My progress as a full-time investor & trader

Hi all,

Just a quick one, further to a previous post on how the EAs try to catch long trend, ( http://www.forexfactory.com/showthre...79#post6992279 ) here is a quick example of using short term intra day entries to pick up the occasional long term trend ... potentially lasting several weeks and in this case resulting in multiple 1:20 Risk Reward trades. Example taken from one of the EA back test.

Attached Image (click to enlarge)

Take Care,

mim2005

21vs7,

"But how to do this? Positively correlated pairs are flattened in their index values. To determine how to intervene on particular pair, one needs values where it would stop trading that pair. How one would pick that? (a difference between them is less than xx in value= no trade) I what time series? We pick average?"
Currently, I don't have a mechanical formula for defining the minimum level of negative correlation needed, between two currencies, that gives (on basis of probability) continued directional bias for their pair. I hope to work toward achieving this, over the next few months, as I experiment with different correlations in selecting pairs to trade with Mim2005's EAs.

In terms of which timeframes to use, I tend to gravitate toward longer term TFs, for macroeconomic reasons (see my comments below). But it makes sense that the TF should somehow reflect the horizon of your own trading, e.g. if you are trading daily/weekly moves, then you should be checking strength/correlation on daily and longer TFs.

"I trade and also EURCHF pair which by each currency is highly positively correlated"
I agree that EUR and CHF are frequently very positively correlated, resulting in a lot of sideways movement in EURCHF. The strength plots that I review show this on a regular basis. There is also occasional positive correlation between AUD, NZD and CAD. AUD and NZD used to be highly positively correlated, but this has changed somewhat over the last 1-2 years, possibly due to Australia's stronger economic ties with China. USD and JPY are also occasionally positively correlated. EUR and GBP used to be positively correlated, but not so much these days. These (longer timeframe) correlation patterns seem to change gradually over time. Possible reasons are explained in threads like this and this.

However, it's perhaps debatable to what extent correlation plots are merely the result of summing the movements of (relevant) individual pairs, or price movement itself is the result of the strength/weakness in the individual currencies themselves. In other words, a "catch-22" situation involving cause and effect: is correlation the effect of price movement, or is price movement the effect of correlation? Because I believe in the longer term effect of macroeconomics (i.e. fundamentals suggest that the economy in one nation is stronger — or at least, perceived to be stronger — than the economy in another), I tend to be swayed somewhat toward the latter position.

One of the reasons Mim2005 and I chose EU for our backtests is that EUR and USD are frequently negatively correlated. (Other reasons included EU's low spread, and the fact that the entry edge of one of the EAs is specifically applicable to EU).

"To plot a change and potential consolidation when we are already inside established trend one could use a higher speed indexes. Meaning lower time series values which could give some kind of detection of a pause."
Agreed; the plots I use lag due to a smoothing factor, hence I'll occasionally look at the H4 plot to get possible advance warning of correlation changes on D1. I think correlations, like price movement, tend to become less predictable, the shorter the timeframe used. And I believe the best can hope for are very rough probabilities, that ideally need to be grounded in a robust 'abstract' (e.g. macroeconomics), otherwise the risk of 'curve fitting' increases.

I will continue to post my thoughts here over the next few months, in whatever direction my research takes me.

David

Yes, it does, indeed, mim2005! Thank you so much.

The question, IMHO, is now to see, coming back to erik1's stimuli, whether there is an evidence of "such kind of" edges to be stable, and to which extent, along time and possibly across markets. When I say "evidence" I attach to this some quantitative variable with some metrics. I am currently "brain working" on that, together with several other things...

As far as I was exposed to algorithmic system generation, so far, I was impressed at least by two things:
1. The underlying theory of markets under continuous evolution. This might be true, but I feel, to a certain extent (again we need metrics). Still more deeply this would imply a long-term market efficiency, if it would be 100% correct. It is demonstrated that markets are not a random walk on every timeframe you consider, and a hint I have is that what really makes it deviating from that model, could be some form of "stable pattern". The kind of explanations you give, i.e. that behind the markets there are "humans", with needs of sleeping and eating, among others, could be relevant. And, of course, we are much interested in the habits of the "deep pockets". The kind of automatic pattern finding I have seen so far is quite "rude" or "simplistic" for getting such kind of patterns, and furthermore, many of these systems are employed on timeframes where those intraday quite "intuitive" patterns will be absent (typically daily bars). If an evidence of stable patterns emerge, it would be relevant for the "whole industry" checking whether the modelled data series, used all over (at least in institutional quant departments) for validating strategies (and escaping data dredging) are really mirroring the market, or just an approximation, eventually filtering out exactly the kind of edges you and Hanover (and others) are trying to exploit.

