The ACD Method

Quote from Lucias:

Thanks for link. I think your explanation is poor though.

There are typically 2 ways to trade spreads: relative strength (you expect one to outperform) or mean reversion (expect spread to collapse). If you play relative strength you buy A hoping it continues to outperform B. If you play mean reversion then you sell the stronger and buy weaker.

The basic idea behind spreads for mean reversion is the law of one price. Basically, similar products are interchangeable. In terms of relative strength, there are many ideas behind that too. One might might look for fundamental reasons, for example, might think company A is stronger then B and will outperform. There are many books on spreads at Amazon.

What Maverick has spoke about in spreads (in relation to me and others), is how proprietary firms will trade spreads. Often for example spreading a cash market with futures. This could be for arbitrage or something longer term. You might for example create a basket of stocks that you think will outperform the indices but you're making a relative bet, so you buy stocks and sell the futures.. Or you could spread a future against a future.. creates a calendar spread.

Typically for pairs for mean reversion.. you want them to have some correlation, a high correlation to be exact. You want a high correlation because it implies the stocks move together. A high correlation isn't enough though, they need to get "out of line" at times to make a profit. FOR relative strength spreads, you want an inverse correlation.

You have just described what most of the prop firms do, spread relative value. I think Mav is showing you can use acd to find pairs that are outside the box. I don't disagree with your description of pair trading.
 
Here is a chart of spy and uso ( the oil etf). Looks like the same trade to me or a correlation nightmare. You would give your risk manager a heart attack with a position like this.
 

Attachments

Quote from kinggyppo:

Here is a chart of spy and uso ( the oil etf). Looks like the same trade to me or a correlation nightmare. You would give your risk manager a heart attack with a position like this.

What do you mean? That's 1000 basis pts!!!! That spread is huge! LOL. True correlation should never spread 1000 bps. Love that spread.
 
Quote from Maverick74:

Thank God nobody on ET is discussing the Yen breakdown right now. Only the biggest move in probably 5 years across that spectrum. :)


that's a moonshot on the daily wow!
 
Back
Top