You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Is Daytrading a Random Event?
- Thread starter LelandC
- Start date
Originally posted by chasinfla
I have done studies that have shown that the shorter the time frame, the less 'trendy,' the more random, price action can be. But it's relative, in a sense.
Care to describe your "studies" in more detail?
Considering the amount of interest and money in very short-term trading, it's surprising to me how little material there is in the public domain that demonstrates much sophistication.
Originally posted by Kymar
Care to describe your "studies" in more detail?
Considering the amount of interest and money in very short-term trading, it's surprising to me how little material there is in the public domain that demonstrates much sophistication.
I was afraid someone was going to force me to be more articulate about this....(If you're looking for sophistication, you've come to the wrong place. The derivation of that word is very telling.)
I took the DM contract (this goes back a decade or so). I figured the percentage of time the close was greater than the open. As I increased the time period (week, month, year, multi-year etc) periods of clear trending were seen. Hardly exhaustive or overly sophisticated, but enough for me.
I also observed, seperately, that a chart of a 100 event coin toss, the purest 'random event,' looks alot like a tick chart. Further, a 10,000 event coin toss looks like a your average daily stock chart (leaving out bubbles and busts). I tried to attach a photo that illustrates this.
Randomness is part of life. I believe it dominates in all time frames, and that two - three standard deviations is the boundary that the Creator has put on the majority of events in order keep things manageable for us (but this is just a hunch). As I said elsewhere, the edge that God gave man is free will. This gives us control over how we react to the 'randomness' of life.
Incidentally, it's instructive to observe how, the more sophisticated mankind becomes, the more he seeks to iron out the 'randomness' of life, make the unpredictable predictable, and somehow take the risk out of everything. This is so contrary to natural law that its results will be (and are) disastrous. God has reserved the control of that for Himself. When man seeks to do it, he is trying to play God, and that NEVER works.
Finally, trends, which traders often define as proof of 'non-randomness,' are made up of a series of random events. The trends themselves are probably just larger ticks of randomness in ever larger trends.
Randomness is here to stay. We might as well get used to it. And, yes, you can prosper in an environment of randomness, owing to free will.
I mistakenly deleted this from my post.
These charts clearly show that purely random events (50/50 propositions) can aggregate into 'trends' (which is how most traders define 'non-randomness'). Charted are the winnings of a coin toss game.
Incidentally, a strong case is made for 'regression to the mean' or contrarian trading, as the trend moves further from the mean line.
These charts clearly show that purely random events (50/50 propositions) can aggregate into 'trends' (which is how most traders define 'non-randomness'). Charted are the winnings of a coin toss game.
Incidentally, a strong case is made for 'regression to the mean' or contrarian trading, as the trend moves further from the mean line.
Attachments
everyone interested in this topic should drop by a bookstore and read "fooled by randomness" written by a guy named Taleb who is an experienced options trader, Phd in mathematics, and a quantitative analyst. He also was a friend of Victor Neiderhoffer and warned him before he blew up. It is a great book with lots of story examples that get the points across. Its one of those books that you wind up reading again and underlining stuff.
Originally posted by chasinfla
I mistakenly deleted this from my post.
These charts clearly show that purely random events (50/50 propositions) can aggregate into 'trends' (which is how most traders define 'non-randomness').
As has been pointed out elsewhere, the fact that a coin toss chart (and, more typically, an interesting section of a coin toss chart) can appear to resemble a tick chart does not really say anything about the possible randomness of the latter. That any randomly generated pattern might appear to resemble some other pattern does not necessarily say anything about the randomness or non-randomness of the forces that generated them.
Without going into too much detail, I think it's also worth noting that general measures of randomness (or of dependency or other kinds of theoretical predictability) may give indications, but don't tell us what most of us need to know as day-traders. Unless we're strictly interested in intraday trend-trading, and to a large extent even then, we're most interested in detecting and acting upon certain moments of high tradability - relatively highly consequential progressions or reactions within or between "local" probability frameworks. Such moments may only just barely register in statistical tests of very large amounts of data - but to a trained eye they can be fairly easy to pick out, and, to an experienced trader, just about as easy to exploit profitably.
Originally posted by Kymar
As has been pointed out elsewhere, the fact that a coin toss chart (and, more typically, an interesting section of a coin toss chart) can appear to resemble a tick chart does not really say anything about the possible randomness of the latter. That any randomly generated pattern might appear to resemble some other pattern does not necessarily say anything about the randomness or non-randomness of the forces that generated them.
Without going into too much detail, I think it's also worth noting that general measures of randomness (or of dependency or other kinds of theoretical predictability) may give indications, but don't tell us what most of us need to know as day-traders. Unless we're strictly interested in intraday trend-trading, and to a large extent even then, we're most interested in detecting and acting upon certain moments of high tradability - relatively highly consequential progressions or reactions within or between "local" probability frameworks. Such moments may only just barely register in statistical tests of very large amounts of data - but to a trained eye they can be fairly easy to pick out, and, to an experienced trader, just about as easy to exploit profitably.
Right. In the mean time, trends happen in the midst of 'randomness.' It's all random. And it works.
nobody wants to believe that it is random, but when they blow up they will say they just had bad luck. Its a psychological thing. The math is real.