2. Another thing that makes me a bit doubtful about algorithmic system extraction is the kind of extreme simplification of trade management available to these systems. This is why I so greatly appreciated your explanations about pyramiding, even partial hedging, the two kind (entry and exit) edges... Frankly, for the time being, I cannot really distinguish, in your (up to now) successful backtesting + something live, which proportion of profitability is due to the "strength" of the edges, and which one to the smart management of trades (which is much more than simple "money or risk management", that, alone, will nowhere produce long-term profitability without the "edge"). The proportion of edge and trade techniques would in fact influence the relative importance that stability of edges would have in system design/extraction.

After all, confronted with erik1's thoughts (and similar elsewhere), there is an interesting basic question of approaches here. erik1 told you're "the old style". It might be. But finally here we are discussing whether it is possible to develop profitable long- (or medium-) term profitability just with either an "inductive" (data mining, exhaustion of available data) method, or with a "deductive" (human smart observation able to get "regularity" in few observation and to abstract it into a generalization), or with both. Surely, as far as I have seen, such automatic system extraction methods are surely useful for institutional traders, for which harnessing a 2% excess profit (over the buy and hold) means billions a year, given the huge working capital. It's far less attractive for us, small fish retail traders. I find it interesting that erik1 avoided to answer my question about average annualized return. And I also suspect (let's say I have some initial "number" or better math concept about it) that the general claim of no profitability higher than that being possible, has much to do with institutional targets, both in terms of time horizon and accountability, besides regulatory as well as liquidity constraints.

Clearly, these are just my opinions, and it might well be that they are substantiated more by my ignorance than by some pieces of "truth". I am grateful in advance to anybody showing me new things to learn to correct my ignorance!

Bye

BL

Yes, its pretty hard to get some value out of this. Flattening is good. Flattening of different speeds of one currency index. Then of course a correlation towards other 6 in my scenario.
Im seeing my index values and haven't figured that out yet. Im in the first phase where im seeing only a broad picture of movement. Seeing specifics of particular currency on its own. I think any out of 7 Im trading seems to have its own movement.

To identify the crosses and intervene:
Im in direction of calculating different speeds of indexes on same TFs and of course also on the different ones. So, have relatively many 4H/1H indexes which show levels where one could consider some pause is happening and where the levels of potential high reward entries hide. This is not automated yet.

atm seemingly "good" pairs to trade imo:
uj s; acad b; au b; cadchf s; eg l; gchf l; eu l

Yes, im seeing that in my results on some occasions. Especially when currency pairs are not filtered out from higher TFs. I would add that one could try entries on the shorter TFs, especially when adding into a trend already established.

Good effort, as long as your equity graph goes from south west to north east, you're doing something right.

Keep it up.

Thanks mate ... North East is definitely my preferred direction too

I'm hoping the EURUSD trends a little more, I don't mind which way but there's almost a direct correlation between the EURUSD trend and the current equity curve.



Cheers,

mim2005

Hi betterlife,

Your comments are excellent and well thought out, it's given me much to think about and possibly set the stage for where I will need to focus my efforts in the future. It's only very recently that I've started thinking along the lines of whether there are quantitative approaches to assessing edges as mentioned by yourself and several other posters.

To date I have been more absorbed in finding 'any' reasonable edges using both deductive and inductive approaches and then assessing whether they have an explainable supporting theory/concept. I guess it can be very easy to 'fool' oneself into explaining why something works - so I acknowledge this is not very scientific. Although sometimes using your brain to work out that for example an institutional trader may need to under certain circumstance take profits at the end of a session can be much easier to do yourself than by any mathematical means.

I have no insight into the real working days of professional institutional traders and I have often thought if I could be a fly on the wall with these guys for a week observing the daily routing and reaction then I would be able to derive a lot more edges.

My strategy in the short term when I revisit the EAs will be to focus on finding a handful of additional edges using the approaches that have seemed to work so far. I figure you cant go go to far wrong in collecting as many as possible (across a few markets) and then focusing on a systematic way of maintaining/assessing/adapting and eliminating.

