I hope that people are still using this strategy successfully.
Can we have some updates ?
Yes still beavering away. May was a great month for me but June so far has been very poor. The market has been choppy and volatile so opportunities do creep up often, but I am finding convergences harder to come by. The beaten down stocks just don't bounce back up easily anymore.
Yes still beavering away. May was a great month for me but June so far has been very poor. The market has been choppy and volatile so opportunities do creep up often, but I am finding convergences harder to come by. The beaten down stocks just don't bounce back up easily anymore.
Thanks for the reply,
Are you still using v2.99 or have you upgraded to V3.0 with cointegration ?
If you have, are there any +/- changes in your returns ?
Cheers
Sri
A great week last week made back half June's loss. When the market turns one way or the other the account usually gets a boost... it is when oversold shares finally converge with their stronger counterparts.
I've upgraded to v3.0. I have re-jigged my database and culled some pairs while added a few new ones. It's too early to tell the impact of the P&L, and I am not using a strict cointegration filter anyway.
One thing you need to watch with cointegration is that a choppy share will show high cointegration with any other share, regardless of whether they are truely 'fundamentally cointegrated' or not. It's up to the trader to decide whether they want true cointegration or just mathsmatical cointegration.
It's merely a statiscal property involving 2 or more time series (http://en.wikipedia.org/wiki/Cointegration). The main (if not the only) reason traders look at cointegration is its stationarity (or mean reverting), if things divert enough in a stationary process, it WILL revert to its mean (but the exact timing of when is not known). It has absolutely nothing to do with correlation, which is another statistical concept.Is'nt cointegration just decorrelation over a shorter period with correlation over a longer period?
I think CBA versus WBC would be my pick. 2 WBC shares to 1 CBA shares would be approximately a neutral spread. Fundamental equals. Maximum divergence 15% over the last 15 years.
Holding both and selling options on them would smooth the equity curve somewhat too.
Cheers!
A great week last week made back half June's loss. When the market turns one way or the other the account usually gets a boost... it is when oversold shares finally converge with their stronger counterparts.
Thanks for the prompt reply.
What do you check to see that the Pair is fundamentally cointegrated?
Do you look for fundamental data like PE etc or the business?
Cheers
Sri
It's merely a statiscal property involving 2 or more time series (http://en.wikipedia.org/wiki/Cointegration). The main (if not the only) reason traders look at cointegration is its stationarity (or mean reverting), if things divert enough in a stationary process, it WILL revert to its mean (but the exact timing of when is not known). It has absolutely nothing to do with correlation, which is another statistical concept.
Generally speaking the big four banks are good candidates for pair trading, but because they are good, they often not diverge enough to generate good profit after brokerage costs. (And more annoyingly, one bad trade will often erode 3 - 5 profitable trades)
It will be interesting to see how people go with the extra "cointegration" feature in PTF 3.0.
Personally I have some reservations about how cointegration can be expressed in a number between 0 to 1 (perhaps 1 - it is cointegrated, and 0 - it is not cointegrated?). Nothing offensive and it may work out brilliantly, but just don't feel like using something I don't understand...
It's merely a statiscal property involving 2 or more time series (http://en.wikipedia.org/wiki/Cointegration). The main (if not the only) reason traders look at cointegration is its stationarity (or mean reverting), if things divert enough in a stationary process, it WILL revert to its mean (but the exact timing of when is not known). It has absolutely nothing to do with correlation, which is another statistical concept.
Generally speaking the big four banks are good candidates for pair trading, but because they are good, they often not diverge enough to generate good profit after brokerage costs. (And more annoyingly, one bad trade will often erode 3 - 5 profitable trades)
It will be interesting to see how people go with the extra "cointegration" feature in PTF 3.0.
Personally I have some reservations about how cointegration can be expressed in a number between 0 to 1 (perhaps 1 - it is cointegrated, and 0 - it is not cointegrated?). Nothing offensive and it may work out brilliantly, but just don't feel like using something I don't understand...
As a game here are the ratio charts of 6 pairs... would people like to guess what the ballpark co-integration value is for each pair according to PTF software?
View attachment 43557
I shall reveal the answer in a few days
Do I get to know the names of the pairs as the prize? I'm drooling over pairs 1, 4 and 5, so they must have a damn high score
I had a short on MCC yesterday and I copped it sweet today.
Never would I have thought that someone would launch a takeover for a coal stock one day after the carbon tax is announced. In fact, yesterday I noticed surprise strength in WHC and NHC - so I avoided shorting them (in case something was up), and chose the beaten down MCC instead. Unbelievable...
This is the first takeover I ran into after over 1000 trades. Thankfully the position was <5% of the account, so the 35% jump today only caused me ~1.6% damage. Not the worst percentage wise, but the worst in terms of $$ lost as I just doubled the capital in my account for this financial year.
Oh well. I guess if you try to pick up pennies long enough in a doggie park you will step onto dog poo eventually.
This dog poo definitely stinks as hell, I guess no one but Peabody & ArcelorMittal could really see it coming in this sort of timing. Glad you have proper capital management in place so your account doesn't blow up otherwise.
Two interesting stocks today:
1. Profit downgrade from DJS today making it plunged 15%, I'm closely watching this one cuz I really don't see fundamentally MYR can be any better if DJS is doing bad in sales
2. FXJ, some heavy selling is happening right now, for no apparent news (still researching), might try my luck if I really find nothing...
DJS and MYR are forecasting the same NPAT now. But DJS is worth ~$1.7B vs MYR ~$1.45B.
Saw the FXJ plunge but by the time the PTF alert came through it's already moved back up... that's why we want live data!!
We use cookies and similar technologies for the following purposes:
Do you accept cookies and these technologies?
We use cookies and similar technologies for the following purposes:
Do you accept cookies and these technologies?