Pair Trading Strategy Journal

Quote from sysre:

This is the equity curve of the pair strategy I backtested. It assumes MOC with 1/2 cent commission per share. It consists of 25 pairs, most of which I got from this thread.

The backtest is, of course, cherry picked - I only chose pairs that had consistent equity curves over the backtest period.

I guess when I go live I'll be dropping out pairs if the equity curve of that pair turns down (maybe 2 stdev below the regression line of the equity curve). And adding pairs in a similar fashion.
More...

Thank you for this. It is excellent.

I'd be interested to know which pairs you are using, and happy to discuss this with you off this thread. No doubt you are aware of the dangers of cherry picking and then extrapolating future performance from that. If your results are consistent with mine (and your graph suggests that) than the last few months would have been flat to slightly negative?

I'm running a larger universe and trying to use a lot of discretion over which ones I trade, so that what I do is underpinned more by the fundamental rationale than a purely statistical approach. This involves including some losing historical trades that I believe make sense from a fundamental perspective.
 
Quote from cipherscribe:

Walt,

I have been going over some of your comments, looking for the comment about the % difference between the ratio and the average ratio. I think this is key for me, as the difference between these two represent the maximum profit for the given trade.

Adrian
More...

Hi Adrian.

Here is the quote from my post earlier this year:

==========================================
To calculate the profit from the change in ratios I think this formula will work.

The (% profit) = (% change in the ratios)/2

The following is from my spread sheet on recent trades:
Example 1: The ratio moves up 9.3%, the profit is 4.7%
Example 2: The ratio moves up 23%, the profit is 12.1%
Example 3: The ratio moves up 13.2%, the profit is 6.9%
===========================================

So, I'll give you my formulas. Here is the Excel spreadsheet formula I use to quickly determine if I'm interested in a pair:
=((B3-A3)/A3)/2
A3 is the Ratio today empty field.
B3 is the Mean Ratio today empty field.

In other words, if the Ratio today moved over the next few days all the way to the Mean Ratio (converged), the formula will give a good approximation of profit.

However, I try to be more conservative and look instead at a point 1/2 way between the Ratio today and the Mean Ratio today as a more realistic profit potential. I expect the convergence to be 1/2 way. The formula for that is:
=((B3-A3)/A3)/4

This spring I was limiting my pair choice to only those with a potential greater than 6%, or 1/2 way at 3%. So, I conservatively anticipated getting 3%. And I was averaging 3% a week. I was one happy camper. Using the rule of 72's, that meant my money would double in 12 weeks! And again 12 weeks later. A millionaire very soon!

This summer I need to drop my profit requirement to 2%. My return dropped, and I had fewer pairs that qualified. Then in June, July, August, I've made little.

I'm now gun shy. I don't know what happened.

Except, I know I'm not a millionaire. YET.

Walt B
 
I did a quick search and after some filtering I had just one pick

VGR/LO They are in a support / resistance levels, ratio RSI is at 30, also if you look in different time frames (6,3,2,1 months) they have been crossing each other several times and in all time frames, LO is above VGR in all 4 charts.
 

Attachments

Quote from waltbx:

Hi Adrian.

Here is the quote from my post earlier this year:

==========================================
To calculate the profit from the change in ratios I think this formula will work.

The (% profit) = (% change in the ratios)/2

The following is from my spread sheet on recent trades:
Example 1: The ratio moves up 9.3%, the profit is 4.7%
Example 2: The ratio moves up 23%, the profit is 12.1%
Example 3: The ratio moves up 13.2%, the profit is 6.9%
===========================================

So, I'll give you my formulas. Here is the Excel spreadsheet formula I use to quickly determine if I'm interested in a pair:
=((B3-A3)/A3)/2
A3 is the Ratio today empty field.
B3 is the Mean Ratio today empty field.

In other words, if the Ratio today moved over the next few days all the way to the Mean Ratio (converged), the formula will give a good approximation of profit.

However, I try to be more conservative and look instead at a point 1/2 way between the Ratio today and the Mean Ratio today as a more realistic profit potential. I expect the convergence to be 1/2 way. The formula for that is:
=((B3-A3)/A3)/4

This spring I was limiting my pair choice to only those with a potential greater than 6%, or 1/2 way at 3%. So, I conservatively anticipated getting 3%. And I was averaging 3% a week. I was one happy camper. Using the rule of 72's, that meant my money would double in 12 weeks! And again 12 weeks later. A millionaire very soon!

This summer I need to drop my profit requirement to 2%. My return dropped, and I had fewer pairs that qualified. Then in June, July, August, I've made little.

I'm now gun shy. I don't know what happened.

Except, I know I'm not a millionaire. YET.

Walt B
More...
Walt, my experience has been similar to yours . That is, profits were much higher earlier in the year than they have been recently.

Judging from the lower volume of messages on this thread, I would guess that many other pair traders have experienced a similar drop in profits.

The reason for this drop off may be the recent lower volatility. Earlier this year stocks were moving in crazy frenzies and reversals in pair relationships were frequent. Nowadays trends are more steady and reversals to the mean seem less common.

I wonder if pair trading now would produce better results if a "follow the trend" strategy was used instead of a "reversal to the mean" strategy.

Don
 
Quote from Don87109:

Walt, my experience has been similar to yours . That is, profits were much higher earlier in the year than they have been recently.

Judging from the lower volume of messages on this thread, I would guess that many other pair traders have experienced a similar drop in profits.

The reason for this drop off may be the recent lower volatility. Earlier this year stocks were moving in crazy frenzies and reversals in pair relationships were frequent. Nowadays trends are more steady and reversals to the mean seem less common.

I wonder if pair trading now would produce better results if a "follow the trend" strategy was used instead of a "reversal to the mean" strategy.

Don
More...

I think you hit the nail on the head, Don. Profits have been harder to come by in pairs the last few months, and the trending nature of the market during that time is one of the likely culprits. One way I have been trying to combat it is not only doing the 'traditional pair trades', but looking to buy pullbacks in trending pairs. I've also started using more ETF's against stocks, with modest success thus far.
 
Quote from bentedges:

I think you hit the nail on the head, Don. Profits have been harder to come by in pairs the last few months, and the trending nature of the market during that time is one of the likely culprits. One way I have been trying to combat it is not only doing the 'traditional pair trades', but looking to buy pullbacks in trending pairs. I've also started using more ETF's against stocks, with modest success thus far.
More...

Applying a momentum strategy in addition to mean reversion should be a good hedge.
 
Hey all. Been away from the board for awhile, but I thought I would try to rejuvanate with some new thoughts and trades that are killing it.

These are not mean reversion pairs but negatively correlated and very trendy. Place these trades and just sit back.

I call this base line investing/trading. Really nothing more than tactical asset allocation.

TLT/VTI
TLT/EFA
TLT/EEM
TLT/IYR
TLT/GSG

You are always long both. However by creating a ratio and than graphing the ratio you can see which one to over weight. Use the 100 day moving average and then enter your trades when the ratio crosses through the moving average. Currently, the ratio is downtrending thus you want to be overweighted stock 2 and underweight the treasury. You'll see what I am talking about when you plot the graph. When the ratio trends up you are overweaight 1, underweight 2.

I always use a 80/20% or 20/80% allocation.

Currnetly I am 20% treasury and 80% the other asset classes.

The strategy works and stems from the CAPM model. And you will always be on the right side of the market based on the simple fact that you are trading with the markets perception of risk/reward versus the risk free rate.

Good luck and hope this adds some new life here. Any questions welcome.

yobo
 
Quote from yobo:

Hey all. Been away from the board for awhile, but I thought I would try to rejuvanate with some new thoughts and trades that are killing it.

These are not mean reversion pairs but negatively correlated and very trendy. Place these trades and just sit back.

I call this base line investing/trading. Really nothing more than tactical asset allocation.

TLT/VTI
TLT/EFA
TLT/EEM
TLT/IYR
TLT/GSG

You are always long both. However by creating a ratio and than graphing the ratio you can see which one to over weight. Use the 100 day moving average and then enter your trades when the ratio crosses through the moving average. Currently, the ratio is downtrending thus you want to be overweighted stock 2 and underweight the treasury. You'll see what I am talking about when you plot the graph. When the ratio trends up you are overweaight 1, underweight 2.

I always use a 80/20% or 20/80% allocation.

Currnetly I am 20% treasury and 80% the other asset classes.

The strategy works and stems from the CAPM model. And you will always be on the right side of the market based on the simple fact that you are trading with the markets perception of risk/reward versus the risk free rate.

Good luck and hope this adds some new life here. Any questions welcome.

yobo
More...

YOBO, I don't understand why you don't like to go short, these days US bonds are not safe any more, gold and oil are much better so if you are long both market and bond they can be cut in 50% and still the ratio is the same. Look at the graph I used the EMA 25 and MA100, to detect the trend, it look safer to me.
 

Attachments

Quote from magon:

YOBO, I don't understand why you don't like to go short, these days US bonds are not safe any more, gold and oil are much better so if you are long both market and bond they can be cut in 50% and still the ratio is the same. Look at the graph I used the EMA 25 and MA100, to detect the trend, it look safer to me.
More...

The reason I do not short the TLT is because first it is negatively correlated to the asset classes thus by its very nature it provides the hedge. Second the TLT pays out its dividend monthly. If you were short, you would have to pay the dividend versus collecting it. You get paid to hold it and it provides a hedge without shorting it.

Its true the correlation is not perfectly negative, so you can have days when both go down, but in the aggregate you don't need to worry.

Also, in regards to oil and gold, I am sorry I did not include the the GLD, but you could do that as well as part of the portfolio.

Essentially you could baseline any asset class or sector against the TLT to understand which one you should be overweighted. The premise is the same. rational investors will seek out riskier assets for investment if the risk reward ratio is greater than the risk free rate which is generally accepted to be the 30 year treasury. Thus simply trade the direction of the ratio and weight accordingly. You can set the weighting to anything you want. I like the 80/20 allocation because it protects me on the big surprising days.

For example watch how the TLT trades today against the SPY,GLD, VTI, IYR, EEM, EFA and any other sector or class you like.
 
You must log in or register to reply here.
Back
Top