This is a mobile optimized page that loads fast, if you want to load the real page, click this text.

ASX Stock Pairs Trade Journal

Jump to new

My :
1. Statistically the two are two different properties of time series analysis
2. In terms of pair trading, the key difference is that mean reversion is part of the co-integration definition, while with correlation you are relying on the back test result as a proxy of the statistical test for mean reversion
3. Co-integration pair trading is a much more computation intensive process than correlation + back test
4. The Correlation + back test approach (PTF) trades dollar neutral, co-integration will depend on the co-integration coefficient, and more often than not it won't be

Some simple ideas behind co-integration pair trading:
1. Find the co-integration coefficient
2. Find the co-integration period
3. Test for stationarity (mean reversion)
4. Once a co-integrated pair is found, make sure the stationarity holds over time

1. is rather simple (a regression exercise) and it's really 2 and 3 (and somewhat 4) that is time consuming and repetitive, and even with a decent computer it might take a good while to do for just one pair. I think that would make it difficult to throw a few hundreds of pairs at it as PTF.
 

SilverRanger has given a much better response than I could. But my thoughts are along the same lines.

To me using correlation and backtest as filter is a pratical way to get around the co-integration maths.
 
Hi all,
This is my first post, been following this thread a lot,
Just thought I'd mention that because cointegration is based on probability there is no single one way to calculate it, and as said earlier the more accurate for it to be the more computing power it would need. Also it's probably not included as they don't want to 'garauntee' a win of sorts (just guessing here)

Oh and while I'm here, has anyone tried to program their own pair trade finding software? Ive noticed on the internet the equations for correlation and cointegration have been written out for some maths programs and are freely available (by that I mean legally freely available). Any thoughts?
 
Market changes are coming http://www.theaustralian.com.au/bus...trading-targeted/story-e6frg8zx-1225937363096 high frequency trading will dramatically increase over the next 2-3 years on the ASX from the current 20% levels of volume probably more closer to 70% like the US markets I think, because of new legislation that will open up the market to more competing exchanges and hedge funds, next year will probably see the introduction of chi-x which will create segmentation in market liquidity, and hopefully brokers can create smart routing options to collate the bids/offers together into one que for us or something like the NBBO in the US.

It's important as a trader to be aware of these changes, because it will affect how stocks trade, how the tape looks, hopefully the ASX forces time on que rules so we don't see those fleecing bids/offers that the US market sees, some stocks over there can have 3,000 bids/offers posted and pulled in under 1 second, not something we want here but let's see how it goes.

Until then the ASX still offers potent alpha from a relatively in-efficient market.
 

Yes, I have been working on this some time ago, basically trying to trade the pairs with low or even negative correlations. The calculations I used were completely manual (hence inefficient), but I can see this could be a lot easier if you are coding in stat packages like Matlab. And FYI, I couldn't prove the stationarity of this pair (albeit getting close in p-value) and the entry and exit signals I got were similar to the ones in PTF.
 
skc said:
PDN/ERA had good correlation a while back, and they are the only 2 uranium producers on the ASX. So I am happy to trade them when things look right.
Click to expand...

I actually entered this pair last week and had a pretty wild ride - up 4%, then down 3.5%. But with the ERA poor qtrly report today I end up having a reasonably profitable trade. Only got ~50% of the full profit available - got most of the ERA fall but none of the PDN gain today. I was expecting negative sentiment on ERA to drag PDN down through the day - but it turned out that money chasing uranium ended up buying PDN instead... Should have just waited to square off the position at the close.

Not sure if this is an example of co-integration or not!


Thanks for the update. The high frequency stuff in US is really quite crazy. I hope it's not going to come to Australia. Having said that, the ASX has some pretty poor spreads / liquidity for even some pretty large companies.

And I wouldn't hold my breath on Aussie brokers offering smart routing... the major online brokers (like Comsec) have very basic web based, non-dynamic platforms that are not designed for optimal trading. Iress platform is more respectable, but I don't expect them to spend money on development until chi-x prove itself with sufficient volume.
 
And PDN keeps going up, while ERA keeps falling.

At current price ratio it is 22% above my entry ratio!!! And how much profit did I bank? 4%

Just have to shake my head and laugh sometimes.
 
skc said:
And PDN keeps going up, while ERA keeps falling.

