Quants Without Context

I love to read the material from two excellent quant services to which I subscribe:

Since I am a visual trader i.e. I trade relying on my ‘understanding of the structure of the market’ and verified chart patterns, I find well researched, digital content very useful. On many occasions, their information has provided either an angle I had not considered or a pattern I could use as a setup.

But in this environment i.e. a rising probability of a full fledged 12-month (yearly trend) bear market, we need to exercise care with quant studies. The reason is this: statistical studies need to compare apples with apples i.e. in using past studies to provide a statistical edge, we need to ensure that the past sample is at least similar to the present environment.

So here’s the point.

If we are in a full fledged 12-month bear market, we need to go back to all the way back to the 1929 to 1932 for a comparable environment; but, most data bases the quants use do not go that far back. From what I have seen, most go no farther back than 1982 (when the most recentĀ  12-Month Bull Market began). Occasionally some go back to the 1966 to 1982 era. The problem is even this bear was not of the same size as the one we are probably in.

So what does this mean for quant studies (at least as far as I am concerned)? I’ll take oversold readings with a grain of salt, as I will with buy recommendations. And, most importantly, I’ll use the studies as a way of avoiding myopia for a bullish case.

Figure 1 is a monthly chart of the DJIA since 1885 courtesy of The ChartStore.

12-m-djia-1884-to-2009.jpg

Figure 1 DJIA Monthly

7 thoughts on “Quants Without Context”

  1. Ray

    On another note:

    Airtime: Thurs. Mar. 5 2009 | 09 13 00 ET

    The most important thing legislators can do is make past bad practices illegal, says Rep. Barney Frank, (D-MA) House Financial Services chairman.

    http://www.cnbc.com/id/15840232?video=1052999377&play=1?__source=CNBC|newsnow|vid4|2009|

    On Sentimentrader:

    Quote

    One of the worst parts about these junctures is that it makes you constantly wonder “is this it?”, “is this it?” as we cascade lower, and it can result in significant losses. The three ways to minimize that potential are to 1) wait for true extremes, 2) use small position sizes and/or 3) wait for confirmation.
    Unquote.

  2. Position size is a good idea. What is confirmation? Can you tell me? Also, what is “true” extremes? Hindsite is a beautiful thing.

  3. Hi Lance

    Thanks for your post.

    You do Ana an injustice by implying she is identifying ‘true extreme’ and ‘conviction’ with hindsight.

    In her comment, she is stating only the conclusions she has found to be true.

    Take a moment to read some of the archives to see how I arrive at these projections.

    You’ll see that Barrometrics seeks to project possible extremes by: a) statistical means and standard deviations; and b) some ratios. It does this for both time and price.

    It then looks for a conviction bar to confirm the projections. As to what is a conviction bar, it is again in the archives.

    Can either or both projections be wrong? Of course, that’s the nature of trading and the reason for exit strategies.

    But for the same reason, both conclusions can be right.

    All the best with your trading.

  4. Hi Ray,

    Thanks for your response. I will do as you said and look through your archives. Though, I wasn’t implying that people cannot look at certain things to try and gain more confidence. I suppose I was saying that with the nature of the markets and how they are, 100% conviction is an illusion. But I suppose this goes without saying in most things in life.

  5. Hi Lance

    I couldn’t agree more with you. There is no 100% with markets and life – except with hindsight.

    I’ll do a S&P analysis focusing on the stats, ratios and conviction bars.

    Take care.

  6. Here in lies the crossover of analysis and psychology.
    Looking for probabilities, whilst accepting all possibilities.
    And I suppose that could be widened to include, not just the probability of a future event, but also the probability of the significance of a past event. IE, was that a significant bottom? If probably yes, then you may have a probability edge to ‘lean against”.
    mmmm, not sure if that sounded a bit Dick Chany-esque?
    Ciao
    Stuart

  7. Hi Stuart

    Know exactly what you mean. Hence the Barros Swings. All things being equal, the higher time frame swing extreme is more important than the lower time frame.

    You can’t have that sort of distinction using traditional indicators.

Leave a Reply

Your email address will not be published. Required fields are marked *