Attached is what the pair looked like when I traded it. I'm surprised I did not have the correlation chart there, however, there you go. Correlation numbers in PTF change very quickly in the console - The correlation was 96.84% on date of the trade - 6th May. It was a hold of over a month, but most bad trades are long holders.
Because it's not correlation that makes you $$, it's the pairs' cointegration. Get a pair that is too highly correlated, they move so close together that you never make money. A cointegrated pair can move apart a nice amount in the knowledge they will come back. Adrian
I know the cointegration issue has come up before and I'd love to be able to measure it easily. Does anybody know a dumbed down formula for it? I'd love to have an excel function =cointegration(rangeA, rangeB)
Nice long Pairs thread - and no time to go through everything. I would like to contribute though. How are you guys measuring and creating trade triggers for your relationships? What are your thoughts about normalising the stock prices using a z-score method? Then creating a composie z-score oscillator to measure the divergence convergence of the normalisations? I have some custom Ninja code to handle the statistical relationships and for backtesting. Trade decisions are then a composite of statistical, technical, sentiment and fundamental triggers. Attached is a sample of the Ninja idea.
New one; Long XOP:IEO @ 0.7726 Managed to get a good entry even though the spreads are a little wacky.
The key to sustainable pairs trading is to know when to cut loss before the correlation/co-integration black swan suddenly appears.... what cut-loss rules does pairstrader software based on? Any suggestions using some quantitative guidelines?
I don't have a good quantitative guideline for you, but one way to achieve black swan protection is to use options. If you have some options savvy look at buying in-the-money calls and puts to simulate long and short stock positions. This is not economical for every stock, but for hi-volume stocks that have tight B/A option spreads it has merit. There are a couple of advantages to using options. The primary one is that your losses are limited especially in a black swan event. Of course there are disadvantages. Primarily options have something called time value which costs you some money. By using ITM options this factor is minimized. IMHO you should know options pretty well before you try this, but if you have the savvy it can be a good strategy. Don
Hi Jonny, Can you advise what was it about this pair that caught your eye, given the correlation is so low (61%)?