Optimal Stop-Loss and Take-Profit Levels

[Gamma_Jammer]

Dacoffey - like the style. Sounds like you're putting some though into this and preceeding fairly methodically.

Out of interest, have you considered setting a risk/reward level (1/2 or 1/3 are common) then just widening the stop and tp levels together while preserving this ratio.

Also if you are able to extract this data, you could consider turning it into a 3d plot with percentage of winners as the other axis.

Hope these suggestions are of interest, keep up the good work.

GJ

[dacoffey]

After collecting and examining more data, I think that perhaps results presented earlier were too optimized.
This is always a danger when seeking to model complex systems, and must always be kept in mind.

I have found that better results can be achieved by reducing the sensitivity of the networks (ie, signalling trades less frequently, when they get a much closer match) and increasing exit levels.

I am now using a stop-loss of 70, and taking profits after 90.

anomaly: Yes, I think I can get even better results if I use individual levels for each pair. I haven't coded this up yet, but plan to research at some point.
Also, like you mentioned, I have found my nets to be most successful at predicting EUR/JPY.

gamma_jammer: I don't collect that data right now, but its a good idea and I will prolly start logging that also.

[Gamma_Jammer]

Interesting to hear you say your results may be too optimised. Admission of this fact is a sign that you're definitely thinking along the right lines imho.

The aim with what I was suggesting was robustness really. I am currently trying to develop a system for my own use that has a high success rate without trying to catch 100% of a given move. Hence the suggestion about winning %. Get that high enough and, provided you've factored in your costs correctly and the few drawdowns aren't excessive, and you're in business.

[dacoffey]

Gamma, I am thinking about how to measure risk...I think I will also add a statistic for average drawdown across all trades.

Is this how you quantify risk?

[anomaly]

GJ--
Six months ago, I would have definitely agreed with your last post. But, my thinking has changed dramatically since then. Here are my thoughts now, several groundbreaking books and five trading systems later.

Success rate, or system reliability is really not as important to the success of the system as it is to the trader's psychology. It can be very hard to trade a system that fails 90% of the time, but that doesn't mean it's a poor system.

My first strong attempts at system development attempted to take just a piece of the move, because most moves went at least X pips. I wasn't going to give those X pips back, because no one ever went broke making a profit. The more I tested the system, the more wrong I found I was. I'm sure it can be done, but creating a positive-expectancy system that cuts profit runs would be difficult to do. Even if you were successful, I am skeptical that you would be able to exceed the expectancy of a system that allows profits to run.

[dacoffey]

anomaly: Exactly. Psychology is 100% of trading.

(Psychology is 100% of everything according to me, but I am a cognitive science major so I may be biased.)

This is the essence of a trading SYSTEM.
The system tells you what to do, and you if you have a system that works, you must believe in it and do what it says.

I don't know how many times I've thought, "oh, I totally see a move there" and poof there goes 50 pips...or I think "oh god this posistion is gonna tank" and I exit, only to have it bounce 100 pips. Succumbing to fear and greed will make you a very broke trader.
I much prefer to just let my system do its thing.

[Gamma_Jammer]

There are many ways to measure risk. Commonly it's expressed in relation to return, and there are a few commonly used methodologies.

Have a look at Sharpe ratios for starters.

http://www.moneychimp.com/articles/risk/sharpe_ratio.htm

Hope this helps

GJ

[anomaly]

dacoffey--

I've definitely found that developing a trading system is easier than fixing my flawed psychology. =) Much faster, at least. In fact, the system I trade now is based entirely on numbers (not chart patterns) and I only look at the chart to confirm what my system has already determined.

Regarding risk, I view risk as the amount of capital I have at stake at any given time. This is usually the difference between your entry point and your stoploss (multiplied by your position size, of course). I always evaluate risk on a worst-case scenario for both risk/reward and money mangement purposes.