Reflections On An Unfinished Experiment

I brought up The Seasonality Problem as a call for more objectivity. Being objective does not mean becoming obsessed with statistics (although numbers often help). Being objective means being obsessed with the truth.

Truth is an important thing, but it’s a tricky thing. Possible evidence of this “trickiness” is the reality that in over the centuries human civilization has lived on this planet, we’ve only figured out just about four ways to arrive at truth:

1. Authority – If someone we highly regard says something, we presume it to be true. Media and the Church enjoys a lot of this privilege.

2. Rationalization – We can use logical arguments to arrive at truth—or the process of deduction—like Sherlock Holmes.

3. Intuition – Occasionally, truth just reveals itself through a hunch, a feeling. This method is hard to verify, but it does exist. (As a wife discovers about her husband’s mistress, by choosing just on a hunch to arrive one day early from a vacation for instance).

4. Scientific Method – We can formulate hypotheses and take observations to disprove them.

All of the above methods are valid ways, and they are all prone to some degree of error. Scientific method in particular takes so much time, effort, and is very error prone that the benefits probably only marginally outweigh the costs. But it’s by far the most objective method there is.

One thing we should always distinguish about our search for truth is whether we are making a scientific argument or a moral one. One is a question on how and what things ARE, while the other is a question of how and what things SHOULD BE. Both are valid questions in the search for truth. But if one is mistaken for the other, that’s when conflict arises.

Let’s bring up the previous questions I raised and attempt to answer them in light of our data:

1. What phenomenon, condition, or factor should I be looking for?
The phenomenon is seasonality. Purchases made in October have a higher likelihood of success in the next 180 trading days than in any other month.

2. How do I spot it?
By checking your calendar date if it's October.

3. What happens after the phenomenon?
Studies have shown that purchases made in the month of October show an average return of 17% over the next 180 trading days, which outperforms similar period returns from purchases made in other months.

4. How reliable is the likelihood of the result following the phenomenon?
Studies have shown that 147 out of 151 days in the last 7 years or 97% of past cases, purchases in October have tended to end up with a profit after 180 trading days.

5. What is the likelihood that the result just happened out of chance?
The results of the study are significant to a probability of 1 out of 6.79 billion-billion-billion.

6. Why does the phenomenon occur?
Unknown.

7. What is causal relationship between the phenomenon and the result?
Unknown.

Moreover, drawing empirical observations and using statistics can also be dangerous. In his book Practical Speculation, Victor Niederhoffer identifies 4 flaws in drawing conclusions from statistics:

• The Error of Randomness – or a failure to account for luck in generating the results. In our experiment, we checked for statistical significance using the chi-square test, so we may have mitigated this error to some extent.

• The Error of Omission – or the Missing Third Variable. Most arguments and market theories essentially draw a relationship between two things: in our case we are drawing a relationship between a time element (the month) and the corresponding returns. The Error of Omission arises when we draw a conclusion based on two variables, but the real driver of the result is an unknown third variable, that is also linked to the two variables. Correlation does not imply Causation—or simply just because two things happen together, does not mean one causes the other.

This is where we our experiment might have a problem. Time is a tricky thing—because in the real sense, time does not exist. It’s a notion, not a thing or an object. Therefore, by capturing time, we are actually inadvertently capturing other things that go along with time. By saying October yields the best results, we may be missing the possibility that there are other things that just happen to occur in October that drive the returns. This is essentially the answers to questions 6 and 7 earlier. Although it may satisfy us for the time being to just focus on October, if there is a third reason that’s causing the returns and that reason stops being associated with October—then our study is useless.

• The Fallacy of Mobility – or a failure to take into account a change in the nature of the population. In our experiment we checked this by chopping up our sample into years and seeing if the relationship holds for each year, which we have seen does. If the trends October’s performance in previous years were different, and the total results just skewed by the later years, then our findings would be slightly different, if not opposite. However, we only know our observations to be true for the sample population we have. We have not checked if the same is the case for previous years.

• The Fallacy of Aggregation – or to associate the whole with the sum of its parts. Since our sample was drawn over a bull market, it would be easy to presume that the individual months of our sample would perform similarly to the aggregate—which we have see is not the case by checking each individual month, so our experiment may have also mitigated this somewhat.

As I have said, The Seasonality Problem is an unfinished experiment as it stands. Further study is needed to probe deeper into the nature of the months to isolate exactly what behavior happens during these months that drive returns. This is something that won’t happen overnight, but if we stick to our guns, and remain ever vigilant and critical of ourselves, there’s no reason why such knowledge can’t be found eventually.

I have attempted to show one way of uncovering truth and debunking myth, which hopefully will inspire others in the forum to be more objective and critical about information they give and receive. As Financemanila expands, it’s only logical that it evolves. As Morpheus said in the Matrix Reloaded: “I do not consider it to be a matter of hope, merely a matter of time.”

I say the same thing for Financemanila’s development into a premier financial website.

It’s only a matter of time.

 <end?>

Related Posts You May Wanna Read Too

• No Related Post