good results in testing by analyzing charts, looking for correlations and devising strategies to take advantage of differences. If anyone has a good resource please let me know.
Statistical arbitrage (or pairs trading based on anticipation of reversion to the mean) works reasonably well in my experience when diversified to multiple 'high liquidity' stocks...
In theory (charts/academic-papers), the assumption for the sometimes grandiose equity-curve, is to be able to buy (and short) shares at some printed market price (open, close or high-low avg of the day), but seldom does that work in reality... The biggest killer in my experience, of an otherwise reasonable trading strategy (which pairs trading is) is 'slippage' - trying to get a fill on 1,000 shares of a low-liquidity stock (to hedge the long position, or vice versa) on a Monday morning - the stuff of utter frustration, and you end up foolishly chasing the shares (remember, you have to buy/short both sides - it's not as easy as the theory suggests)...
Lesson: Stick to high liquidity pairs (eg., KO/PEP, or XLY/XLP which works well on sector rotations), test them using mean-reversion charting, and (most important) don't rely entirely on statistical reversion - be aware of fundamentals/goings-on in each of the two businesses - and spread out the risk with multiple [high-liquidity] pairs. It's as good a strategy as any, without any guarantee of success.
Best,
Shiraz
In theory (charts/academic-papers), the assumption for the sometimes grandiose equity-curve, is to be able to buy (and short) shares at some printed market price (open, close or high-low avg of the day), but seldom does that work in reality... The biggest killer in my experience, of an otherwise reasonable trading strategy (which pairs trading is) is 'slippage' - trying to get a fill on 1,000 shares of a low-liquidity stock (to hedge the long position, or vice versa) on a Monday morning - the stuff of utter frustration, and you end up foolishly chasing the shares (remember, you have to buy/short both sides - it's not as easy as the theory suggests)...
Lesson: Stick to high liquidity pairs (eg., KO/PEP, or XLY/XLP which works well on sector rotations), test them using mean-reversion charting, and (most important) don't rely entirely on statistical reversion - be aware of fundamentals/goings-on in each of the two businesses - and spread out the risk with multiple [high-liquidity] pairs. It's as good a strategy as any, without any guarantee of success.
Best,
Shiraz