Successful Automatic Trading - Like Chasing Bigfoot?

Some of you believe automatic trading is viable, others don't.

If you stand on the negatively biased side on this topic, maybe your stance will shift a little in coming months and years. I hope it does.

I've spent eight years on a personal journey that has turned me into what I hope is a decent spot forex trading strategist who can turn nearly any trading strategy into a fully automated Expert Advisor (EA).

It took a while to figure out how to code EAs that can deal with any number of currency pairs, any account type, any number of decimal places, etc.

I'm NOT hugely into indicators. Many traders are searching for the magic indicator but what we must always remember is that ALL indicators are a reflection of what has occurred in the past and that they do NOTHING to deal with upcoming events. All traders must have an arsenal of well practiced (or automated -- I prefer automated!!!) tactics that can deal with any and all market events.

I prefer simple multi-pair strategies, which is why Looser's trading strategy caught and held my attention. That strategy is similar to ones I've been pursuing and consider most worthwhile.

Offer trading ideas that are intellectually honest and show you've been through the darkest trading forests and maybe I will help you automate your ideas. Please don't take offense if I don't respond to your ideas. I've tried almost everything and have very specific ideas about what has a chance of working and what does not.

If you ask basic coding questions, I'll just point to examples in my own source code. It's up to you to dig in and figure it out. I'll just say that the mq4 language is fantastically rich and powerful. It has sufficient depth to build true artificially intelligent trading robots. The burden is on us humans to write great code that encapsulates brilliantly elegant trading strategies.

As far as code, I will never post ex4 files, only mq4 files. Source code offers full transparency and that's the only way we can have honest, open interchange of automation ideas.

See you around, happy trading.

Good luck with your coding biz.

However, I do feel like EAs are like chasing Big Foot. I use to be a heavy believer/dabbler in EAs, but I just don't believe that there can be a long term successful EA or else everyone would have a copy of it. Then everyone would be rich, and we know that just can't be possible in any market. To me, simple lines of codes are no match for the human brain. Hell, I think the world's fastest super computer can't even mimic a mouse's brain.

Yoda Glenn, remember that the human brain is also what makes people smoke crack, inject heroin, drink too much alcohol...and make bad trading decisions.

If you were a bit more than a dabbler, you might have some understanding of what avoinics software has been doing in military and commercial airplanes for the last 20 to 30 years -- performing tasks quickly and perfectly that no human brain, nor network of human brains, could accomplish. Most people have no awareness of it.

Watch if you dare, learn if you are able.

Well, when you get something together that shows promise, please do share. I'm all for making money, so if you can demonstrate then I'm on board.

Until then, my experience with EAs has been a lot of junk and snake oil.

I was interested in modifying a martingale EA that I have, and using Binary Options as a hedge against its position. I may contact you about modifying this particular EA.

I'm going short on the probable terminal length of this thread - 2 pages max. stop loss at 3.

Let me try this another way.

First, any trading strategy, whether executed by hand or via program/script, COMES FROM THE HUMAN MIND (if you have software that someone claims CREATES A STRATEGY, I'd like to see it).

Second, any trading action a human can take, can also be taken by a program/script. There is no arguing that, it is demonstrably true.

Third, any trading strategy that is COHERENT, COMPLETE and RATIONAL can be readily automated by a competent software engineer. I maintain that "trading on news" is not rational, therefore I exclude that from consideration for automation.

Where we fall down is in having weak strategies and/or weak implementations of strategies. But BOTH come from the human mind.

Some traders claim they make trading decisions that CANNOT BE AUTOMATED. They believe their decision making processes are so sophisticated that "a computer" could not replicate them. More likely, they are unable to coherently, completely and rationally explain their trading strategy. They "just do it" and claim they engage in decision-making that is beyond the capacity of a computer. I assure you, their decision-making is NOT beyond the capacity of the computing power in my wristwatch, but their decision-making may be beyond their own ability to fully and precisely explain it. What my wristwatch CANNOT do is feel emotions and take irrational, self-destructive actions. Nor can it synthesize trading strategies.

I believe that the "sophisticated traders" do not understand their own thought processes, nor how computers work. I also point to automated strategies that perfectly and tirelessly track up to 14 currency pairs, and take efficient trading actions on them, around the clock, all trading week long. Can YOU do that? If the trading actions do not yield long-term profit, the fault lies in the human that dreamed up the strategy, and/or in the skill with which the strategy was automated by a human.

