Odds Czar: Simple Biases in the Futures Markets 2006

A

Art Collins

We're going to start a new thread for 2006, because I'm making a few changes to the biases and their presentation. The next posting is the explanation of the biases for 2006. The original thread is here:

http://www.elitetrader.com/vb/showthread.php?s=&threadid=57164

Also, I'll repost my original introduction for new readers.

I'm Art Collins, a published book author, lecturer, and a frequent contributor to Futures Magazine. I'm also a 17-year Chicago Board of Trade member, professional futures trader and mechanical system advocate.

I'm here to share system ideas and for camaraderie, something I miss from my days on the floor. I've decided to keep a daily journal on Elite Trader as a way to generate more feedback on my system ideas. Each day, I will post my daily market biases to keep the discussion going.

Here are some of my core beliefs. Feel free to add feedback, call foul on me, or whatever.
1. The average trader can't succeed unless he or she is 100% mechanical. Most of us can not "hear what the markets are telling us" because for all practical purposes, the markets ain't saying squat.
2. Human psychology tends to be drastically out of synch with what is needed to trade successfully.
3. When one combines mechanical with discretionary, one tends to get the worst of both worlds.
4. Simple is best.
5. Basic elements can be combined to create greater wholes.
6. Day trading can work, but there are inherent problems compared to other types of trading. The main problem lies in the relatively small trading arcs compared to trading costs.
7. Ideas have to test well over a fairly wide array of times and markets in order to be considered trustworthy.
8. When you do decide to trade a mechanical system, you must adequately budget for it. You must also follow all signals exactly as mandated.

How do I go about building mechanical systems? I identify simple market biases, some of which are not good enough to overcome trading edges. By assigning them +1/-1 strength/weakness values, however, and combining them with other such indicators, you can often get a whole significantly greater than the sum of the parts. I will be showing this on a step-by-step basis, using TradeStation performance summaries to substantiate my contentions.

Undoubtedly, my programming skills are short of the levels many of you are at. I intend to keep expanding my abilities rather than justify my limitations. Just to reiterate though, I’ve been privy to some very esoteric complex systems and some very basic ones. The lion's share of profits has come from the latter, at least in my humble experience.

Art
More...
 
In line with the yearly rollover, the Daily CzarCharts will be presented in a somewhat modified fashion. There are now three sets of biases.

Either-Or Biases

The first set of biases includes six indicators that individually signal either long or short on a daily basis, except for the rare tie. Each indicator has a +1 value for long bias, and a -1 for short. The bottom line is the sum total, which can range from -6 to 6. Positive totals are bullish; negative are bearish.

The six either-or biases are listed below. You’ll recognize the first five as part of last year’s eight biases. All biases are generated after the market close for the next day’s trading. For bullish signals (opposite is bearish):

  1. The 2-day average is below the 5-day average.
  2. The close is above the 40-day average.
  3. The highest close of the last 50 days occurs before the lowest close of the last 50 days.
  4. The day’s trading range is smaller than the 10-day average range and the day’s close is higher than the 10-day average close OR the day’s range is larger than the 10-day average range and the close is lower than the 10-day average close.
  5. The close is above the midpoint of the average 15-day range. (The 15-day high average plus the 15-day low average divided by 2.)
  6. Fade the majority direction of the last three open-to-closes.
    [/list=1]

    This last one is new, although we have discussed the concept. An open-to-close direction suggests a fade — either because it reverses more often than not or it produces bigger returns when it does or some combination thereof.

