What was the random entry/chandelier exit study?

1

1a2b3cppp

I recall hearing something about some test they did where they randomly entered various uncorrelated markets long with a huge trailing stop, and over time the account balance became very positive. I don't remember if they added to open positions or not.

I don't remember the name or any other info. Does anyone know what I'm talking about?
 
Random entry system and a atr stop. It was teached by chuck lebeau, van tharp and some others...

It is used to prove that entry is not so important then having a solid exit strategy. And they proved it by entering the market at random and using only an atr stop. The markets they used for this proof where the futures markets like gold, oil, wheat, and so on with exception of the s&p. The starting balance must be 1 million dollars. A lower balance did not work... No piramiding or re - entry...
 
Quote from 1a2b3cppp:

I recall hearing something about some test they did where they randomly entered various uncorrelated markets long with a huge trailing stop, and over time the account balance became very positive. I don't remember if they added to open positions or not.

I don't remember the name or any other info. Does anyone know what I'm talking about?
More...

Its talk of people who don't know what they are talking about.
 
There was a discussion thread about this on the blox forum, a few years ago. It spurred me to run some tests of my own. I found that there are ways to make all three of these variants profitable, when applied to a large, diversified portfolio of futures markets:
  • (random entries) with (trailing stop exits) --> profitable
  • (classical trend-following entries) with (random exits) --> profitable
  • (random date entries) with (random date exits), however entries are only permitted in the direction of a classical trend-direction indicator such as MACD or Percent R --> profitable
What are the ways, you might ask? very distant trailing stops, loooong time-horizon trend capture (set "Number of Days" parameters to large values such as "200").

I've copied one of the figures from somebody else's testing in that discussion (thread name = "Test results 4 different entries + random exits"). He ran the simulation 50 times with a different random number generator "seed" each time, and overlaid the results of all 50 runs on the same graph. All 50 of them are profitable. In this case, you might be tempted to conclude that Exits Are Unimportant since profits were made with random exits.

attachment.php


He used a portfolio of 10 futures markets; I tend to use more than 50.
 

Attachments

Instead of futures markets, what about using different currencies. Are those uncorrelated enough?

And you could use a small lot size so that each position was a small percentage of your account, that way you could have a really big trailing stop without running out of margin.
 
Quote from Marky33:

I did run a few tests and also found it was profitable.
Here's a pic of the equity curve:
More...

Looks like that hasn't been profitable since 92.
 
Quote from 1a2b3cppp:

Instead of futures markets, what about using different currencies. Are those uncorrelated enough?

And you could use a small lot size so that each position was a small percentage of your account, that way you could have a really big trailing stop without running out of margin.
More...

bump
 
  • Instead of futures markets, what about using different currencies. Are those uncorrelated enough?

    And you could use a small lot size so that each position was a small percentage of your account, that way you could have a really big trailing stop without running out of margin.


What prevents YOU from trying it out using currencies?

What prevents YOU from trying it out using a small lot size so that each position is a small percentage of your account? That way YOU could have a really big trailing stop without running out of margin.
 
I use a similar idea to check I am not just lucky, I have a universe of stocks I trade say the S&P 100. when I enter a trade I choose a random number between 1-100 flip a coin heads=long tails=short. I then "notionally" record a random trade at the same time I enter my trade, I record the trade at the bid or offer in favor of the random trade so I sell at the offer and buy at the bid I do the same thing on close out. When I close my real trade I notionally close the random trade.

It does pretty well but will big volatility, Its up overall, but thats because it shorted AIG just before it fell apart. So one trade has contributed its profit. To my mind it should yield the risk free rate with the volatility of the market.
 
You must log in or register to reply here.
Back
Top