Pippa is excited to be here with the Trading Gurus

[pippa]

Hi,
here is pippa. I discovered the forex universum last year and think that it offers awesome and unique advantages.

Compared to other (financial) investments or businesses,
the high liquidity,
the low startup costs,
the automation potential,
the easy scalability,
the non-existent overhead (no employees, no store, no inventory)
the free trading platform (ie MT4),
free demo accounts with
free real-time data,
free historical data,a
continuousley trading market 24/5,
a high margin (1:100, 1:200 etc), and live accounts with
low initial funding, as well as a
lot of already published indicators, strategies and EAs make it look very promising to me.

Some time ago I was looking at money-makes-money ideas/strategies (versus: my gruntwork makes money), and forex seemed so much better than stocks, indices, futures, reral estate etc.

I come from a (self-employed) business background, and not having to deal with communal permits, licenses in general, and,

most important, NO EMPLOYEES to babysit, this seems perfect.

So I started out with demo-accounts, running MT4.
Trying different and, in hindsight, bewildering ideas from forums and (e-)books,
I managed to blow about one demo account per month.

This was the time when I really learned to appreciate the free data feed, the free platform, the free demo accounts.
No real costs, apart from time invested. Perfect! Try this with a brick-and-mortar business idea, where it takes ages to set up shop, sign a lease, get all permits... you easily spend thousands if not more till some revenue starts.

So I was not deterred the slightest with my virtual "performance".
And I read a lot of books (real ones), large part of which are recommended in this site´s book store.

And I began to shift my thinking and acting re the forex markets.
I tried to apply position sizing, stay with one currency pair, look at different time frames, test out the very basic ideas

(versus highly complicated rules I have seen on forums), apply stop loss.

All this led to my current thinking:

I will open 3 live accounts with licensed brokers, more or less 0.5k each (w/ real license, and offering at least 1:100 or 1:200 margin)
I will get a (hosted) Windows server where I will install Metatrader4.
I will run automated trading strategies as this is the only way to go (IMO), will elaborate on this further down.
I will use one of these live accounts to forward test new ideas/strategies at 0.01 lotsize, as forward testing using a live account will give much more accurate figures of what to expect.
I will start out with just one pair: EURUSD, but with different timeframes, from M15 - H4.
I will use the following lotsizes: 0.1 lot/1k or 0.2lot/1k, will explain why further down.
I plan to compound, then withdraw a large part of profits (? yes, right), every six months to fund living expenses after initial time period to let accounts grow to something more serious than 0.5k... and to spread risk by opening similar live accounts at different brokers with same or different strategy.
My plan is to stop my daytime job by the end of the year if things go according to plan,
continue compounding account equity and living off savings till spring next year, then implementing my 6months take-money-out.
I am flexible to move from my current location, once taxes are higher than moving (and associated) costs to change to a very tax favourable locale.


OK, why 3 brokers?
Because just having looked at and used demo accounts, I experienced GREAT differences between brokers, generally, the more they advertise, the more unpleasant are the details of trading there. Like stoplevels, spikes, trade context busy et al.
And I expect live servers to be another beast entirely (ie exiting at preset stoplevels, ease of entering during fast markets etc), and maybe there are also differences in processing my payout requests later.... (see Jim´s post on "London due diligence" on his blog)...that´s why I want to diversify risk and spreading my business over different brokers.

Ok, why the 0.1-0.2 lot/1k capital?
My thinking goes like this: to open a position, I have to commit opening margin, to sustain a sudden 200pips spike or "move", I need some reserves, all this while a position can be in the negative, if I use the H4 timeframe I need a SL that makes sense, ie percentage of ATR, NOT something like 10 pips. So, all this added together, will result in 0.1-0.2 lot/1k. (And I have found out in real life, that I can stomach hefty drawdowns). I am looking at the mentioned timeframes from M15-H4, because trading smaller timeframes normally goes hand in hand with smaller pip gains, but I´d have to keep my lotsize more or less the same because of possiblity of spike and required margin, so: fewer pips translate into less money gained.

Ok, why EURUSD?
I have been looking at this pair for months now and from M15-H4 the charts now look familiar to me. On this setup I have tried an amazing range of different things, and different indicators.
Primary reason was low spread and high volatility,also important: it seems to be trending well, so next pair to add (when diversifying) will most likely be GPBUSD.

My trading idea is to determine when the market is trending or ranging,
when trending in H4 use simple trend following method (not determined yet, see further below)
when ranging go to lesser TF like M15, M30 and try to get pips using BollingerBands.
When trending and position open in H4, possible to open hedging position when trending indicator/EA in M15, M30 indicates opening of position opposite direction. I know, this goes against common knowlege to trade in direction of larger TF, but my idea is to HEDGE, therefore being (partly) insured against spikes (partly, because there still will be SL for each and every position), and having a chance of making pips in lower TF as well.


My joining this forum and where my abilities are lacking, is in devloping automated EAs.
I have settled for MT4 because there are a lot of indicators, EAs, strategies and snippets of code around, which will make it easier for me to get started.
Learning a foreign language took me appr 9 months, so learning MQL should be possible (for me).
Albeit currently I feel like a mason who shall perform brain surgery: Just not up to the task.

When reading source code mql4, I do understand the more basic concepts,
So what you will read from me in future will be posts with code (snippets),
where I have taken existing code, applied my ideas and changes and then got stuck
on mundane things like parenthesis or semicolon etc.

Things that are completely alien to me at the moment is writing/reading to/from arrays , files, dll etc.

But I have to learn and conquer this, as I need EAs to backtest strategies, mildly optimize,
and then of course to put them to work on live accounts.


Thanks for bearing with me and reading that far.

PS: I will stop by this forum most likely once or twice a week at weekends, when time allows, so don´t get upset if I seem to respond slowly.

As regards learning Metatrader4 language, I am into "the book" at mql4 site and discovered the Coders Guru pdf, but at least as helpful are the code and reasoning behind the different versions of Ray The Random Robot, as it shows me WHERE and WHY to insert/change/adapt code.

This is why pippa is here in RAYs proximity. Thank you!

[Jim]

Hi Pippa,

Welcome to the Trading Gurus community forum. Ray has asked me to send you his thanks for your kind words.

Your plan sounds like a good one to me, and here's a few of my initial thoughts for you to consider:

1. This is not in any way a recommendation for this broker in any other respect, but you can trade 0.01 lots with high leverage on an IBFX Australia mini account, which corresponds to a size of 100. Apparently that is referred to as a "nano-lot" in MT4 marketing speak. That's the cheapest way I know of testing an EA on a "live" account.

2. Trading longer timeframes certainly makes a lot of sense, particularly when you're just starting out. As you say, apart from anything else there is much more margin for error. However if you can ultimately eliminate most (all?) of those errors then smaller stops = greater size = more money gained.

3. What MetaTrader brokers refer to as "hedging" seems to me to be an utter waste of time and money, as far as those brokers customers are concerned at least. Long one lot plus short one lot equals flat, except it uses up margin and attracts swap charges.

4. Whilst there are indeed "a lot of indicators, EAs, strategies and snippets of code around" for MetaTrader, most of them are an utter waste of time also. In addition whilst MT4 might be free, unfortunately MQL4 is not a very productive programming environment. The MT4 Strategy Tester might be of some use on the longer timeframes you're thinking of using, but once again it's an utter waste of time for backtesting and/or optimizing "scalping" style strategies.

5. When you have coded a reliable trending/ranging indicator please feel free to upload it here!

6. KISS is my motto too. Here's a slightly longer expression of that idea, attributed to Albert Einstein:

Make everything as simple as possible, but not simpler.

I hope all that hasn't put you off the idea too much, and I trust you will enjoy the ride over the next few months and years!

Cheers,

Jim

P.S. I've taken the liberty of adding a few relevant links to your post. I hope that's OK with you.

[pippa]

Death By Data
Death by Data

Thanks for the friendly welcome, and your thoughts:

1) IBFX Australia
Upon your comment I have opened a (demo) account with them, and was quite pleased with all the additional indicators & EAs.
Seems they look at it from a trader´s perspective. Have taken a look at the mq4 code of the some of the EAs, LOTS to learn. Very interesting.

2)Longer timeframes, I certainly agree on this aspect.

3)RE: Hedge useless; Thanks very much for your imput. Now that you said it and I thought about it, seems logical!

Links/moderating etc. I am o.k. with that, no problem,
now let me get into 4)

when I originally posted my introduction, I had been looking at charts on demo accounts for long months.
But I have never used the built-in strategy tester (Ctrl-R). Just read about it. Crazy, ha?

Now the last week I really dived in(-to backtesting // strategy tester), and first it felt all good to be able to zoom through time and see, how RAY (and "lesser" ones) would have performed, if let loose on the markets.

But then I noticed a lot of strange things (at least to me) in strategy tester:

mismatched data erros, holes/missing bars, optimizer simply stopping after a few iterations, different spread applied, etc.
So I started surfing around for solutions, and this is what I came up with:

I need clean data for a long period of time, using reasonable spread that stays the same for every test (within EURUSD).

1) I installed a fresh instance of MT4 platform, solely for testing, and followed the procedures outlined in the thetrademachine.com blog
Let me explain: I read your post about Birt´s eareview.net and tick-data from dukascopy, but I failed in achieving the download from dukascopy, and I was unable to get the "great resource of alpariUK" historical data, that was mentioned somewhere else.
2) I downloaded, also from a link at the trademachine (now, if this sounds like advertising for other site, just delete it, I don´t mind), alpari M1 data from 2001-2009 (zipped download 129MB, extracted 900MB, at megaupload site)
3) Then I imported that data into history (F2). This is where I lost my remaining hair under my wig: a weak computer and lots of data just don´t side well. In the end, I settled to import ONLY EURUSD_M1 data. Otherwise, my computer (1GB RAM) couldn´t handle it. Or maybe, it was me.
4) I then attacked the problem of unrealistic spread backtesting. The solution I found, was going to MT4i.com, downloading the spread-controller. While there, I also downloaded the data-integrity-checker. Nifty.
4a) To apply the spread-controller (which comes with PDF instructions), one has to restart that Metatrader instance; so I found out, that the spread data is stored in SYMBOL.SEL, and the spread-controller (when in charge), backs up the previous symbol.sel into symbol.sel.bak. I renamed ***.bak into ***.bak2 and then copied the actual symbol.sel into symbol.sel.bak ; The effect is, once you gonna finish with the spread-controller, it will restore the original symbol.sel, in this case no funny weekend spread, but the chosen 2pip spread for EURUSD.

But it doesn´t end here, wait:
Applying the MT4i data-integrity-checker, I found GAPING holes in my data, which sent me into another depressive spiral.
For very recent data, one might be able to download history data (in an instance different to the backtesting MT4, which is NOT connected to the internet); for older data, I don´t know.

So, all in all, I think first step for me should be reasonable data quality and data quantity, before backtesting and "The Search" begins.

I plan to backtest in EURUSD, over the eight or ten years I will have data.
I plan to always apply the same spread (2pips), and also use the same data, to make results more easily to compare.

While surfing for data solutions, I found the TickLoggerForFXT (on mql official site), which I foolishly applied (for a day or two) to all the symbols one of my demo accounts offered: This considerably slowed down my PC, it logs appr. 3MB per day per symbol, and I contemplate running it on the (hosted) server I will get.

What are your thoughts on this: Is this worthwhile, or just a waste of time?

Thanks!

[Jim]

Hi Pippa,

Using consistent historical data of "reasonable" quality and quantity for your backtesting is a must. For the timescales you are talking about logging ticks seems to me to be slight overkill though!

In case it solves your problem I've uploaded a script that exports historical data from MetaTrader 5 in a format that can then easily be imported into MT4. Applying a consistent fixed spread to that will then certainly help you to produce consistent backtest results. Please bear in mind though that in real life spreads can widen to 30 pips or more when the NFP data is released, should you happen to have any trades open at such times.

Also see this thread, if you haven't already!

Cheers,

Jim

[pippa]

Thanks for the script! Downloaded immediately, then downloaded MT5 from Metaquotes site. Opening of demo accoun t fails. MT5 downloaded from IBFX Australia. Running of that instance of MT5 fails due to 64bit incompatible to my 32bit architecture...... but I will get the data!

Let me disclose where RayThe Random Robot was possibly conceived in the next post!

[pippa]

Where RAY The Random Robot was conceived !

Actually he is a son of Black Jack, as Stridsman aptly writes in the book: Trading Systems That Work (pg. 70).
More info and all the gorey details in chapter 11 (same book, pg.187 ff).

May I suggest another useful book for your store:
The Cartoon Guide to Statistics by Larry Gonick& Woolcott Smith ISBN: 9780062731029

[Jim]

Hi Pippa,

I was dabbling in random entries before Stridsman published his book, so perhaps it would be more accurate to say that Ray the Random Robot is in fact a cousin of Black Jack?

Ray in his current incarnation was actually conceived here a couple of years ago.

Thanks very much for your book suggestion. I'll rush out and buy a copy! If it does indeed provide an easy to digest introduction to "Statistics for Traders" I'll certainly add it to the reading list.

Cheers,

Jim