Probability - or, when is it better to stay in bed?

[robster970]

Most people are saying every trade is an independent event so trade it as a unique trade, regardless of previous results. I am quite severely chastised for poor maths.

However, some of the same people are lately saying, that actually they amend their trade management rules depending on their success / failure. I don't think you can have it both ways.

That's humans for you.

 

[Martinghoul]

Intuitively, it seams resonable to assume that if you where trading something like a trend re entry system, taking multiple small bites out of a longer term trend then you'd see some degree of serial correlation. In practice, I certainly havnt been able to measure this, and people with way more experience than I claim that trades are independant.

Presumably, the statistical methods that are commonly used to measure this are inappropriate (or maybe the way we pre process the data). It always worries me when something that seams intuitive isnt bourne out by statistics.

Well, all statistical methods are, essentially, models that attempt to describe the behavior of a well-specified system. The mkt, in the context we're discussing it, is anything but well-specified. So that means you actually have a b-ass-ackward problem, in that your job isn't to describe the behavior of a system, based on a set of assumptions, but rather deduce the likely basic characteristics of the system from its behavior. In a sense, it's like instead of the usual "how many heads you get when you flip a fair coin n times" you're trying to answer the question "if I flip a coin n times and get k heads, how likely is it that the coin isn't fair".

So that's really all I'm saying... The problem, as we're discussing it, is about offering various correct answers to the original question, based on a variety of simplifying (and overly simplistic) assumptions. The problem is that the devil IS in the assumptions.

 

[intradaybill]

The problem is that the devil IS in the assumptions.

Knowing the assumptions is equivalent to having the solution to most problems in the world.

 

[Martinghoul]

Knowing the assumptions is equivalent to having the solution to most problems in the world.

Well, not necessarily... There can be interesting problems presented by well-specified systems with all known assumptions, 'cause otherwise most academic scientists would be out of a job. But yeah, I catch yer drift and in real life knowing the assumptions is half, if not more, of a solution.