    Infrequent Biases

    The five infrequent biases are listed below. You’ll recognize them from last year. All biases are generated after the market close for the next day’s trading. For bullish signals (opposite is bearish):
    1. Four successively higher closes were followed by yesterday’s down close. Today’s action was irrelevant.
    2. Five successively lower closes were followed by today’s up close.
    3. CUP trade. For the last three trading days, the middle day had both the lowest low and the lowest close. In addition, the low on the middle day must also be lower than the lows from the previous three trading days before the middle day. (CAP is the reverse and bearish.)
    4. The highest low minus the lowest low of the last three days is less than or equal to 20% of the highest high minus the lowest low of the last three days.
    5. For the previous two days, the market closed lower than it opened.
      [/list=1]

      Calendar Biases

      The third box gives the calendar biases in the indexes. To review, days of the week produce a bias.

      Monday: Buy or sell in the direction of Friday’s close-to-close net change.
      Tuesday: Go opposite Tuesday’s close-to-close net change.
      Wednesday: Fade either Monday or Tuesday’s direction, depending on which move was larger. (Plus or minus).
      Thursday: Buy if the weekly high minus Wednesday’s close was greater than the close minus the weekly low (and vice versa).
      Friday: Fade the largest close-to-close move (plus or minus) of the last four days.

      Second, the day of the month can be broken down into two equally sized long-short biases. Specifically, the bias says to go short on the 7th day of the month and hold the short until the 22nd day of the month. On the 22nd day of the month, go long and hold the long until the following 7th day of the month.

      Third, the month of the year can be broken down into two equally sized long-short biases. This bias says to be long November-April, and short May-October. Specifically, the bias say to be long November 2-May 1, and then short from May 2-November 1. I’ve recently verified this while writing my technical indicator book, which will be released this fall. I say “verified” rather than “discovered,” because it incorporates some widespread beliefs about stocks. Stocks tend to do really bad in September and October. They also routinely experience malaise in summer months. The rest of the year, we should look for rallies, particularly in and around the year-end rollover.

      I had one ground rule in optimizing Month-of-the-Year long-short time-frames. The time-frames had to be equal — six months apiece. I’m sure we could get better results by making the long side bigger than the short, but that’s too much targeting for my tastes. As always, we want to mitigate the chances that we are merely uncovering randomness that bunched up a certain way. As with the Day-of-the-Month bias, the theory is that some times of the year tend to over-perform. This guarantees there will be some corresponding periodic slacking off or even give-backs of profits. It can’t hurt us to know whether our given longs or shorts are flowing with or fighting that prevailing trend.

      Art
 
Here are my futures biases for January 3.

A "1" means bullish bias. A "-1" means bearish bias. The total is the sum of biases. A positive sum will be long bias. A negative sum will be a short bias. A sum of zero will be a neutral bias.
 

Attachments

Hopefully by tomorrow, I should have the TradeStation summaries for all 14 biases available for viewing and discussion. If not tomorrow, it will be posted by the end of the week.

Art
 
Quote from Art Collins:

Hopefully by tomorrow, I should have the TradeStation summaries for all 14 biases available for viewing and discussion. If not tomorrow, it will be posted by the end of the week.

Art
More...

Art,

I missed it. How are you interpreting these Total Biases. For example, your gif table shows SPX Net Total Biases as o (-2+1+1).

What do you do with that information? How do you enter, exit? Stops?
 
Hey Art,

This would be a heck of a lot more interesting if we could see if these recommendations would actually make any money.

A P/L of trading one contract on your net bias would make this thread worth reading.

Cheers,
Fletch
 
fletch2 wrote
Hey Art,

This would be a heck of a lot more interesting if we could see if these recommendations would actually make any money.

A P/L of trading one contract on your net bias would make this thread worth reading.

Cheers,
Fletch
More...
you can. just look at the historical performance. from there, you're free to judge whether you think there's any validity to the idea that the trend is going to persist. i maintain i tested prudently and the biases are worth noting.
good luck for the new year
 
Here are my futures biases for January 4.

A "1" means bullish bias. A "-1" means bearish bias. The total is the sum of biases. A positive sum will be long bias. A negative sum will be a short bias. A sum of zero will be a neutral bias.
 

Attachments

You must log in or register to reply here.
Back
Top