My goal is to essentially maintain a farm of EAs (hopefully 10-20 if I can find this many edges) and then focus on the problem of weeding out the poor ones and introducing new ones as markets change. The job then becomes the constant searching and maintenance of edges and survival of the fittest.

These two observations you've made in regards to system generation are very good and are providing me with much food for thought. I think it will be something that I need to jump into at some point and I will be really keen to come to you and possible collaborate in the "brain working" your doing

Your last paragraph in regards to the two approaches in finding edges, old style vs. automatic ... well I can't help but think that ultimately and edge is an edge ... it exist there regardless how you find it. Whether you pick it up from another system, mine huge quantities of data, establish concepts based on observation or hear it from an old experienced uncle ... use any approach you can - the real question becomes, are there effective quantitative approaches that you can use to assess edges (or possible edges)? and if there are then is it more efficient to use these than to use more qualitative approaches like establishing whether the edge makes some sense along the lines of the "stable patterns" you mention.

Thanks for your great quality posts, I have only been involved in trading for a little over a year and there is always so much to learn around the corner. Look forwards to keeping in touch.


Take Care

mim2005

Depends on many factors and the way you do it. Complex money management on each system is not fundamental, you can simply pile up genetically generated approaches that have been properly screened for edges and you can obtain very complex "smart management" just from the behavior of the portfolio. If you have 10,000 USD you can probably harness a 5 to 1 CAGR to maximum draw down till you hit 100,000, after that you will be restricted to about 2 and after 1 million it becomes hard to beat 1 to 1. However the above is not trivial it requires a lot of smarts. Cheers.

Great posts guys! Keep them coming

This is exactly what we (me along with 2 buddies) are trying to do. We are continuously building and improving a portfolio of EAs. It's never ending work.

Hi erik1,

thank you so much for this very nice reply. Well: I am ignorant, as I told. I have never seen anything like 5:1 CAGR/DD generated by algorithmic system selection so far.

...Cannot tell you how I would like to!

In the meanwhile I am making some significant progress in finding the formal solution to the problem of the 1 chance in 18 millions of the other thread!

Best regards

BL

Hi vox dei,

Thanks for the encouragement.

Completely agree in regards to the never ending work ... I think some people expect to find an edge, build a system and sit back and reap the rewards for several years ahead, but I think given the dynamic nature of markets this is unrealistic. It takes constant study and adaptability so you need to get used to this and enjoy that side of things or it will be way too difficult to persevere.

It's funny because one of the posters (permully) said recently "I put my bet against this system in 3 years time" ... and the first thing I thought is that it would be a gift from god if the systems could consistently make money from the same four edges over a three year period!

My biggest challenge so far is not lack of ideas but rather lack of time to pursue and work on everything. Not many people realise just how extraordinarily time consuming it is to program things. When I lead development teams I remember how slow things were even with 20 developer teams. When your working on your own with this sort of stuff it can take months full time just to accomplish basic things. Thats why you invariably need to team up with people to get through the work.

Anyway, thanks again for your comments and hope you become a prosperous farmer of EAs.

Take Care

mim2005

Hi mim2005,

Appreciate your kind words.

Completely agree with you. Profitable EA trading is not just about finding that illusive edge. The edge must be continuously maintained. Not only that, but one edge alone is simply not enough to allow a consistent return with a comfortable drawdown. The market can throw a number of different trading conditions at us whenever it feels like. Edges will surely suffer erosion, and may be completely nullified during long periods of time, even forever, with a enough change in market conditions. Thus, edges for different market conditions are required, and all these edges must be kept sharp, the same way risk needs to be assessed and adjusted periodically. A smooth ascending equity curve is a function of a lot of hard work, and a little luck.

Concerning my team journey, it is just starting. Last year was a great year (ar:dd about 6:1). This year has been rough (ar:dd 1:1 with dd exceeded). We were trading breakout strategies for the most part. Recently, our EA portfolio has been tweaked to be less market condition dependent. Atm, I feel very confident about the future

Wish you the best,
d

Hi all,

IMPORTANT: First and most importantly ... I watched this excellent video today, it's a relaxed interview with a highly experienced professional forex trader and I have found it absolutely EXCELLENT. It has confirmed many of the ideas that I've been thinking about in regards to how institutional traders work and how Algorithmic trading works and also provided me a little more personal insight into why some of the edges that my current EAs use work also.

MUST WATCH ->

Inserted Video

Secondly ...

