The next long term trade
- Thread starter nitro
- Start date
IMHO,
Once you get the basics of the "mathematics" of _SHORT_TERM_ pair trading, there isn't much to doing it. The key is finding really good pairs, and not necessarily looking for them in the standard ways.
For longer term pair trading, the situation is very different and some "advanced" (advanced from the point of view of traders, elementary otherwise) mathematics may prove to be useful. I have not seen any book discuss this in length.
nitro
Once you get the basics of the "mathematics" of _SHORT_TERM_ pair trading, there isn't much to doing it. The key is finding really good pairs, and not necessarily looking for them in the standard ways.
For longer term pair trading, the situation is very different and some "advanced" (advanced from the point of view of traders, elementary otherwise) mathematics may prove to be useful. I have not seen any book discuss this in length.
nitro
I just saw the following on Briefing - LOL 
nitro
======
14:27 ET
Beta Bets : The equity market has been on some kind of roll since March, but for all intents and purposes, the fun started last October - October 10 to be exact - when the Nasdaq and S&P 500 bottomed at 1108.49 and 768.63, respectively. At their current levels, they are up approximately 70% and 34% from those low points with broad-based sponsorship behind the move.
Some stocks, however, have trailed the rally in noticeable fashion. In some cases, the underperformance has been warranted as company-specific issues have sullied their fundamental appeal; in other cases, the underperformance is to be expected as the improved outlook for economic growth has prompted participants to shed their risk averse mindset and to favor higher beta names.
Whatever the cause, the laggards stick out like a sore thumb in this market environment. We've seen time and again, though, that there can be fortune in misfortune, and with concerns increasing about the market's overbought condition, there is a good chance that the underperformance gap may narrow in the weeks ahead. The question is, though, will the leaders simply correct while the laggards continue to languish or will the laggards benefit from a rotation of capital in the wake of a leadership correction?
The answer is probably somewhere in between. Companies like Schering-Plough (SGP), which have been weak due to fundamental issues, have a long way to go to restore investor confidence. Companies like Procter & Gamble (PG), meanwhile, which have good fundamentals, but have underperformed because of the renewed appeal of higher beta names, stand a good chance of benefitting from a rotation of capital precipitated by a correction in momentum plays.
Whatever the outcome, a few issues, we think, will stand in the way of a notable resurgence of year-to-date laggards. First, September is the final month of the third quarter, and barring a disastrous event, fund managers will want to be holding the best performing issues in their portfolios for marketing purposes. Secondly, as we move into the fourth quarter, tax-loss selling will come into play as investors look for ways to offset capital gains. In recent years, tax-loss selling hasn't been much of an overhang because there have been fewer gains to offset. However, with the Nasdaq and S&P up 41% and 17% on a year-to-date basis, and many stocks up much more than that, tax-loss selling could be more of an issue this year.
So, while there may be some opportunities in lower beta names over the near-term, we wouldn't expect them to get too far ahead of themselves, for as long as the market continues to roll, it will be the higher beta issues that continue to lead.-- Patrick J. O'Hare, Briefing.com
===
nitro
======
14:27 ET
Beta Bets : The equity market has been on some kind of roll since March, but for all intents and purposes, the fun started last October - October 10 to be exact - when the Nasdaq and S&P 500 bottomed at 1108.49 and 768.63, respectively. At their current levels, they are up approximately 70% and 34% from those low points with broad-based sponsorship behind the move.
Some stocks, however, have trailed the rally in noticeable fashion. In some cases, the underperformance has been warranted as company-specific issues have sullied their fundamental appeal; in other cases, the underperformance is to be expected as the improved outlook for economic growth has prompted participants to shed their risk averse mindset and to favor higher beta names.
Whatever the cause, the laggards stick out like a sore thumb in this market environment. We've seen time and again, though, that there can be fortune in misfortune, and with concerns increasing about the market's overbought condition, there is a good chance that the underperformance gap may narrow in the weeks ahead. The question is, though, will the leaders simply correct while the laggards continue to languish or will the laggards benefit from a rotation of capital in the wake of a leadership correction?
The answer is probably somewhere in between. Companies like Schering-Plough (SGP), which have been weak due to fundamental issues, have a long way to go to restore investor confidence. Companies like Procter & Gamble (PG), meanwhile, which have good fundamentals, but have underperformed because of the renewed appeal of higher beta names, stand a good chance of benefitting from a rotation of capital precipitated by a correction in momentum plays.
Whatever the outcome, a few issues, we think, will stand in the way of a notable resurgence of year-to-date laggards. First, September is the final month of the third quarter, and barring a disastrous event, fund managers will want to be holding the best performing issues in their portfolios for marketing purposes. Secondly, as we move into the fourth quarter, tax-loss selling will come into play as investors look for ways to offset capital gains. In recent years, tax-loss selling hasn't been much of an overhang because there have been fewer gains to offset. However, with the Nasdaq and S&P up 41% and 17% on a year-to-date basis, and many stocks up much more than that, tax-loss selling could be more of an issue this year.
So, while there may be some opportunities in lower beta names over the near-term, we wouldn't expect them to get too far ahead of themselves, for as long as the market continues to roll, it will be the higher beta issues that continue to lead.-- Patrick J. O'Hare, Briefing.com
===
First, thanks for the answer, nitro.
