“Trading is a probability game”. You might have read it a thousand times. I think many of the traders have not understood this or have misunderstood this. There is much difference between betting on the probabilities of a method with an edge and using statistical probability to find an edge.
To trade the Markets profitably, we need a method with an edge. Edge is nothing but a positive expectancy. As we all know well, there are no certainties in the Market. Anything can happen at any time. We must trade the method long enough, ignoring the individual trade results, to make this positive expectancy to work in our favor.
Working on statistical probabilities to find an edge in the market is altogether a different ball game. It is not easy as we think. For example, after an elaborate study, we find that 70% of the gap ups fill on the same day and we conclude shorting the gap ups at the open could be a profitable strategy. Sure it is not going to work.
There are many logical fallacies that could distort the studies. Many statistical tests calculate correlation between variables and in many cases correlations does not imply causation. It will be totally wrong to conclude two events that occur together are taken to have a cause-and-effect relationship. To cite an example all the momentum moves in the market will be preceded by an MA cross over. But acting on all the MA crosses is not going to make you money
Spurious relationships are much more dangerous. Some times two events or variables have no direct causal connection, yet it may be wrongly inferred that they do. This could be due to mere coincidence or there could be an unseen factor.
Market is a place where countless people executing countless strategies at any point of time. Each and every trader is a variable who can change the direction of the Market, at least theoretically. Every moment in the Market is unique and probabilities themselves shift from moment to moment.
In my humble opinion, statistical probability studies alone are not enough to find an enduring edge in Markets.