Computers do exactly what they are told to do, nothing more, nothing less.

We might not know WHAT to tell the computer to do, and we might not know HOW to tell the computer what to do once we THINK we know what the "WHAT" is.

ALL poor trading strategies, and there are many, come from the human mind, whether they are executed by hand or via automation.

A couple steps ahead of you.

http://www.forexfactory.com/showpost...&postcount=316

My mission here is to find strategies I think are good, then implement them as perfectly as I possibly can. The members of the forum can determine if I have picked good strategies and whether I have implemented them excellently. If I fall short, help me do better.

Just be sure to have a reversal strategy and well rehearsed tactical execution ready to go for when your stop loss gets hit.

can you code this system -- 50EMA 1hr chart -- sell below lower band once candle close -- buy above once candle close -- tp is optional feature--
SL will be - when the new signal come in the opposite direction-- (most of the time is around 50 pips)
see chart

well first I will have to be sure not to post aga...damn it! Got suckered in by the head fake...

Very interesting subject

Let's be frank : I don't think that automatic trading is viable, as per the usual definition of trading. However, automatic order management can produce very nice results no matter what the trading strategy used. I tried using random entries with a good order management system and guess what, I had better results with random entries than with my usual, developed with my blood trading entries.

Let's be frank a second time : managing orders isn't easy to understand, it is NOT easy to think of, and it is absolutely not what the trading industry sells us. We want indicators, easy money, 100% hit-rate trading systems or EAs and they try to sell us that. Everyone is culpable but that's fine as long as you take this extra step and start making money. The absence of interest shown by the public what will protect our profit margin and that's fine by me.

I'd be really glad to know your opinion about this matter, and even though my words might not be the ones you would have used I think you get the idea. But even that isn't a new concept : FF "stars" like 7bits or RR or others went on with that idea or even gave us working EAs but people are not interested. I don't know why but I think it's too abstract for them. Instead, they relish on threads like Jacko's house of pleasure and pain.

Now now, I'm using Oanda's API and I can't backtest my EA on that platform. Is it possible to backtest on Metatrader ? I mean, things like entries, additional orders, reversals, hedging, time specific tradings zone, time specific entries-exit (after the order was created), % of account invested, daily/weekly profit, was the last trade a success or not and how much time it took to hit the stop loss, etc etc. It took me time to be able to that on Oanda and even though I don't have the backtesting I am quite happy with the results. Would learning to code in MT4 be worth it ?

I'd say that you are not specific enough. How are the bands calculated ? a 50 EMA is easy, but what are the other 2 lines ?

nothing to do with the setup-- those are grids of 00 and 50 levels --

Two things I have almost no time for: indicators and backtests. I explored both extensively and saw that they were steering my thinking in the wrong directions.

If you are good with the Oanda API and getting satisfactory results, MT4 might be of academic interest only.

50 ema high low close 50 exponential shift 0

Not the grids : what exactly are the pink, yellow and greenish lines ? It seemed to me that the yellow one was the 50 EMA but that you wanted a BUY order executed when the price closed above the pink line or a SHORT order executed when the price closed below the greenish line.

I would buy a straddle with a strike around 3 pages, that's usually when the bickering starts.

I understand your opinion concerning indicators and backtests but what do you think of order management/position management/sequence trading ? (I prefer the latter term) This concept sounds like what you did with you looser trading system but you might have a different opinion

Thanks for your answer concerning MT4

A fully or semi automated method for order management/position management/sequence trading makes great sense and my hat is off to you for your disciplined approach to it.

To see snapshots of risk and health of a trading strategy, I like to show the following on-screen: largest floating loss, lowest margin, account type, leverage, current balance and equity, last flat balance and equity, lot size being used, number of buys and sells open, biggest winning and losing trades, current PL of open trades, profit target...there are a few others but those are the most important ones I can recall off the top of my head.

Looser's "Kiss Me" strategy is pretty simple and not too different from one of my own strategies, so it caught my attention. Every trading action he describes is fully automated in the EAs I posted. The version 03 that I posted today opens trades around the London open and will hold trades open as long as 9 hours, which isn't too different from what eTrader is doing. He uses an indicator to determine inital all-long or all-short bias, but I haven't incorporated that feature yet. I don't believe it's critical to the success of the strategy.

the 3 lines are 50ema high low and close -- when 1hr candle close above all 3 lines buy when close below all 3 lines sell the chart is clear