Nuts and Bolts of Pairs
- Thread starter bungrider
- Start date
Bung,
I have been working this into my strategy with some success, so I will just share a few ideas.
With ETFs, it makes sense to pair these with another ETF since the moves are about the same. I have hedged some shorts against some etfs, but the problem is that on an up day, my shorts go up more than the dollar value of the index.
It would seem on the surface and this is the next leg of my research this weekend to find stock to stock pairs in different industries. For instance I noticed that truckers hit a new high. So I plan to go long the a strong co and short a weak one.
It has also helped to set them up at the same time. Perhaps even writing out your entry, exit and stop loss goals. For instance on Friday I anticipated a sell-off much sooner than it came. As a result I sold part of a long at the open for a gain, which was painful until about 3.
The other thing is to take a profit when you can. When the pair becomes profitable it makes sense in some cases to take it off. Since correlation is not perfect, a negative divergence can blow up the pair.
Regards
I have been working this into my strategy with some success, so I will just share a few ideas.
With ETFs, it makes sense to pair these with another ETF since the moves are about the same. I have hedged some shorts against some etfs, but the problem is that on an up day, my shorts go up more than the dollar value of the index.
It would seem on the surface and this is the next leg of my research this weekend to find stock to stock pairs in different industries. For instance I noticed that truckers hit a new high. So I plan to go long the a strong co and short a weak one.
It has also helped to set them up at the same time. Perhaps even writing out your entry, exit and stop loss goals. For instance on Friday I anticipated a sell-off much sooner than it came. As a result I sold part of a long at the open for a gain, which was painful until about 3.
The other thing is to take a profit when you can. When the pair becomes profitable it makes sense in some cases to take it off. Since correlation is not perfect, a negative divergence can blow up the pair.
Regards
I started trading pairs a few months ago and it has worked out very well. Since it is commission intensive, and spreads can run pretty tight on the highly correlated pairs, more capital leverage/less commissions is helpful - so this is well suited for prop traders.
I also have not found a great wealth of information on the Internet regarding pairs, but if you can find some experienced pairs traders (especially at the big prop firms), they can be a great resource. Also, books such as The Complete Arbitrage Deskbook give some good insights and background.
www.market-topology.com lists the most highly correlated pairs, the subscription would probably be worth it. I got the free trial and am still trading the pair ideas I got from that site three months ago. RD/SC, AMR/DAL are a few examples.
I also have not found a great wealth of information on the Internet regarding pairs, but if you can find some experienced pairs traders (especially at the big prop firms), they can be a great resource. Also, books such as The Complete Arbitrage Deskbook give some good insights and background.
www.market-topology.com lists the most highly correlated pairs, the subscription would probably be worth it. I got the free trial and am still trading the pair ideas I got from that site three months ago. RD/SC, AMR/DAL are a few examples.
I'm not sure how much the subscription is, you can send an email to them to find out I believe.. With the free trial, you can pull up a list of the most highly correlated stocks in the entire mkt. Was much more time efficient for me, than loading data from individual stocks looking for correlations.
I just started tracking the pairs on this list, found the ones that looked good to me and started trading those pairs, and have been working those pairs since, so I had no need for it past the three day free trial.
I just started tracking the pairs on this list, found the ones that looked good to me and started trading those pairs, and have been working those pairs since, so I had no need for it past the three day free trial.
I had some difficulty with market-topology. I can't seem to find out how they run their correlations on the market-topology site. Is it an EOD correlation (assuming you are willing to hold the position for the duration of the day) or would you be better served by a 60 minute correlation? Also, how big of a sample are they correlating? (30 days, 90 days?) There are too many unknowns for me to rely on the information in that website. (HUGE DISCLAIMER: I've used the free trial *only*, so I may be missing stuff that comes with the subscription.)
I think your approach would tend to suit you better. You know the timeframes you are trading and how far back you want the correlation to hold--you could even see how the correlation varies over time to really get to know that pair. I run mine in Tradestation, but I think Excel will work equally well, especially since you have it all set up.
The secret to making money with a good pairs strategy is to not get too bogged down in the math and the correlation. Use the basic "price difference" function available on Redi and other software, check back 3 years and also 3 months. The intraday trading of the pairs is where the money is....not buying and holding the pairs at differing price intervals.
Many times you will never even do both sides, just buy the stonger "leaning on the weaker"......
It takes a few hours in class to get the concepts out...but I thought a little help here would be ok.
Don
Many times you will never even do both sides, just buy the stonger "leaning on the weaker"......
It takes a few hours in class to get the concepts out...but I thought a little help here would be ok.
Don