Which strategies work, and which don't

trhudson · May 1, 2012

Quote from jcl:

Eliminating losses is easy, just set a tight profit target, but all tested strategies so far got worse when a profit target was used at all. Which confirms that one should let the winners run.

I get the exact same results as well implementing profit targets. I have attempted numerous types of combinations of profit target / stop-loss and these are not as effective as letting the winners run.

I have found that profit-targets may help smooth the pnl equity curve, but it does not produce higher profits in the long run.

As for tight vs. loose stop-losses, this depends. I have found that large stop-losses often cause larger draw-downs, but then again it really depends on a few factors.

I always test implementing a strategy that would exit a position if the risk is less than my initial risk if a high / low is broken on the day. This typically takes the cake in both reducing draw-downs as well as generating higher profits.

I will add, that I do not use typical technical analysis as the methodology that I created using a top-down scientific approach was developed around measured moves.

I test time-framed based entries and exits as well and most strategies if they work trading in the first hour are generally more profitable if held until the close of the day. However, this also comes along with a slightly larger draw-down. It's really about risk vs. reward. While the profits are there in the morning, the trader must ask himself this - would the additional profits be 'worth your while' to hold the position to take the extra risk.

failed_trad3r · May 1, 2012

what is sharpe ratio? I don't know and I don't care. i only care about making money and sharpe ratio is the worst! you are much better of with profit factor. My profit factor is 5!!!

failed_trad3r · May 1, 2012

Quote from trhudson:

Before the book even begins, the introduction discusses the issues that impair the discretionary trader.

This discussion involves a belief system and this is a major reason why it's difficult for a trader to switch from discretionary trading to algorithmic / automated trading.

I have read many times from discretionary traders that there is no way a program could give you insights to how the market works. I guess these guys have no idea what data-mining is. They also do not understand that many successful automated traders use a scientific top-down approach to developing trading systems and prove trading ideas.

Rule induced data-mining is a way of proving if a trading method is valid. This is not AI, but a method that allows a trader to see market tendencies and determine by computer simulation if the tendency is tradable. If you read a few other books, such as The Talent Code or Talent is Overrated it has been proven that practice creates talent. A computer simulator can practice years of trading in minutes that would take a trader years to complete manually. The market is changing and the new breed of day-traders will be automated.

DISCRETIONARY TRADERS LOSE BECAUSE THEY USED MANUAL BACKTESTING, WHICH IS BEATEN BY AUTOMATED BACKTESTING.

a guy i know printed out markets since 2005, each intraday was one A4. He then manually looked for his setups. After 2 weeks he has tested his strategy. LMAO. he traded discretionary.

jcl · May 1, 2012

Quote from failed_trad3r:

what is sharpe ratio? I don't know and I don't care. i only care about making money and sharpe ratio is the worst! you are much better of with profit factor. My profit factor is 5!!!

If I had a profit factor 5, I would probably also not care about the Sharpe ratio. My best profit factor is 2. With which method do you get 5?

failed_trad3r · May 1, 2012

Quote from jcl:

If I had a profit factor 5, I would probably also not care about the Sharpe ratio. My best profit factor is 2. With which method do you get 5?

wait, do you daytrade?:eek:

dom993 · May 1, 2012

Quote from jcl:

While working on a couple different trade strategies, I found an interesting trend. The more time I spent with implementing an algorithm, the less profitable it turned out to be. Quite frustrating. The most profitable algorithms were also the simplest.

Have you made similar experiences?

BTW, these are the algos tested so far:

Simple frequency filters - Sharpe ratio ~ 2
Seasonal trading, AI based - Sharpe ratio ~ 1.5
Curve pattern detection, Frechet algorithm - Sharpe ratio 1.2 .. 1.5
Support/resistance, AI based - Sharpe ratio 1.0 .. 1.2
Price action, AI detected bar patterns - Sharpe ratio 0.9 .. 1.2
Conventional TA, f.i. RSI, MACD etc - Sharpe ratio below 1

For comparison, I've read that professional human traders achieve a Sharpe ratio in the 1.03 area.

Not sure what a simple frequency filter trading strategy is ... would you mind elaborating on this?

dtrader98 · May 1, 2012

Quote from dom993:

Not sure what a simple frequency filter trading strategy is ... would you mind elaborating on this?

It likely refers to simple moving average based strategies (i.e. if ma1 cross ma2 then buy, etc...).

Incidentally, in my experience, SR increases at higher frequencies ...
unfortunately, so do commissions and slippage.

jcl · May 2, 2012

Quote from dom993:

Not sure what a simple frequency filter trading strategy is ... would you mind elaborating on this?

Highpass, lowpass, and bandpass filters. This is the lite-C code for a simple second degree lowpass filter:

Code:

var smoothF(int period) { return 2./(period+1); }

var LowPass(var *Data, int Period)
{
	var *LP = series(*Data,3);
	var a = smoothF(Period);
	var a2 = a*a;
	return LP[0] = (a-a2/4)*Data[0]
		+ 0.5*a2*Data[1]
		- (a-0.75*a2)*Data[2]
		+ 2*(1.-a)*LP[1]
		- (1.-a)*(1.-a)*LP[2];
}

This is no moving average. MA crossing strategies are normally not profitable because of lag.

Using such filters in trade strategies is described on a website to which I can't post a link here out of consideration for the sponsors of this forum. I already got chastised by a moderator for posting too much info.

trhudson · May 3, 2012

Quote from trhudson:

I always test implementing a strategy that would exit a position if the risk is less than my initial risk if a high / low is broken on the day. This typically takes the cake in both reducing draw-downs as well as generating higher profits.

Just a little follow-up note as to different tendencies in different types of markets:

This method for reducing risk is applied only to trend-following methods and not reversion to the mean methods. When you compare the results trading individual stocks vs. market-based ETF's it's obvious that they do not have similar tendencies as this is not a very effective stop-loss on the SPY, DIA, or Q's.

However, it is effective on many momentum stocks.

SteveNYC · May 3, 2012

Quote from jcl:

I got that from the book "Thinking, Fast and Slow" by Daniel Kahneman. According to his statistics, the profit difference between professional trading and throwing a coin is very small. For the traders that was also a shocking surprise.

Is Kahneman pronounced like Con Man?