Skip to Menu Skip to Content Skip to Footer
  • Banner
Join Mailing List:

Show Search

Friday, Oct 24th

Attention: open in a new window. PDFPrintE-mail

Last Updated (Tuesday, 24 January 2012 00:47) Written by Carl Weiss Wednesday, 22 September 2010 22:07

WIND [Weiss Inter-Market Neutrality Detector]

The genesis behind developing WIND percolated up from a question that had been nagging at me for quite some time.

The question arose from an understanding of market behavior that I had absorbed that comes from designing thousands of trading systems.

One common output of just about any trading system that has ever been developed is that if you run a system over a long enough period of time, you will find slugs of time, when the system had a great winning streak.

Hence one of the banes of the automated trading industry is that folks find themselves dealing with two re-occurring phenomenon:

1.    They build a winning system in the lab only to find out that once they implement the system with real money, it falls apart.  This is due to either to the timeframe of their back-testing sampling being randomly lucky, or they built a system predicated on what they have recently been witnessing in the market.   What happens in these instances is that many traders witness the same events occurring in the market and build similar to take advantage of these recent events.  The wide-scale implementation of these systems tends to eradicate the kind of behavior that they were intended to profit from.

2.    They subscribe to a trading system that is on a winning streak only to find it fall apart right around the time they invest in the system.  Can you figure out why this happens?

What stands out as a good recent example of this are the winning systems that traded the long side of silver in the beginning of 2011.  These systems did very well. Do you think that this is this because the system was so brilliant or rather because silver went up in price?

With this lesson burned into my head, I wondered if instead of building automated trading strategies that made money in any market, why not just build some algos that could tell you what type of market you were in, and then, simply take a strategy off the shelf that did well in this type of market and let her rip?

For instance, if you knew that the market was going to go up today, you would dust off your ‘market is going up today’ strategy and then click the ‘On’ button.

On the other hand, if the market was going to go down, you would implement your ‘market is going down today’ strategy.

What this boiled down to was if you knew which way the market WIND was blowing you would know how to set your trading sails for the day.

Hence we developed, WIND.

The first thing that WIND does is to take in a lot of information across many, many markets.   It then focusses on one behavioral aspect across all of these markets and asks the following question:

Is there harmony, or is there discord across all of the order flows across all these markets?

If there is harmony then it should be pretty easy to decipher which direction are all of these trading computers are heading in….they will create their own WIND.

HFTs, Automated Trading Programs, Trade Desks herd just like animals herd and people move along in crowds, hence it is not uncommon for all of these disparate groups to be synchronized.

Furthermore, it is not uncommon for this synchronization, this widespread trading ‘intent’ to be detected in the pre-market.

Which direction will the WIND be blowing is what we want to know as we are drinking our morning coffee. To the Buy side?  To the Sell side?  Or are there cross breezes that will make for choppy trading?

Either way, as long as we know what way the WIND is blowing, we can now implement our trading strategy for the day, or until the WIND shifts.

Here is an approach to using WIND to trade S&P 500 Binary Options.

...sceeto's WIND - Developed by Carl Weiss of Algo Futures