Thanks for the feedback and encouragement many of you have provided over the last few weeks.

I have decided to remove the robot explorer from public view for a few reasons ...

1) Please be clear, I do not know if these robots will be profitable in live systems in the future. I am just testing concepts like most others.

2) I generally like to maintain a low profile and I'm not sure it's very smart providing too much information on live accounts.

3) I've been made more and more aware fairly recently that public trades can very typically be reversed engineered relatively easily and although I don't know if the systems I use are of any value to others I prefer to play it on the safe side and not have edges reverse engineered.

4) I'm not sure it provides a lot of additional value to me to have the trades public - I am a fairly honest and sincere trader and I plan to keep this journal updated on the robots progress manually (explorer screenshots etc.) but obviously precise trades will be hidden. Note: I will let readers know if the robots perform poorly, I have already acknowledged that this is very possible.

5) I'm sure people will say that I don't show trades because I am worried the system will perform poorly ... I hope you've read enough of my posts to realise I'm fairly honest and genuine about my work, learnings, struggles and challenges, there's no reason for me to hide from my systems performance whether they do well or poorly.


I plan to continue posting for a little while longer (a couple more weeks) but once I've documented everything I think I can give back I will likely start to contribute a lot less as I'm finding the posting a lot more time consuming and distracting than I ever imagined.

For the record I will leave a screenshot as to where the EAs got to by this weekend (17 days of trading)

Attached Image (click to enlarge)

Attached Image (click to enlarge)

Take Care,

mim2005

I'm starting this testing today. The four pairs I've manually selected from this morning's H4/D1/W1 correlation analysis are GC, GA, GJ and CF (note: the shorthand I use is A=AUD, C=CAD, F=CHF, E=EUR, G=GBP, J=JPY, N=NZD, U=USD). See attached screenshot.

One of the three EAs has been designed to run specifically on EU, but I'm going to run it on all pairs regardless. After all, this is just an experiment, and I can later segregate the results by EA, pair, etc.

I have no idea how the results will compare with mim2005's tests on EU. We'll have to wait and see.
I'll post trading statements periodically, to keep everybody who's following this thread informed.

David

THOUGHTS ABOUT WEALTH IN GENERAL

As I mentioned in earlier posts, I invest and trade my own capital full time for a living. But this is something I only started doing full time for the last two years. Prior to this I have been investing for at least 20 years.

One of the biggest challenges facing most people in this field is that they don't have the capital base to really invest and be able to live off. Of all the successful and relatively wealthy people I've met, regardless of the industry or business they work in I have found that almost all of them have grown their wealth slowly over time through hard work and perseverance.

I personally do not know any way of consistently making a lot of money quickly.

Just as importantly I would never trust anyone to manage my money if they hadn't gone through the slow grind of growing wealth. I honestly believe that most people are set to experience life at a certain level of wealth, this is most likely the current level you are experiencing life at the moment. If life was to re-shuffle the deck and the population was all given an equal amount of starting wealth I wouldn't be surprised if those previously wealthy would be much more likely to come wealthy once again ... why ? because they behave and think differently to most others. In order for most people to reach the next level in wealth they will need to change the way they behave and think.

I was lucky that when I was a young man I read several books regarding wealth building, but one that stuck in my mind was "The richest man in Babylon". From these teaching I learned some very simple rules which I have followed all my life.

There is nothing magical in these rules, they are simply what I have leaned and used successfully over time. Most people don't want to hear these rules ... they are too slow

1) Maximise your income in every way possible (go for a high paying career, regular promotions, earn income on the side somehow).

2) Keep your expenses fairly lean (don't get caught into fancy cars or materialistic possessions etc.)

3) Always save and grow the difference (between income and expenses) ... this capital is the engine of your wealth, each coin you accumulate will help you to accumulate more.

4) Become knowledgeable enough to learn how to regularly invest your capital at a return that is reasonably more than inflation (I aim for at least 10%-20% per annum across all capital)

5) Once your confident you can consistently achieve rule 4 ... Start leveraging and using other peoples money to also increase your wealth (within reason)

6) Never, ever, ever, ever lose any significant part of your capital


Thats my journal entry for the day

Take Care,

mim2005

In the last post I spoke about how I had worked for a long time to build up a capital base (and continue to do so). I think many traders probably don't realise how important this is.

