Really interesting way of looking at it - I never thought about it this way before. I was not sure whether to agree or disagree until I completed the next sentence. Perhaps waiting for the breakout is a skill (patience), perhaps exiting only after clearly being proven wrong is a skill (giving a trade enough room), or perhaps taking profits only after a a sufficiently large move is a skill (letting profits run). Perhaps the element of luck could be assigned to the actual outcome of the trade rather than the process. However, there is no correct answer to this - we can merely have our own views.
The first answer is quite good and I have not looked at any of the others. When I did my stats course, the optimum sample size was determined at 30. Seems like that no longer is the common view. However, when using statistics, your population needs to conform to certain criteria. For instance, it is quite easy to get an answer when trying to determine the average wealth of people in a certain suburb, but it is quite difficult to determine whether saturated fats cause cholesterol. In the first instance, the test is quite simple whereas in the second a large number of variables will influence the result. For instance, do people drink red wine, do they exercise, are they obese, etc.? In such a case even a sample size of 1,000 or even 10,000 may not provide sufficient confidence. Just look at how views have changed on this topic in the last 20 years.
Are you certain that your setup can be limited to a simple parameter, or are there multiple parameters to consider? Just remember that statistics provided a way to handle large amounts of data in an efficient manner. Nowadays with computers you can run back-tests over decades of data, test out-of-sample, perform Monte Carlo analysis, etc. Unfortunately, no matter what you do, you are not guaranteed that the setup will work going forward. Perhaps this is why Ed Seykota calls it mathturbation.
While the article is called "There are known knowns", I refer to the "unknown unknowns". Firstly, how do we know that we have all the data? I refer to the previous belief that all swans are white. Secondly, reading "Fooled by Randomness", I was struck just how medical professionals got survival rates wrong and it seems like these are not the only statistics that they get wrong. Therefore I do not agree with the view that we will make the correct decision even if we have all of the time and relevant data.
I do however agree with the skill in making a decision in a timely manner with limited information comment. I have observed that most successful traders operate using certain rules, i.e. heuristics, but they also incorporate feedback from their most recent results into their trading. An example would be where I trade breakouts, but only 1 of the last 5 worked. In such a case, I will definitely be trading smaller size and observing current trades to see whether I should scale down further or scale back up.
The following user says Thank You to grausch for this post:
No, I am not certain, of course. I just thought that the 20-30 number sounded like a 'common sense', sufficient set of trades that, if executed in generally the same scenario, with generally the same execution, could give you a sense as to whether they are working or not, as a whole. But of course each trade is unique because of the different participants in the market each time.
It sounds like we actually agree. The 'making right decisions with all the info' is I suspect a 'provocation' by the author to lead the audience into the concept that the real skill is making decisions with a limited amount of information in a limited amount of time. I suspect what he meant was that even in the event of 'unknowns', given sufficient time they would become 'knowns' but, again, I think he was merely trying to make a point.
As a side note, I love heuristics, I think you are right that trading should incorporate a mix of heuristics and feedback from recent results provided (as Mark Douglas put it) that my objective observation of the market as well as of those results is unimpaired.
Last edited by xplorer; January 12th, 2016 at 07:29 AM.
The following user says Thank You to xplorer for this post: