A challenger appears

Hello.

I had never heard of ForEx until two days ago. I've now installed a few trading stations (including apparently 3 copies of MT4. can't have too much of a good thing, I suppose) and I have some mad demo account money to burn.
I've also bought "currency trading for dummies" along with "technical analysis for dummies". Those titles are such a tight fit for me it's almost uncomfortable.

And so I've been spending a couple of evenings so far trying to write an EA that can consistently survive a strategy tester run across various timelines and pairs, and failing, unsurprisingly.

Well, no rush. I won't play with real money until I have generated enough reproducible data to convince myself that I may not be wasting my time here.

Essentially, I refuse to believe in voodoo, so any indicator I don't understand, I won't use. My current EA works purely with flat averages, and attempts to detect localized trends as they start and piggyback on the ride. It does a terrible job so far, but I understand exactly why.

Reading "currency trading for dummies" so far made it plain that the market may be 24/5, but not all hours are equal, so an EA worth a damn will most likely have to watch the clock and know the expected volatility impact for a given pair, at least. Too bad I have no idea what to do with predictable volatility changes yet.

At this point, I'm still not convinced this isn't a giant sucker play, but the two books I've bought so far have cost me just as much as Civilization 4 so as a worst case, I can write it off as playing a fun make-believe game.

If that's alright, I'll keep posting updates on my ever-so-slightly diminishing ignorance levels in this thread. I might even throw in a graph every now and then to keep people interested

I had the same discovery about a year, this big, bad, mad world of retail forex. Best of luck! You will get find plenty of advice on FF, lots of it contradictory I´m afraid. I´ll chuck my two bits in, but take all advice (maybe especially mine) with a measure of caution.

Markets are primarily social constructions. Those charts are graphically representations of multiple human decisions, traders all reacting to information and to the market itself (so therefore each other). I don´t really understand EA´s but with this in mind I can´t imagine they work for this reason. Big banks and hedge funds employ flesh and blood traders to play the market, not programmers.

It´s a game of chance. Sure, with good analysis and lots of chart time you came skew the odds in your favour say 60/40 but the market WILL go against you on a regular basis. This is why good money management is crucial; just like a coin will come up heads 5 times in row if you throw it enough if you trade enough you WILL have a losing streak. If you´ve been risking 10% of your account 5 losing trades will cut your account in half.

But don´t listen to me, read, read, read and read some more, spend every waking moment on the charts and then maybe, just maybe you will be one of the 5% of retail traders who make it.

Go for it!

http://www.forexfactory.com/showpost...05&postcount=3

Oh don't waste money on books. There is so much free info everywhere. All of us have some pdf s somewhere, just ask. I will look around for you.

Thanks for the kind words. Peter's advices for n00bs is the first thing I read since it's linked from, well, everywhere in the forums. It's a fairly straightforward read, if a little depressing at times.

E-books are fine, and I do intend to beef up on knowledge online, but there's something to be said for plain old paper, conveniently bound together.

So I read a bit about those awesome forex EAs that win 100% all the time OMG awesome. Too bad the web sites that sell them all sound like complete snake oil salesmen, destroying their credibility pretty much entirely.

So that seems to indicate one of two things:

 

1. It's impossible to build an EA that consistently beats the market, or
2. Whoever built one ain't sharing.

And then there was http://www.forexfactory.com/showthread.php?t=109589 which gives you an mq4 file that does some crazy "grid" voodoo, which makes the graph go up like mad, right up to the point where it flattens down horribly. Maybe it's awesome and the final flattening could be avoided with a simple tweak, but I just don't understand what it's doing at this point, so I'm filing that until I can go back to it.

I'm a coder by trade, so my natural instinct is to express my understanding and hunches into code. Right now, that results in mostly consistent losses.

Just for giggles, here's a chart where the EA does (very) roughly what I want:
http://imgur.com/GgteA.gif
The red line is a sell order, the blue line is a buy order, awesome!

Except right after that, it starts to look like this:
http://imgur.com/H5BVo.gif
Which is complete crap, and will pretty much destroy any portfolio in a frenzy of terrible decisions.
Note that I'm not using any stop loss on purpose here, as it would only hide the true quality level of the EA. In a "real" trading session, if I was dumb enough to use that EA, I suppose I'd put some really tight stop losses to slow down the bleeding a bit.

Hey, I said there'd be graphs, I never said they'd be good.
This is from a recent EURUSD M1 slice, the red lines are some SMA5 and SMA20 lines, which my whole logic is based upon, stealing some basic concepts from SPGM (ok, stealing whatever little bit of SPGM I understood.)

So based on the graphs, you might tell that the EA only does one trade at a time, and tends to alternate sells with buys, except when it has no idea where things are going, in which case it waits. The criteria for starting an order probably needs to be a lot stricter, which may result in a lot more waiting. The criteria for ending an order right now is just "whenever SMA5 crosses SMA20 at a nice angle" which isn't totally bogus, but happens a bit later than ideal. I'll try to use a sign reversal in the SMA20 derivative as an earlier clue and see if that helps.

Which reminds me: Is there a builtin way to approximate a numerical derivative of SMA indicators in mt4?
I keep having to do those manually right now.

The federal mimimum wage is at $7.25/h. Take an 8 hours day, that's $58 a day.
Now let's say we have a seasoned trader who's able to consistently extract 20 pips a day out of the market.
For that trader to reach the minimum wage, he'd need to extract at least $2.9 per pip. That means trading in with volumes of at least $29,000.

If we have a 100:1 leverage account, and use a 2% risk policy, that'd imply our seasoned trader needs to use an account size of $5,800 or greater, or he won't make the minimum wage, even with his performance averaging 20 pips/day.

Unless I severely botched my math here, that doesn't seem like an encouraging proposition for small traders, who either need to make a amazingly high amount of pips consistently, or need to come up with a lot of start up capital to make this a viable past-time.

Which brings me to my next question: What is the lowest consistent spread one can get from a reputable brokerage firm?

Well, I think I've come to terms with the futility of writing an EA at my stage.

It's an old habit. Coding is my big hammer and everything looks like a nail. It usually work, but probably not here. Not yet.

This is the part where I stare at live graphs for the next few months and hopefully build up an understanding of a few common patterns in the process.

Hi, I think we have similar hobby. coding an EA.

Try look into martingale strategy and try coding it with many2 different idea in the forum. You will find it fun and challenging to make it profitable without margin call.

For backtest your EA, get at least 90% modelling.

That's one of my problems too. Where can I get actual sub-minute-resolution ticks for historical periods?

It looks like most of the historical data the strategy tester is getting is, at best, one tick per minute. That's very different from the tick resolution I get from a live feed, and the EAs as a result end up behaving quite differently.

Am I just supposed to keep a trading terminal open for a few years and hope it saves every tick it sees so that I can start backtesting meaningfully?

Strategy-wise, I've been looking at a bunch of EAs, some of which claim to be martingales (usually the crazy grid ones). Some others claim to be "neural networks" too, although they're usually so rudimentary they're barely justified in calling themselves "perceptrons" (and of course, they casually skip over the training part, providing instead a few magic numbers instead.)
The cold truth is, there's still a lot I don't know, and since I refuse to use stuff I don't understand, that means martingales are out for now.
I will get to it though, just as soon as I'm done failing with my other ideas.

Right now, I'm sticking to the idea of watching MA crosses and seeing how far I can take it. I've read many times that it's a dead-end and that I'm therefore wasting my time, but I find that reading up others' conclusions is not nearly as valuable as coming up to my own conclusions. Something about the journey being more meaningful than the destination.

Oh well, I remember these days as I used a calculator to calculate when I will be endless rich. Unfortunately it doesn't work that way.

Can I ask how you determine to enter a trade? it looks like your using MA's.

