Intraday Pair Trading/Stat Arb

[Trader13]

Quote from Shanb:

I just haven't seen alot of discussion about this on an intraday basis. Factor models used intraday will be alot different than on a longer tf given the difference in market participants that operate on that tf.

Also some nuance may lie in what factors to consider when creating the basket to trade against the ETF. I am noticing that the different Sector ETF's have some different allocation of holdings. While one sector ETF may work well with using the varying industires. Something like the XLE may work better by differentiating according to market cap and industry

The ETF market makers are probably executing arb strategies continuously intraday to keep the ETF aligned with its components. Of course, they will have market maker advantages. So that's tough competition.

There may be a brief opportunity at the session open when overnight gaps throw the ETF/components alignment out of whack until market liquidity and pricing settles down. But I imagine all this transpires in seconds in our electronic markets.

Sorry I don't have any better insights to contribute to your quest. It's an interesting topic which caught my attention.

 

[Bob111]

I don't have a quantitative background, so going through this stuff has been slow and grinding. Now I've got the basics down and need to gather everything for testing. What do you use for intraday testing. None of the basic software out there even deals with intraday data!

i don't have it either..i didn't even finish my high school..but hey..that's how it's goes pal..slow and grinding..

----A journey of a thousand miles begins with a single step---

regarding software-like you said-excel. vba..vb6..something flexible yet simple. or whatever you know already.
to make your journey less painful-ETF vs basket of stocks of same weight=dead end. as been said above-they been arbed to death...you are not going to find any inefficiencies at all. unless you do something more creative than just trade basket vs index fund

 

[Shanb]

Excel and VBA will have to do then!

Also I'm not looking to arb the individual components against the ETF. Its more like looking for a distinct group of stocks within an ETF that should move together. When this basket gets out of whack with the overall sector...play it against the broad ETF. Like Internet retailers against the XLY. These are all just ideas, we will see how it plays out...there may not be anything here.

 

[reid5525]

Quote from Shanb:

Excel and VBA will have to do then!

Also I'm not looking to arb the individual components against the ETF. Its more like looking for a distinct group of stocks within an ETF that should move together. When this basket gets out of whack with the overall sector...play it against the broad ETF. Like Internet retailers against the XLY. These are all just ideas, we will see how it plays out...there may not be anything here.

There is something there. But you will have to let your selected symbols and sectors play out each day until you find what matches up.

 

[Shanb]

Quote from reid5525:

There is something there. But you will have to let your selected symbols and sectors play out each day until you find what matches up.

Thanks, that is what I was looking for.

 

[reid5525]

Quote from Shanb:

Thanks, that is what I was looking for.

Placing symbols and sectors in QT portfolios with the inline mini charts open to intraday shows clearly which way everything is going.

 

[EEUT84]

Been trading pairs for four years now, have had both good and bad years. Let me give you a bit of advice:

1. Cointegration wont matter much intraday. High correlations (>90-95%) won't give enough opportunities to trade either.

2. Keep it simple when it comes to weighting. Feel free to regress, beta-weight, z-score, etc. You will likely find that dollar-neutral works best.

3. You are looking for intraday reversion. So look for trades that have the highest possibility of short-term reversion. Gaps, emotional moves, etc. Problem is that most emotion has been 'softened' by algo trades nowadays.

4. Your entire strategy will be dependent on managing your losers.

Good luck.

 

[reid5525]

There is opportunity for index arb ( subset of stat arb). A nubie will need time to arrive at high prob matching.

 

[Trader13]

Quote from Shanb:

I'm a trader at a prop firm and we have a group that runs a pair trading approach and does quite well. Some of their guidance and insights have led me to investigate this type of approach...

Can you share some of their insights? If they are doing well, why not just follow them?

 

[Shanb]

Quote from Trader13:

Can you share some of their insights? If they are doing well, why not just follow them?

Well, Its A bit hard to execute without large overnight BP and is dependent on good Knowledge of the companies they trade. Think smaller banks, reits, utilities.