ASX Stock Pairs Trade Journal

[TradePro]

I have been looking into pairing some of the Real Estate Investment stocks such as SGP, GPT, MGR, WDC etc. These stocks have historically shown high average correlation, around the 80% level for the last year.

This is what I am finding a bit strange that the current correlation between SGP and the rest (GPT,MGR,WDC) have gone down drastically. For example, current correlation between SGP and MGR is 36%, between SGP and GPT is 37% and between SGP and WDC is 30%

Does anyone know why this is?
And also does this mean the above pairs should not be traded given such low correlations?

 

[Sugar Dunkaton]

I am going to say without looking at anything that SGP, the common factor in all those low correlations, has probably had some significant news or a drastic change in sentiment towards it to warrant its fall out of sync with the others.

According to the rules I trade by, we would not trade stocks with a falling correlation given that it is that low. We normally try and stick to stocks >70%

 

[TradePro]

Do any of you use 'Percent From Mean' value for deciding which pair trades to take? For instance take a trade only if 'Percent From Mean' is greater than a set percentage, etc.

Does 'Percent From Mean' gives an idea of what might be the expected returns if a pair trade was taken?

 

[skc]

Do any of you use 'Percent From Mean' value for deciding which pair trades to take? For instance take a trade only if 'Percent From Mean' is greater than a set percentage, etc.

Does 'Percent From Mean' gives an idea of what might be the expected returns if a pair trade was taken?

The maths in PTF is actually quite simple. Percent from mean is the difference between current ratio of the two stocks (R_now) and the moving average of the ratio (R_MA).

It gives an indication of what the returns would be if the ratio is returned to the MA on the same day. For example:

Stock_short = 2.1
Stock_long = 0.95
R_now = 2.21
Say the R_MA is 2
The percent from mean is 2.21/2 -1 = ~10.5%

If the ratio was to return to the mean on this same day, you would have made 10.5% x size of one leg. E.g.

Cover stock_short at $2 (+5%)
Close stock_long at $1 (+5%)

In terms of taking the number into account. I usually take a look at the Percent from Mean chart and see if the current value is significant compared to the past. E.g. if a pair regularly goes to +/- 15%, but the current value is only 2%, then may be there is more divergence to come before the pair converges.

Note however that the R_MA changes every day so the a percent from mean can change quite quickly even if the stock prices are not changing in your favour.

 

[skc]

Thank you skc, that makes things clearer. I have watched charting using an 11 day exponential moving average so this is similar - my PTF is set to 150 days and 14 day lookback, is that normal?

My numbers are similar but can't tell you what is normal. Best ask guys at the website.

 

[TradePro]

Stocks affected by news are avoided when pair trading. When I check the news for many of the stocks in ASX100 that I want to pair, there seems to be quite a few ASX announcements. Are there any particular kind of news that needs to be avoided or any news at all for the stocks that are to be paired?

Also, how long would you wait for the effect of news to subside?

 

[n5032245]

Trade Pro,

you need to take the news on a case by case basis, apart from when results are release IMHO. i do not trade one week before results or one week after. full stop. happy to let trades go through that occur in this time frame.

as for other news, you need to determine for your self if there has been a material price movement in the stock resulting from the news (or built up to the news). some news will be clasified as price sensative, but won't materially move the market. so disregard.

remember that we are looking to make a profit from an statistical arbitrage situation where two similar stocks have diverged for no underlying reason and are predicted to revert back to the mean.

hope this helps.

 

[Trade to win]

Guys, no trades open for me at the moment as I'm traveling, thanks for keeping the thread alive and interesting, pleasure to read, will comment more on discussed topics when I have time.

Hi
Intresting software, but the problem is that I trade forex and Futurs, have you ever thought of upgrading your software to include these ?

As you may know there are many opportunities in forex and futurs in pairs trading and the amount of forex and futurs are much more worldwide, which they can use your software, because mostly people who trade US shares are living inside the US but the forex and future's market is global.

Thanks.

 

[samgribbles]

long NAB at 29.47, short ANZ at 23.33. SD 1.63, % 1.87

not > 2xSDs from SMA but % from mean is high. I will take this one on the open if it is as above or better and will add to it if the SDs from SMA gets above 2

 

[n5032245]

is pair trader going to start his live journal back up again?

 

[bochman]

re NAB & AMZ pairs
interesting, seem to hit 2 StdDev on a shorter time frame (45 & 60 days), and confirmed with 1 year chart.
May be a play.

 

[awg]

is pair trader going to start his live journal back up again?

There are posts from Tech/A and Nick Radge recently

the upshot is that unless you hold an AFSL then you cant post regular trading live on forums such as ASF..due to ASIC requirements.

This apparently included Tech and Rozella

A complete load of BS in my opinion, but the speculation was PT would also fall under this issue.

Comment PT?..mods?..Joe?

 

[Sugar Dunkaton]

thats a definite niggle, it was a very interesting journal to watch.

 

[Barndat]

Yes it was an interesting journal to read and I followed it up by downloading the trial prg. Now however it would seem that interest has moved on to the next bright trading idea whatever that is. I'm disappointed as I thought the prg had great potential. The PT dedicated forum is also completley dead...

My trial will expire today and I will be none the wiser....

 

[samgribbles]

closed NAB/ANZ trade that was opened 22 Sep
went against me straight away so doubled up

15 Sep opened / 17 Sep close
long NAB 28.420 / 29.54 = +3.9%
short ANZ 22.190 / 22.85 = +3.0%
=> profit = 0.9%

2nd trade:
long NAB 27.980 / 29.54 = +5.6%
short ANZ 22.030 / 22.845 = +3.7%
=> profit = 1.9%

1st trade w/leverage, return net of fees = 5.8% ROC
2nd trade = 14.7% ROC

(apols for late post, forgot my password)

I see PT has disappeared, I'll try and post my trades as I enter

 

[samgribbles]

FYI for those interested in the strategy. It does work.

On live trades, my IRR since commencement of pairs trading strategy (1 Jul 09) is 120%pa from 24 trades (19 winners, 2 x breakeven (ROC<1%), 3 x losers).

Using real $ but only small bikkies still. If I can keep the IRR above 100% after 50 trades I'll significantly increase my capital at risk.

 

[Sugar Dunkaton]

I know it is irrelevant to the thread, but I am fairly sure that IRR is not the term you are looking for. IRR is usually used for budgeting and planning for capital projects and their feasibility. What you are referring to is your "rate of return [ROR]". Excuse the pettiness.

But yes, I have followed the system too, and I have found that it can make a solid return.

 

[samgribbles]

Sugar - I'm always eager to improve approach. You'll have to explain the difference to me though; I'm pretty sure IRR is approriate for average return over time with capital going in and out over time

 

[bochman]

yes, it does work.
However, please bare in mind that it is all depending on the pair you are choosing, you have to be really careful about news, etc as well.

As in anything, you do need time to practice before going crazy.

 

[samgribbles]

29 Sep.
long NAB at 31.21, short CBA at 52.10