Depressing, bloody depressing??!! Well, no one has ever said that before and I was just going to post and say what a pleasure it was to find a new member with such a good sense of humour and a lack of rose-tinted glasses.

But now I won't - err, I just did - oh well, it couldn't have depressed you too much.

I'll pop back in when I get back from holiday the end of next week and see if you're still smiling.

Currency Trading for Dummies, indeed.

Very poorly.

So far, the best I have is to stare at SMA5 and SMA20 crosses (which lag) then decide based on the shape of the cross and the outgoing angle if it might be a good time to put an order.

It tends to be wrong slightly more often than half of the time (and of course, thanks to spread, flipping the decision logic still keeps it wrong more than half of the time.)

So that's what one might charitably call a "work in progress"..

Argh. Noooz! I didn't mean to upset you! That was a compliment, I swear! You throw a healthy dose of reality at all of us star-eyed n00bs, and all that.

In truth, I'm a big fan. I've read your guide twice so far, and a number of items that I had mentally filed under "later" became "Oh yeah, I think I've seen that one" items. I'm probably going to read it a few more times until everything sinks in.

The secret to my high spirits so far is my utter lack of financial commitment. The moment I've deluded myself enough to believe I can make it with real money is likely to be the moment where my mood might darken a tad.

It will most certainly take longer than two weeks though. I'll be lucky if I'm done assimilating all the knowledge I need by the end of the year.

Well, I'm kinda doing the opposite here: Calculating that even if I somehow become super-consistent in my trading results, I won't even be able to pay my internet bill with it unless I use a large enough account size to start with.

BTW, your oanda chart is most impressive. I have no clue how you do what you do yet, but the fact that you can gives me hope.

Alright, inspired by Xaron's link, I've opened my very own myfxbook account to share my glorious results too:

http://www.myfxbook.com/members/krak...ount-evar/3124

I'm done for a while with EAs that auto-trade. and MA crosses too.

I'm still primarily looking at EURUSD M1, but I do check the longer timeframes to help evaluate the odds I could stick around longer if short term moves doesn't work out.

Next up, learning s/r, trendlines and stuff like that.

Well, I got slaughtered. On the plus side, I became intimately acquainted with the joy of margin calls. Yay, experience!

Alright, my plan for the week-end:

 

1. learn a few T/A things ( this seems nice.)
2. decide on a simple methodology to try for next week.
3. write buy/sell scripts that enforce lots size, stop and limits using some risk and MM logic.
4. setup a much smaller demo account.

You are doing a great work!! Welcome in the forex jungle

@tunera: Thank you for the kind words, even though I don't think I've earned them quite yet.

Well, mid-week-end update:

 

1. I've acquired some minimalistic rudiments of technical analysis. Not a lot, but enough for my next bullet.
2. I'm going to try to follow Jacko's approach (pdf, thread). It looks really boring, in part because of the slower timeframes. Oh well, I can still do fast-paced stupid trades in another demo account.
3. I hacked a bit at money-managing scripts. I'm tempted to automate various bits of Jacko's strategy, but won't. Not yet anyway. It's better to try new things manually, I think.
4. I'm going to use a new demo account with a much lower size. $2500.

I'm not sure whether to be excited at the possibility that I might have a reasonable edge, or terribly bored at the prospect of making probably less than 1 trade a day.

Another week-end update. I used Jacko's approach on a few trades. The AH stuff is probably my favorite part. The use of slow timeframes is the most bothersome, since I'm still at the early tinkering stages where I have to keep messing with things continuously, rather than put a trade in and walk away for a week.

I now have some rudiments of Pivot Analysis, S/R and trendlines. It's handy.

Alas, I have gone back to my sinful ways of scalping M1 charts. Still, pivots and S/R help for that quite a bit too.

And I haven't used my smaller demo account yet, as I first want to see if I can bring back my first account into the green. Given that I blew off half of its value in a few hours early on, it seems like a good challenge.

This week, expect more night scalping. And probably a bit of Python code.