Pair Trading Strategy Journal

Hello everybody,

I have some news about the pairtradinglab website - it now contains a searchable database of about 350,000 pre-screened pairs (NYSE/NASDAQ stocks and NYSEARCA ETFs). There are 3 studies per pair and it is updated every month automatically.

You can take a look on it here:
http://www.pairtradinglab.com/prescreened

You must be a registered user to try. Please note that real stock names/codes are hidden in the search result until you are a premium member (paid service). For now, the feature is in test stage, so I am going to give free premium accounts for testing to limited number of members. There is no way to pay for the service atm, it will be launched later.

If you are interested in participate in the testing and get premium member for free for limited time, please:
1) register on the website
2) drop me an email (to the address on the website!) with your username and I will activate the premium account for you

This is a limited offer so hurry if you are interested :)
 
Why has this thread stopped?

Some negative comments by blowinsky and some recommendations of other sw seems to have killed it.

Blowinsky was highly critical, yet he was still interested enough to keep revisiting the journal. He suggests a number of times that you shouldn't put on a trade which is less that .9 correlation. This is what johnny and others were sticking too anyway and they were doing ok trading pairs.
One thing i'm not sure of is whether blowinsky meant correlation or co-integration (he says correlation).
He also mentions that any sw worthwhile must factor in splits which suggested that PTF didn't so. Having looked at their website the yahoo feed that they use for free does factor in splits.

I'm not sure if other services similar to PTF are better. PTF seems to be very useable in the data it shows. I am running it on an old mac with windows virtual machine allocated 512MB memory so it does crash sometimes but if i had a newer computer and were not running it inside virtual machine then im sure it would be fine.

So...if the method seems to be still intact after some criticism, and the sw is doing what it should then are people still trading this strategy sucesfully??. I have adjusted all trades to be 5k rather than 10k and all the average winning trades seem to be quite small. Not sure if it is worth the risk, especially as I was looking at an RBS pair when it went tits up. (A lot of average winning trades are $60 that i can find) Anyone long RBS would have lost £50k! Slightly concerning in this maket :(

So my question is what sectors are you finding good pairs in today that aren't too risky?

Any thoughts on the stockpair trading platform that jarred recommends, it seems to use options and looks very easy to use. Don't fully understand it though and not sure if trading costs are higher than cfd/spreadbetting.

ps. I am in the UK but looking at US stocks because there doesn't seem to be that many UK pairs from the sectors i have crunched.
 
Hi Tony501,

I have just started pairs trading using the updated PairTrade Finder that also includes cointegration so to see this thread come to an abrupt stop is a little bit disconcerting.

I am trading pairs via a spread betting account and I have a small profit after one month of trading, the current "in-trade" losses are cancelling out about 70% of my profit taken from trades that I have closed so far.

I am taking signals from PTF on stocks with cointegration of greater than 0.9, looking at the slope of the ratio chart and making sure that I trade in the same direction. I am also checking that the PE of the long stock is less than the PE of the short stock. I also check the two stocks on Share Scope to look at the 2 period RSIs and to check for upcoming earnings dates. Having done all of this, the exit signals from PTF seem to show good results on stocks that have moved due to earnings and those trades that go against the slope of the ratio. So am I wasting my time doing these checks?

I have struggled to get anywhere with directional trading and have only made small profits over the last three years.

I would be interested to know if people still think pairs trading has an edge, my back testing shows that certain sectors and markets can return good profits but back testing doesn't always tell the whole truth.

I have found that PTF has limitations due to it's memory management but I am able to find plenty of pairs candidates. I am running PTF on a quad core 2.7Ghz CPU with 8Gb memory and a fast hard disk.
 
Trying to capture spread between two stocks is a crazy game to play. You are essentially trying to scalp pennies nickels and dimes. You are better off trading a directional trend with pairs and using the relative weaker stock as a hedge to help eliminate volatility. Get say from the idea if capturing profits in a day and find pairs you can ride using a bit of leverage to goose returns.

Otherwise you playing a high frequency trading game against robots that are pinging limit orders trying to get hit.
 
