Significance – The important implications of statistics for investment

Author: Not in this mountain

Source: Snowball

The stock market is a world full of uncertainty, people always hope to find some certain and graspable laws to make a profit, so there are a variety of research methods, such as technical analysis, fundamental analysis, macro analysis and so on.

However, no matter what the method, it is hoped to summarize from the complex phenomena that can be used to make a profit, such as the W bottom bullish, high return on net assets of the company's stock long-term yield.

Incidentally, no matter what seems to be regular, it is only a probability in the field of investment.

So how do we know if a conclusion is a true law or a conjecture, an accident or an inevitability? Is there a way to identify?

In statistics, there is a very important concept called significance, which is used to measure the reliability of a certain conclusion, or to judge whether a phenomenon is the influence of inevitable factors or accidental factors.

In the academic papers of various disciplines, the word will appear almost always, those who have not studied statistics may be a little dizzy, simply put, to say that a conclusion is significant, which means that the conclusion does not appear by chance, but is regularly reflected.

Conversely, it is not significant, that is, it is impossible to determine whether the conclusion is caused by chance or necessity.

So, how is significance calculated? Here we will not dwell on the complex statistical calculations, but only the ideas.

If you observe that the frequency of a certain phenomenon is much higher than the frequency that should occur in a random state, the so-called anomaly must have a demon, and there is usually a certain law at work behind this phenomenon, we will say that it is significant.

For example, if you toss a coin 10 times, 6 times up 4 times down, or 7 times up and 3 times down, this can not say that the coin is problematic, but if 9 times up, 1 time down, we have a relatively high probability that the coin is unbalanced, that is, significantly unbalanced.

Or toss the coin, this time 10,000 times, if 6,000 or more of them are up, we can say that the coin is significantly unbalanced.

As can be seen from the above examples, whether the coin is significantly unbalanced is related to two factors, one is the probability of positive occurrence.

The second is the number of coin tosses. The fewer the number of throws, the higher the positive probability required to be significant, for example, 10 tosses may take more than 8 times to prove their significance.

The more times it is thrown, such as 10,000 times, 6,000 times is enough to prove its significance.

Back to our investment decisions, first of all, technical analysis, we often hear about head and shoulders top, double bottom, red three soldiers, volume breakthrough and other K-line patterns, can these patterns really indicate the trend of the later stage?

In the era of various "masters", we should keep an eye on it, in the spirit of being responsible for our own money, we may wish to do a significant test before investing.

For example, if a stock has broken through, do you want to follow? You can count how many K lines have broken through in the history of the whole market, how many of the trend markets have been accounted for after the breakout, how many have gone back after the breakthrough, if there are enough samples in the history of the volume breakthrough, and the probability of going out of the trend market is large enough, it can be considered that there is a regular existence, that is, significant, and vice versa.

In fact, according to a large number of tests I have done before, the profitability of the vast majority of classic patterns only exists in the analysis of the "masters".

This is also a significance of quantitative investment, and some seemingly reasonable conclusions may come to different conclusions from a statistical perspective. Some conclusions may apply only to one period of history and are invalid in other periods, which is also the difficulty of investing.

So, does fundamental investment need to consider significance? For example, in the past two years, high return on net assets stocks have risen well, is this a general and lasting law?

We can count the performance of high ROE stocks in the past history to see what percentage of the stages that perform well are significantly higher than other stocks.

In fact, from the perspective of historical backtesting, the simple consideration of high return on net assets when selecting stocks is often not good in the later period.

In the market is full of all kinds of specious views, arguments, experts, teachers will use vivid examples to tell you that after a certain form of stocks soared, or because of a certain financial indicator how to be optimistic about the future market, and those failed cases are often avoided, investors are easily blinded by the survivor bias.

The market is complex, chaotic, and evolving, no one can predict the future, and humility is the attitude of science.

For those masters or pseudo-masters who speak with an overconfident style, I personally usually look at them with a skeptical eye.

In order to improve the ability of rational thinking and independent thinking, it is very necessary to properly learn some basic principles of statistics. Among the introductory books, I highly recommend Neil Sarkind's Love statistics, which uses simple examples to explain the basic ideas of statistics in an easy-to-understand way.