I believe growing a capital base is just as important as knowing how to trade profitably. The capital base is just as important if not more important than any tool in trading. In most cases, without this it's very difficult to make a significant income and when mixed with having to live from your own trading income it becomes a very stressful exercise. The larger your capital base the more diversified you can be in your investment and trading approaches and also the larger the safety buffer you have when things don't go according to plan.

I'm sure some of you will think it's too late and you don't have time to build up your capital or you think you can trade with a small account taking high risks, get lucky and build a large account and then start protecting and growing your capital. I would advise that this is not optional and that if you don't have the discipline to simply grow your capital then how will you have the discipline to trade effectively? Even if you do get lucky in the short term there is a good possibility you won't have the skill developed to simply protect it and grow it at the later stages.

So ... learn to be disciplined and build your capital base and never ever lose it.


PORTFOLIOS, DIVERSIFICATION AND LEARNING DIFFERENT STYLES AND MARKETS

When starting to trade and invest full time, one of they main challenges I had was in knowing what I should be investing and trading in. I am a firm believer in not limiting yourself to any single style or market but rather explore as much as possible until you find the styles that work for you. Given I was relatively new to many types of investments and trading I decided I would try many different investment and trading styles.

The approach that I used was fairly straight forwards, I listed all the different types of investments and trading that I would study and undertake with what I thought the realistic percentage growth rate I felt could be achieve by an experienced investor/trader. Furthermore I looked for courses/mentors and tutors in each area to try to learn an already working strategy rather than to try to come up with something myself which I knew would be too time consuming. The types of things I had on my list included things such as:

a) Long term equities (local and global)
b) Short term equities swing trading via CFDs
c) High interest/yield type investments
d) Options trading and writing covered calls
e) Forex trading (intra day trading, longer term trading), discretionary, mechanical, technical, fundamental.
f) Futures trading (eMini ES etc.)

Over the course of a year and a half took courses on all of these (except options) and studies several strategies in fairly extensive detail. I invested/traded in each one with small live accounts until I became comfortable with the styles and was able to achieve a level that I felt I could be profitable. Over the course of a year I worked on building a portfolio with specific asset allocated to different approaches. Note: There were several approaches which I could not be comfortable with or I couldn't be profitable with - with these I simply discarded them.

My asset allocation is divided into 5 major categories:
1) Cash (mixture of yield/interest approaches)
2) Australian Equities (3 different approaches, one being a fund manager who I compete with as a measure of my own performance)
3) Global Equities (Primarily through hedge funds)
4) Property
5) Speculative (Forex EAs, eMini S&P discretionary and others)

The bulk of my capital is held in areas I am more comfortable with and typically these produce lower rates of return. I place relatively small amounts of capital in areas I am still studying and researching such as EAs etc.. but conversely these also offer some of the best returns. It's only when a strategy proves itself that I decide to start allocating more capital.


This all sounds a little elaborate and over the top but one of the biggest benefits I have found is that the diversification brings me a lot of peace of mind. There are many instances where one of the strategies or approaches has performed poorly and I have taken a reasonable hit but it doesn't bother me too much as there is often another strategy or asset thats performing well over the same period. This has enabled me to persevere with many strategies that if I was only trading them alone would be very hard to persist with psychologically.

So ... try to diversify your strategies and approaches as early as possible ... Personally I found it way to hard to psychologically handle the ups and downs of only one strategy/market and I would never have the courage to invest all my capital in one area.

As with all of this ... remember I am only journaling my experiences ... and I consider myself to be relatively new and inexperienced in this career so please don't take what I say as being the only approach ... it's simply the approach I chose to follow.

Hope this helps a little.

mim2005

I took some time to write what I thought was a valuable post, but FF decided to 'eat' it. Oh well :/

I'll just say thanks for the great posts, mim2005. Keep them coming!

Hey vox dei,

You can't just leave us hanging like that ... I am now really curious about what you had to say!

I've had the same thing happen to me before and it's bloody frustrating ...

I'll be waiting patiently

Take Care,

mim2005

Hi people,

very nice and important thoughts, surely. Keep them up.

When writing important things, make it on notepad and then copy-paste them here. Strange things happen much too often...

Bye
BL

Hey guys,

Last post resulted in unintended suspense and high expectation. Should have seen that coming. I'll get back to it given time and especially the mood. I'll just stress that it was a valuable post from my perspective. Not necessarily from anyone else's




Thanks for the advice 20/20 for me in hindsight.

Later,
d