I beg that those of the odd ones here scream my name 5 times out loud...
Enough already! It's not random
- Thread starter kut2k2
- Start date
Quote from kut2k2:
Dear fellow technician, this thread is for YOU.
The next time some random walking idiot tries to rain on your efforts to find a technical edge while talking down to you like you're a small child, just direct his arrogant ass to this thread. Maybe it's not too late for *him* to learn something.
I never thought it would be this evident, given all the preaching about "proof" that the markets are random. Guess what? It's all BULLSHIT!
I just ran a little robust statistical tool called the runs test that was specifically designed to test a sequence for randomness.
First I collected 10 years of data on the S&P 500, from Jan 2 1998 to Dec 31 2007, from yahoo.
The results are:
2,514 data points
2,513 daily price changes
1,315 ups
1,197 downs
1 zero (ignored)
1,332 run sequences
Applying the runs test yields a z-score of 3.111.
The chances of that z-score coming from a random sequence of ups and downs is
ONE OUT OF 1,073 !
I got your coin-flip "chart" right here, random walkers.
just calculate the average return of your returns, then define an "up day" as a day with performance over that average and a "down day" as a day with performance under that average.
And have a look at your new z-score.
Sorry for you if it hurts... you've just demonstrated that on that time period the market has been going....significantly up.... which is something we all know from plotting the index...
As for stock index futures, I have found it useful to consider them as behaving with both random and non-random movements. The key I believe, is that the shift from random to non-random is not random.
*SIGH*
Allow me to clarify, and sorry for assuming too much that people would do a google on "runs test" (just press "I feel lucky", it takes you straight to wikipedia):
The chance of getting 1,332 run sequences from a random sequence of 1,315 ups and 1,197 downs is one out of 1,073(!)
In other words, the upward drift is already discounted in the runs test.
The guys who designed the test made no assumption about the ratio of ups to downs, they designed simply to test if an observed sequence of runs was random in a given pool of X ups and Y downs.
Also, the runs test is one of the few known as nonparametric. No assumption about the actual underlying distribution of ups and downs is made.
All it tests for is randomness, and it didn't find a reasonable degree of randomness in the S&P 500 for the ten years that ended last year.
You are an amazingly simple person. People have pointed out numerous flaws in your logic, and you keep banging your drum.
You did not prove anything. But keep saying you did and you can sleep better at night.
The market moves up. The period you chose was strongly up. You found that more days were up than down.
All of your logic after that was gum flapping, from someone who thinks he knows the math, but has serious deficiencies.
It hurts, just to watch.
Perhaps he uses technical analysis to determine random v. non-random periods...
Quote from kut2k2:
Speaking of childish, why can't YOU post actual counter evidence?
P.S. I don't do .doc
You started the post, you have made a false statement, you exhibited no understanding of the subject, you failed to comprehend the math, you were unable to see the hard evidence that I have already provided and now you are trying to use someone else's posts to back up your infantile theories? Stop embarrassing yourself and go back to school, read some books develop some knowledge then come back prepared and we will talk. My goal is education but I only provide knowledge to the ones who want it. You clearly are ignorant and unwilling to learn, so no efforts from me will change your thick head. Useless!