I've been pair trading for twelve years and its nothing like it used to be. Low intra-day price movement and low volume has really put a damper on it. Years ago I was holding dozens of pairs for weeks but was also day trading them around the core positions with high success. That barely works now, if at all. Pair trading can still be a profitable strategy on a longer timeframe, but in my opinion it has a very limited return-on-capital. Its has also become very boring and is not likely something that you'll want to stare at full-time, day after day.
 
you stock traders call them pairs. Over there in commodities we call them spreads. You could do the same with options, but your constant enemy (or sometimes unfaithful friend) is time decay.

All it really does is slow the market down to a more human level.

So you don't get shaken out on a big move after an unexpected announcement.

It's still directional, but you get a break on the margin.

I pay for it with twice the commissions and twice the spread.

But it gives me some staying power I wouldn't otherwise have.
 
Hi Yobo and Mike Trent.
Thank you for your postings, you make some very valid points but it sounds like you are primarily day traders and may have a different view of trading to me. I am trading just in the evenings after work and happy to hold positions from a few days to a couple of months.

I am trading the minimum amount I can on each pair for the moment just to see if the idea works and will be worth it in the long run. What is so frustrating, with the sectors and markets I have chosen and back tested, the ParTrade Finder program is generating plenty of entry and exit signals with the vast majority of exit signals being very good and definitely showing a high win rate. But.. quite a few of the pairs I am currently trading don't seem to want to re-converge, they are just hovering around at the extreme of the deviation. I do need to give them more time though.

I like the idea of trading a directional trend using the weaker stock as a hedge, that sounds interesting and I will definitely look into it. As I have plenty of pairs with suitable hedge stocks all I need to do is to find the trending ones - hah, if only it was that easy!

Most of my gains over the last few years have come from directional trades on spot gold but it has been quite scary at times especially as I have been trading it without any hedge. Now that I have got into pairs trading it may make sense to find a suitable hedge against spot gold, it would help me to sleep a bit better I think!

I am also trying out pairs from the PairTrade Finder signals service and so far the pairs have worked out. How they come up with them I just don't know, they often have very poor co-integration, correlation and deviation but they seem to work.

Thank you for your comments and feedback.
 
Hi Yobo and Mike Trent.
Thank you for your postings, you make some very valid points but it sounds like you are primarily day traders and may have a different view of trading to me. I am trading just in the evenings after work and happy to hold positions from a few days to a couple of months.

I am trading the minimum amount I can on each pair for the moment just to see if the idea works and will be worth it in the long run. What is so frustrating, with the sectors and markets I have chosen and back tested, the ParTrade Finder program is generating plenty of entry and exit signals with the vast majority of exit signals being very good and definitely showing a high win rate. But.. quite a few of the pairs I am currently trading don't seem to want to re-converge, they are just hovering around at the extreme of the deviation. I do need to give them more time though.

I like the idea of trading a directional trend using the weaker stock as a hedge, that sounds interesting and I will definitely look into it. As I have plenty of pairs with suitable hedge stocks all I need to do is to find the trending ones - hah, if only it was that easy!

Most of my gains over the last few years have come from directional trades on spot gold but it has been quite scary at times especially as I have been trading it without any hedge. Now that I have got into pairs trading it may make sense to find a suitable hedge against spot gold, it would help me to sleep a bit better I think!

I am also trying out pairs from the PairTrade Finder signals service and so far the pairs have worked out. How they come up with them I just don't know, they often have very poor co-integration, correlation and deviation but they seem to work.

Thank you for your comments and feedback.
 
Quote from yobo:

Trying to capture spread between two stocks is a crazy game to play. You are essentially trying to scalp pennies nickels and dimes. You are better off trading a directional trend with pairs and using the relative weaker stock as a hedge to help eliminate volatility. Get say from the idea if capturing profits in a day and find pairs you can ride using a bit of leverage to goose returns.

Otherwise you playing a high frequency trading game against robots that are pinging limit orders trying to get hit.
More...
I agree with your comments but my crystal ball has always been extremely cloudy. Whenever I think stock A is going to outperform stock B ... it usually reverts to the mean!

Could you provide an example of a recent trade or a possible trade?


Regards,

Slave2Market
 
You must log in or register to reply here.
Back
Top