Automated Trading Theory Discussion

[permanentjaun]

Here's my point. MA's absolutely follow price in the direction. They don't necessarily diverge. In ranging and trending markets they still follow them. Do they lag, yes. Do they predict, no. But, are we concerned with predicting movement or following it? I would say we want to follow it. It's similar to playing breakouts. You can predict that a currency will break a channel and begin a new trend or you can wait for it to confirm a breakout and then enter. Is that not a certain bit of lag as well?

You argue that MA systems only last a few months at most before failing. I can't understand this. MA's have been following price ever since they were conceived.

I started the thread on this idea and the idea that the success in MA automated trading is finding an efficient time frame, period of MA, type of MA, and strategy to analyzing what its showing. I noted that I believe markets are always moving in that we can't designate a ranging or trending market. This is because of time frames. I could argue GBP/USD has been ranging since 1985 on a monthly chart. Or I could say it has been trending up since the 10th of october on a 2 hour chart. Which is it then? Ranging or trending? It's all relative.

My idea is that automated systems will always need us, yes. What works on a 4 hour chart today may not work the same in 4 months. The idea is that we may apply the same strategy, such as the Open/Close MA strategy, but alter the variables of time frame and period on the MA's as the market shifts. Perhaps playing a 20/21 WMA open/close strategy on a 40 minute chart works for a few days and then begins to fail. Therefore you take a look at the market and decide to shift the strategy to a different currency or just change the time frame to 2 hours and a 35/36 WMA open/close strategy.

Do we not do the same thing for manually traded systems? Even price action systems? Do you play the breakouts on a 20 minute chart or a 4 hour chart? It's personal preference and sometimes is more obvious on one chart compared to another. You adjust your exit strategies accordingly. Someone looking for 400 pips off a breakout on a 10 minute chart isn't practicing safe risk management.

If you say the MA is lagging too much, make it faster and take a look at a longer chart time period so the movement is more substantial and enough for you to exit with your pips.

Why am I putting up such a fuss about this? It's because I believe highly in finding efficiency. If something is lagging, what chart is it lagging on? If it's producing too many whipsaws, what chart is it whipsawing on? Those are signals to me that I'm not using the efficient settings and I should reconsider the variables.

I'm trying to figure out why MA systems, which follow price, have failed in the past. I admit they fail, but do all of them fail? Many don't know the rituals of Buddhist monks in Tibet, but that doesn't mean they don't practice them and lead successful lives in their own sense. Therefore, I can't imagine it is impossible to create a successful automated system so long as it is monitored.

Why bother with an automated system? I could sit here all day looking at the charts and manually trading the system. I want automated so that my emotions are out of the equation. That is already a huge benefit to my investing. Secondly is to make trading multiple pairs, even just the majors, less time consuming in terms of my sitting in front of the computer.

Sorry for being so stubborn.

Besides those thoughts, I've had this one. Has anyone tried programming a trailing stop loss where it closes a trade if the currency closes in the loss? For example you buy at 100, it races to 200. You have a 25 pip trailing stop loss. The currency falls to 160, but closes at 180. Then it runs to 225 and falls to close at 200. It would trigger the stop loss at 200. Anyone tried this? Matt

[craig66]

But if you talked to the monks, would you believe them if they contradicted your beliefs? In trading you really need all the help you can get, it's not a easy exercise, being stubborn in trading is going to cost you $$. People are telling you to forget MA's for your automated system but you keep grinding on and on...

If you really want to use MA's I would check out the 'Hull' moving average, it seems to be the best MA I have seen, but I still can't turn it into a profitable EA, but don't let that stop you, I could have missed something...

There is debate on all forums concearning MA systems, going back years, I suggest you check it out.

[permanentjaun]

I wouldn't believe them no. There's a difference. Buddhism works for them, but does that mean Catholocism doesn't work? In forex terms it's like saying, using an automated price action system works, but does that mean a MA system doesn't work? Would a system based off of RSI not work? I'm sure there is a system based around it, I just care not to go in that direction since it doesn't directly follow price. There is more than one way up a mountain. MA's follow price and are easily determined by computers so that is why I think it is the best for automation.

I'll have to read a lot of the MA EA threads. What do you say to my idea of efficiency and always adjusting around it? Adjusting chart time frames, MA periods, etc.? I still believe it depends on what chart youre looking at. Does a 20 WMA on a 10 minute chart not quickly show the early beginning of a 6 hour down trend? It can, but will whipsaw on the way down.

Also, how much of the movement are we interested in capturing? When you wait for confirmation of a breakout you give up the pips possibly gained by entering before confirmation. Are we not doing the same thing with MA's? Matt

[permanentjaun]

craigg66 - Talking to others using the Hull MA. One method they suggest is to write an EA that changes the color of the HMA as its slope goes from positive to negative, thus indicating an up or down trend. Then program an entry/exit signal at this point of color shift. Did you ever try that for an EA?

[craig66]

Yes, marginally better than normal EA.
I think the major point you seem to be missing is that it's all very well to lay indicators on past price movment and think "If only I had sold when blah blah", it's a whole different game in realtime. I would hope you have the ability to code up an EA to try this yourself, if not, this would be a good place to start learning.

[permanentjaun]

Yea I need to learn. Unfortunately GFT doesn't have EA capabilities yet. They will sometime in the near future is what I've been told. Not sure if I'll switch to someone else to try things out. I was just hoping to get the theory out of the way now and worry about the scripting later.

[astro]

Hi permanentjaun,

Look, I don't know how long you are in this business, but the enthusiasm you demonstrate talking about how an MA system does exist and everyone is wrong saying it doesn't - telling me you are a beginner.
Also the way you describe your money management theory. You talk in terms of lots, rather than % risk.
This reminds me of myself when I started trading.
Anyway, I don't intend to fight you and try to convince you that MA systems don't work. They don't. I know that. No real trader use them since they don't work.
In any timeframe the market will sometimes range and sometimes trend and you won't know that until it already happened and you lost $$$. You can't find a timeframe and/or MA which produces only winning entries. Period.

Now, do you want to take this advice? That totally depends on you.

I wish you good luck in your trading, and I do hope that you find what you are looking for.

Astro

[Soul-Trader2004]

Astro is right, MA systems do not work.

however using an MA as a moving support / resistance is a useful activity.

But i have tested hu8ndreds of combinations of a cross, double cross, triple cross with filters without filters etc - none of em work and that includes the famous ones.

The all work if used wioth discretion of course but not for auto trading.