Reality based coin-tosser method that beats 95% of traders in the world.

TAILS today again. Short.

 

i have studied extensively market occurrences and there are certain events that happen fewer times than a normal distribution would predict and certainly much less than a fat tail would suggest.

for instance, number of days the market goes down 5 times in a row is about half of what the normal distribution would suggest (based on 100 years of DJI historical records)

i have pinpointed this in my earlier post alluding to the fact that market occurrences are strictly dependent from each other while in the flip of a coin they are not. this is one of the most interesting aspects of market behavior and probably one you should study thoroughly.

 

Whisky,

It seems you've touched a nerve with some posters here. I'll go ahead and pinch that nerve.

"Fooled by Randomness" is not only a good book but the term is perhaps nowhere as readily apparent as it is in this forum. It is somewhat satisfying to know how many completely clueless people think that they understand a market dynamic, let alone the foundation of a true edge. Let them trade, the more the merrier I say . That said, I know exactly who is on the other side of my trades and why.

My earlier system result posting was meant to demonstrate how easy it is to create a percieved edge. In fact, if you open the attachement below you'll find another manifestation of this random process at work. All I've done is introduced a couple serial dependencies and trivial ones at that. The "results" are neat, eh?

Some poster from 2003 who shall remain nameless mentioned that this exercise proves nothing. Frankly, I'm suprised he's still around trading... understanding randomness is way more important than attaining an edge. In fact, if one doesn't understand randomness to the point of being able to understand the attached results, then, one will never attain a real edge.

Mike

To everyone else here: would you trade this system if you saw only the results report? Be honest.

 

My anecdotal experience is that markets last about the same number of days going up and going down, but on average the SP goes up more per day than when it goes down. But as I do not intend to prove it, you can take it as a factoid.

 

I like the idea of this thread. Whisky plays the role of Socrates. Getting everyone to ask and answer questions creates new information.

The market is the sum of an enormous amount of human decisions. Even the black boxes were programmed by humans and are therefor human decisions.

Human decisions and actions are not random. Buying and selling is not random. The market is not random.

The enormity of decisions being made is beyond comprehension. Humans by nature don’t like to admit or think about things that are beyond their comprehension. Often times, we’ll judge something quickly when we don’t understand and be done with it. “It’s dumb.” “It’s unimportant.” “It’s random.”

The number of stars in the universe is beyond comprehension. Just because its impossible to think about all the stars in the universe, does not mean they are random and the same is true of the market.

Look at this thread, a collection of human decisions, are the responses random? No, most appear to be motivated by:

Ego - [example - the name calling]
Need for attention - [example - the irrelevant jokes]
Need for validation - [example - the arguing]

95% of the people posting here are motivated by a desire to feel special, get attention, and appear "right."

Humans only get up in the morning because we feel these emotions. You trade because you feel. You buy and sell because you feel.

That is not random.

Let’s say Whisky is the market. When he buys the market goes up; when he sells, the market goes down. We sit next to him and watch his face as he trades. Do you think he makes an expression when he’s panic selling? Do you think he makes a face when he is averaging out of a profitable trade? Do you think you could, with time, learn his expressions and match them to what he’s likely to do next?

 

Whoa - we got a winner here! (see attachment)

Where's BWolinsky and TraderZones? I think I want 1mil for this system:eek:

LOL... I'm bored today...

 

Correct me if I'm wrong (I'm a noob trying to educate himself on stats and probability) but is this just an example of data snooping?

 

An interesting thing happens when I remove the random variable.

Is this an edge? Any takers?

Guesses are free to everyone except TraderZones

 

No, there is no data snooping going on here.

http://en.wikipedia.org/wiki/Cumulative_distribution_function

 

Sorry, I just don't see it. As already pointed out by someone else, you can have 5 wins totaling 50 ticks and 5 losses totaling 50 ticks, you think that's the real world? The real world says you can just as easily have 5 wins totaling 5 ticks and 5 losses totaling 100 ticks. What then? You really think that will just average out over time? Based on what?

The upward bias of the market has also been mentioned. If you actually toss enough coins that the head and tail count evens out, you're still far from break even because the market is biased long.

What kind of draw downs are you going to have during strong trends? You're tossing coins for a (theoretically) 50/50 call on long or short, all while the market flies upward or downward. Again, you're far from breaking even. But you feel that this will average out over time - based on what?