At current price ratio it is 22% above my entry ratio!!! And how much profit did I bank? 4%

Just have to shake my head and laugh sometimes.
Click to expand...

you will always kick yourself sometimes for missed opportunities, just think of all the trades you closed that went into deep loss after you exited the position, you saved $$$ on those....
 
Pairs Trader said:
you will always kick yourself sometimes for missed opportunities, just think of all the trades you closed that went into deep loss after you exited the position, you saved $$$ on those....
Click to expand...

Indeed. I always take the ticker off my watchlist after I closed a trade. It's like not looking at price catelogues soon after you bought a new TV .

Closed a nice trade long PTM short IAG. Thanks KKR for the bid on PPT. Trailing a small PTM just to see how far it can run.

With the program, PPT/PTM shows poor correlation - but ultimately they do the same thing and when one rises when the other is under a takeover offer, you know there is some underlying correlation / co-integration happening.

Opened to day IFL/AMP. IFL has potential takeover appeal, while AMP has potential cap raising appeal... let's see how it works out.
 

I tend to keep the closed pair in watchlist for a day or two, since I do take early exits from time to time.

Closed DXS/ABP, ABP was a pain in the ar$e when it comes to entry and exit, I could have got out 2 days earlier with similar profit. And given all the reverse splits happening lately, I will probably think twice about entering pairs with both legs sub $1.

Opened today:
BLY/AIO

On my watchlist:
AAX/TPI, this one came up late in the afternoon, didn't have time to research on the AAX sell-off today, so I gave that a miss
 

This is how you miss the boat 6%!!
 

That's some pretty volatile shares you have there?!

How do you do your position sizing with these more volatile pairs?

I would probably do it slightly differently... e.g say the pair can go 20% out of whack and I want that to mean no more than 2% of my account. That way each leg is max 10% of account.

Actually entered AAX yesterday and didn't get the best entry. Bit of pain yesterday afternoon and this morning but the reversal candle formed now looks very promising.

Closed out REA/WTF today - well before REA spiked up. Only got the WTF spike down... Another profitable but poor exit in the books.
 

Attachments

  • 20101020 Ausenco Ltd (-).png
    18.3 KB · Views: 6

Must admit I try and avoid anything with a price below $ 1.25. Find it too volatile otherwise.

Not having fun this month...picked way too many losers for some reason and had some issues with stops being triggered as shares just fell into liquidity holes
 

How do I position size volatile pairs? I don't What I do instead is I apply more stringent criteria than normal when pulling the trigger and do more sanity checks. But that's just me willing to take more risk to get some boost in return.

I guess a more scientific way is to adjust the position size relative to the average beta of the pair, so if normally your leg is 10k for a pair with average beta 1, then simply take a 5k leg for your volatile pair with average beta 2.

Closed MGR/SGP at market close today, currently sitting on 6.7% for the month
 
skc said:
Oh no!!! You had it open before the announcement?!



I feel your pain... seriously. Hopefully that pain is eased a little bit today.
Click to expand...

Put it on about 2 hours before the announcement. The annoying thing of course was that I almost cetainly sold to someone who knew....

Sold IRE out yesterday and covered ASX ard $ 40. In the end a trade that was only about 50% more than my standard stop

Gotta have a sense of humour. In this case its the fact that ASX can't even keep its own discussions confidential
 

It would be funny to see a speeding ticket issued to itself as well, just pure unlucky I guess
 

Agree position sizing is always key

As for the avoid shorting the problem I have had with that is there are always rumours. Always hear TSE is a takeover target for UGL but have traded in and out of that on numerous occasions. Ironically have avoided shorting IRE as ASX is supposed to buy them!

The ultimate pain of course is that doesnt seem the bid will succed - at least in this form
 
You must log in or register to reply here.

Similar threads

T
Trender's Price and Volume trading
2
Replies
22
Views
2K
trender
T
XJO, Gear, Spi, Overwrites & CC's (no not the chips)
2
Replies
21
Views
3K
Rosscoe62
  • Sticky
March 2025 Stock Tipping Competition Entry Thread!
Replies
5
Views
443
farmerge
  • Sticky
February 2025 Stock Competition Entries!
3 4 5
Replies
80
Views
5K
frugal.rock
  • Locked
February 2025 Stock Tipping Competition Entry Thread!
2 3
Replies
44
Views
3K
farmerge
Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more...