Maybe you can explain the "mathematics" of _SHORT_TERM pair trading a little more detailed. I'm not sure if I'm with you.
There are many ways to trade pairs, but two ways seem to stand out amongst traders, short term pair trading, and longer term pair trading.
Short term pair trading is where you may or may not do both sides of the pair at the same time, but are "leaning" on the other leg and trading the other. This is a favorite of traders that like to go flat at the end of the day.
The other way to trade pairs is to put them on both at relatively the same time, and then manage the position by adding or removing some on both sides or one side of the pair, but leaving it in days and sometimes months.
These two ways require vastly different skills and IMHO, different types of anlysis. In the old days, all people used to do was to pick two equitites in the same sector and eyeball the pair for comovement. They then dove in and tried trading it. This worked because there were few that were doing this kind of trading and mistakes were not usually punished.
As more and more people got into it and the pairs became more and more efficient in relation to each other, the analysis started to get more and more sophisitcated, with things like correlation analysis and fundamental analysis being done on the pairs. This too worked for a while. However, recently the money that is left in the markets these days is largely Hedge Funds, and this is a staple strategy for them.
So on the short term, you have high frequency data analysis that is being done because of the highly non-linear relations. And on the longer term, you now have cointegration analisys being done.
This could get technical, but it comes down to the "fact" pairs analsisys is:
"1) Co-dependencies between asset returns may be highly non-linear in nature, but correlation is only a liner measure.
2) Unconditional correlation is essentially a static measure, but dynamic relationships between markets may exist with a non-stable lead-lag nature.
3) Unconditional correlation only exists when the returns are jointly covariance stationary. In other words, correlation estimates jump around because they are being measured on non-jointly stationary series.
To summarize the above, correlation is not a stable measure. The observed instability (in todays pairs) is not (necessarily) due to the lack of joint stationarity, but is a result of non-linear relationships.
At the longer time frames, cointegration analysis allows for the dynamic co-dependencies in financial markets. In the shorter time frames, you probably need chaotic (attractor) type analysis to deal with the same non-linearity..."
I am developing a software that uses some of these concepts and I hope will allow me to get an edge not only on pairs, but on sector trading (which is a kind of flip side to pair analysis) as well. We'll see if I can couple the software that hones in on good pairs together with my skill in trading...
nitro
Short term pair trading is where you may or may not do both sides of the pair at the same time, but are "leaning" on the other leg and trading the other. This is a favorite of traders that like to go flat at the end of the day.
The other way to trade pairs is to put them on both at relatively the same time, and then manage the position by adding or removing some on both sides or one side of the pair, but leaving it in days and sometimes months.
These two ways require vastly different skills and IMHO, different types of anlysis. In the old days, all people used to do was to pick two equitites in the same sector and eyeball the pair for comovement. They then dove in and tried trading it. This worked because there were few that were doing this kind of trading and mistakes were not usually punished.
As more and more people got into it and the pairs became more and more efficient in relation to each other, the analysis started to get more and more sophisitcated, with things like correlation analysis and fundamental analysis being done on the pairs. This too worked for a while. However, recently the money that is left in the markets these days is largely Hedge Funds, and this is a staple strategy for them.
So on the short term, you have high frequency data analysis that is being done because of the highly non-linear relations. And on the longer term, you now have cointegration analisys being done.
This could get technical, but it comes down to the "fact" pairs analsisys is:
"1) Co-dependencies between asset returns may be highly non-linear in nature, but correlation is only a liner measure.
2) Unconditional correlation is essentially a static measure, but dynamic relationships between markets may exist with a non-stable lead-lag nature.
3) Unconditional correlation only exists when the returns are jointly covariance stationary. In other words, correlation estimates jump around because they are being measured on non-jointly stationary series.
To summarize the above, correlation is not a stable measure. The observed instability (in todays pairs) is not (necessarily) due to the lack of joint stationarity, but is a result of non-linear relationships.
At the longer time frames, cointegration analysis allows for the dynamic co-dependencies in financial markets. In the shorter time frames, you probably need chaotic (attractor) type analysis to deal with the same non-linearity..."
I am developing a software that uses some of these concepts and I hope will allow me to get an edge not only on pairs, but on sector trading (which is a kind of flip side to pair analysis) as well. We'll see if I can couple the software that hones in on good pairs together with my skill in trading...
nitro
wow, good stuff, nitro.
Maybe one more question. Do you place price stops for pair trading or do you take time as a stop? Or don't you use stops at all with your pair trades?
Maybe one more question. Do you place price stops for pair trading or do you take time as a stop? Or don't you use stops at all with your pair trades?
Looks like the fake break scenario previously discussed might become a reality here. It is textbook for large operators to stimulate a breakout (especially given sentiment, it's so *easy*), sell into it, and sell the rest of the way down during the seasonally weak period. We've come back under the breakout level (1015) on the S&P on a strong move, and barring any kind of real bull retaliation over the next session or few, it might be a prudent choice to consider hedging, exiting, or even better, shorting the index or your favorite weaklings. I'm waiting for the bull's to make a stand, and if it's weak, I'm going for size and initial target will be around the 960-965 area (bottom of previous range). Any specific stock idea's or other analysis is always welcome!
Goodluck,
-TD80
Goodluck,
-TD80