You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Pair Trading Strategy Journal
- Thread starter jonnysharp
- Start date
Do you trade both styles so you are kind of hedged against trendy markets? Like jonny said right now the pairs just seems not to converge (revert to the mean) but they will if the market stop it's short term downtrend. I can imagine that trendy pairs are doing well right now? So your trendy pairs compensate for the mean reverting that dont revert. And the same thing in a choppy market, the mean reverting ones are going to do well and the trendy ones may stall. Do you split your portfolio 50/50 between mean reverting and trending or is it just not a concern or you never thought about it?Quote from yobo:
...Just want to add one thing to Walts comments. There are two types of pairs. Ones that trend in a direction and the ones that dance up and down around a mean without really going in any direction...
Quote from total_keops:
Do you trade both styles so you are kind of hedged against trendy markets? Like jonny said right now the pairs just seems not to converge (revert to the mean) but they will if the market stop it's short term downtrend. I can imagine that trendy pairs are doing well right now? So your trendy pairs compensate for the mean reverting that dont revert. And the same thing in a choppy market, the mean reverting ones are going to do well and the trendy ones may stall. Do you split your portfolio 50/50 between mean reverting and trending or is it just not a concern or you never thought about it?
Mean reverting pairs are tough now. Im having trouble with those too. The trends in the market are very strong. I don't split my portfolio per say with 50/50 but just am aware of both types and trade both. Never really thought about....
My spread sheet is set up for mean reversion alerts as well as 50 day cross overs. So I try to catch a pair for reversion as well as dispersion. I also track all the sector and asset class etf's and pair them off against the spy to see which areas are out performing on a relative basis. For example, technology is gaining some momo and the qqqq/spy is beginning to trend. Create a pair ratio for all those and calcluate the the 50 and 200 day moving averages and graph it and you will see what I mean.
Quote from thetrendfollowe:
LOL thats exactly what the Amaranth and LTCM guys said!
But that's cool, thanks for your response.
FYI Amaranth loss big on a huge one way bet on nat gas that everyone on the street knew about and took advantage of, LTCM loss on exponential betting on Russian bonds that defaulted, I have previously stated I only commit a small % of my a/c to each trade so even if the stock went to 0 I wouldn't be wiped out. Im am currently however testing a time based stop and if it improves system robustness I will implement it.
Quote from jonnysharp:
Yes I wait as it will always come back eventually.
Maybe it is not clear: A pair reverting to its mean does not necessarily imply the trade is profitable. That "mean" is a moving average with limited length, e.g. a few weeks. So if the pair diverges and the ratio increases, the moving average goes up as well; a trade that was open for a long time could close by reverting to a mean value that is significantly higher when it opened.Quote from thetrendfollowe:
LOL thats exactly what the Amaranth and LTCM guys said!
But that's cool, thanks for your response.
Quote from total_keops:
Do you trade both styles so you are kind of hedged against trendy markets? Like jonny said right now the pairs just seems not to converge (revert to the mean) but they will if the market stop it's short term downtrend. I can imagine that trendy pairs are doing well right now? So your trendy pairs compensate for the mean reverting that dont revert. And the same thing in a choppy market, the mean reverting ones are going to do well and the trendy ones may stall. Do you split your portfolio 50/50 between mean reverting and trending or is it just not a concern or you never thought about it?
In my experience diversification with as many different pairs as possible is the best way to go. There are a lot of pairs out there right now that are so stretched out from historical averages, that even a little mean reversion can mean 10 points + in gains. One in particular that has already come in nicely was TSO/VLO 200 x 100, as the crack spreads came in, it ran signifigantly and trended nicely for about 15 points before breaking down. There are plenty more pairs like it that are stretched out and can be trend traded with a trailing stop or a large layering interval. The fallout from the dislocations we have experienced gives us oppurtunities with pairs that seem almost too good to be true, I am keeping position sizes small and stops tight on these pairs and watching technical indicators for trend reversal for entry and exit. I have also noticed that many of my bread and butter mean reverting pairs have begun trending against me. My course of action is to lighten up, discontinue layering and set stop loss levels, some at spread intervals, others at $$$ loss levels. Over the last month, I have been adding more and more of the stretched out pairs, those have helped me make up for the losses that my mean reverting pairs have taken over the last 3 weeks or so. As a pairs trader it has been very hard to let things trend as I am used to getting in and out over and over again, just need to adapt a little and go with the flow.
Quote from saratur:
Maybe it is not clear: A pair reverting to its mean does not necessarily imply the trade is profitable. That "mean" is a moving average with limited length, e.g. a few weeks. So if the pair diverges and the ratio increases, the moving average goes up as well; a trade that was open for a long time could close by reverting to a mean value that is significantly higher when it opened.
A very good point.