Magic of ADX

Oh and yes, had you adopted the premise of the EMH in your trading system, the number of trades should have not been more than 1. Why? I suggest you find out from your sources

dear simple and oxy,

instead of beating around the bush, let me clear some pts regarding backtesting vs trading with the test in real time.

1. i have already mentioned few points in my earlier post. i am not going to counter other's views, which are theoretically correct, even if they may be flawed in real life execution. having said that, yes, if you want to make use of past data, you have to backtest. BUT BACKTESTING WITHOUT OUT OF SAMPLE WALK FORWARD TESTING MEANS that you are trading YESTERDAY's MARKETS, not today's. We can say that there is serial dependence on future financial time series data from the past behavior. BUT THERE IS NEVER A MIRROR IMAGE (CV has put it very well somewhere using non linearity and mutivariate arguments), and this dependence is very very difficult to calculate (almost impossible). you can use different methods only to ESTIMATE the same. Use a Hurst Exponent estimation and you will see what I mean. moreover mkts are changing every moment, more so in todays world. so even though BACKTEST is a must, PURELY RELYING on BACKTEST is never the answer.

2. again overoptimization is NOT GOOD, as i said yesterday, but that does not mean that you should not optimize. because whenever u r using a canned indicator, someone has already optimized it for u and u r just using the same, eg 14 for RSI and ADX etc Why 14? did you ask? again multiple range value optimization on the same data with a surface chart analysis can give you a good idea.

SimpleStuff, yesterday u asked about behavioral issues. unfortunately if u are trading, and dont have integrated platforms connected to brokerages, you will have to build your own discipline, or give the order entry to someone who will blindly follow the system's buy/sell signals.

really enjoyed that!
I wish CV the guru would be free this weekend to put his comments and enhance the thread manyfolds!

Uh,oh..did someone call me.

PS - Man, some huge posts here

Journey of a Scientific System Developer :=

http://www.traderji.com/98313-post112.html

Let us Traders concentrate with our job so as to buy in future Systems from this fellow member.

eh? What so special about genetic algos and NNs? They have been around for some time now and I have been using em for a while
http://www.traderji.com/attachment.p...1&d=1183035268

Abt T3-Fibs, I think the whole program is quite primitive.I have coded my own Fib Confluence detectors that work 10 times better than T3fibs.How about you start a thread posting the levels and I will chip in and lets check the accuracy.I also own the T3 Fibs program so comparisions can be documented.
Wall Street is run by the bots for some time now, nothing spl about the article.

Self Optimizing/Adaptive Systems have been around for quite a while too See http://www.clayburg.com/introduction.htm

Also could you post some results/ perf stats etc of the systems you are talking about, all this sounds interesting.

Cheers

Attached Images

NN.png (20.2 KB, 44 views)

This is for uninitiated like me;

Capital Asset Pricing Model (CAPM)
this model was originally developed in 1952 by Harry Markowitz and fine-tuned over a decade later by others, including William Sharpe. The capital asset pricing model (CAPM) describes the relationship between risk and expected return, and it serves as a model for the pricing of risky securities.


CAPM says that the expected return of a security or a portfolio equals the rate on a risk-free security plus a risk premium. If this expected return does not meet or beat our required return, the investment should not be undertaken.


The commonly used formula to describe the CAPM relationship is as follows:


Required (or expected) Return = RF Rate + (Market Return - RF Rate)*Beta

let's say that the current risk free-rate is 5% ( Bank Fixed Deposit)
and the Nifty is expected to return to 12% next year
Let us say Satyam's Beta is 1.9

Hence
Required (or expected) Return = 5% + (12% - 5%)*1.9
Required (or expected) Return = 18.3%

Now := Efficient Market Hypothesis
"An 'efficient' market is defined as a market where there are large numbers of rational, profit-maximizers actively competing, with each trying to predict future market values of individual securities, and where important current information is almost freely available to all participants. In an efficient market, competition among the many intelligent participants leads to a situation where, at any point in time, actual prices of individual securities already reflect the effects of information based both on events that have already occurred and on events which, as of now, the market expects to take place in the future. In other words, in an efficient market at any point in time the actual price of a security will be a good estimate of its intrinsic value."

Raja's satement :
"Perhaps I wasn't precise enough
Take emh/markowitz/capm add it all up "

Now for uninitiated like me all these emh / Markowitz / capm adding up means
when the actual price is almost identical to it's intrinsic value ,then ,Markowitz / capm
,raja is trying to say Beta is used as a parameter to stock selection.

To understand this basic thing ,(which as a trader few yrs back ,i used to see in Dalal St
the highest Beta stocks for my Intraday) that raja uses CAPM as an indicator,i took so much time initially i actually
became dumb & thought maybe something gr8 ,in plain simle term his system selects
stocks to trade on Beta criterion.Which a dumb fool like me used to do in my Dabba Trading/ Kacche ki sauda ,days 7 yrs back.

Dear SS

I meant Performance Data/ System Statistics of a trading system which summarizes its historical performance like an equity curve, win/loss ratio etc

Attached Eg.

Cheers

Attached Images

Stats-Perf.png (31.7 KB, 33 views)

For CV ,catching up.

Nice!

A truer test of the system would be on intraday data, try it out