METHOD vs SYSTEM and the destructive EA

I am not a forex trader at present, so it will therefore not come as any surprise that I am not intending to present you with a "system" in this posting. (How many will close this thread at this point, I wonder). I am, however, an experienced and profitable index futures trader who is happy to net his 5-15% return per month, consistently, from around 3-8 high probability, low risk trades only, withdrawing some profit as income, whilst continuing to grow capital. So what the hell am I doing talking about systems on a forex forum?

Well, doing only 3-8 trades per month does give one some additional research and education time and the world of forex does interest me for reasons I wont elaborate on here. I have read dozens of threads, some of them end to end, hundreds of postings, and then recently decided to quietly participate in discussions here and there. My humble view is that successful trading is all about "learning how to learn," and I have been doing quite some learning in recent weeks here at forexfactory.com, which I should add is a brilliant resource for traders and whose founders, administrators, and major contributors deserve all the credit in the world.

The thing that has struck me the most about the postings I have read is that there appears to be an incessant desire to automate everything that looks like it MIGHT be a profitable system just because someone can produce 3 months of profitable demo account results (at best) - "Have you written an EA for that?" "Billy blogs is great with EA's - let's get him him to write one for this system." And then... you know how it goes... the focus becomes on automating an untested system and "getting the EA bugs fixed," and the opportunity to properly develop a potentially profitable trading methodology becomes totally lost.

Result? Learning = 0. Profitability = less than 0. Frustration = off the scale.

Don't get me wrong, I do think there is a time and place for EA's - in fact I recently started to use some very simple ones for things like closing multiple orders on hedge trades (no - I never hedge the same instrument against itself!!!), but at the same time could quite happily continue to trade manually without ever caring for an EA again. And indeed I do have a lot of respect for those who have made the effort to learn how to write complex EA's and assist others who are seeking help with automation. But let's be honest about this - is the majority of EA writing and development going on here assisting well intentioned traders to become better traders, or is it distracting them from developing the competencies and experience required to become a life-long profitable traders? I would argue very strongly that it's not difficult to develop a profitable system for yesterday's data, or that work well for the previous 6 months, but it is significantly more challenging to develop the skills, experience and competencies required to become a life-long profitable trader...

So what's the point of all this rambling from a smart-arse derivatives trader who calls himself quiet, modest and humble? Well I think it comes down to the difference between a trading METHOD and a trading SYSTEM. This, I believe is a topic of massive importance to the process of learning, and therefore to the process of successful trading (though I would argue they are part of the same process), but one which is also difficult to fully elaborate on in a single posting - so hence a whole new thread!

So as briefly as possible (difficult for me) - you all know what a SYSTEM is, right? Well, I define it as the kind of thing most people start talking about when they say - "so, tell me about your system." ie the actual mechanics of and conditions for when you enter and exit a trade. It is also the thing that everyone wants an EA for when they have a feeling that a system might be profitable - oh yeah, and a “good” SYSTEM will also have EA's which cover things like "money management" - eg the robots which people buy on the internet from expert marketeers - strange then how people still manage to blow an account with a good system and a good EA. Utter BS, to put it mildly.

So what's the alternative, and where does profitable trading reside? (ok - call it "holy grail" if you must) The only word I can think of is METHOD... A trading METHOD, as opposed to a system, is an end to end process which encompasses things like research, self-education, documentation, exploration and definition of theories, testing, critical thinking, further development, further testing, further analsysis, good record keeping.. and the list goes on - of course live trading is just one part of the overall METHOD of successful trading. And, therefore not surprisingly,


PROFITABLE TRADING IS JUST THE OUTCOME OF THE TOTAL METHOD - IT IS NOT AN ACTIVITY UNTO ITSELF,


as the incessant desire to EA someone else's system would suggest. As an ex management consultant, I could probably construct some sort of dynamic model to depict what I mean by METHOD, but that would make us all want to puke, and probably describes in a nutshell why I left that world for the joys of trading for a living - of course screwing up massively at various points in the journey to date.

So why I am posting all of this? Well I have no intention of setting up a formal coaching programme and have absolutely no commercial interest in anything associated with the above rantings and ravings but I do think my comments are perhaps relevant (call it advice if you want to), based on the prevalence of EA-ing crappy SYSTEMS that have not been properly tested and which are not part of an overall trading METHOD that I see littered all over this (otherwise brilliant) website.

So by all means, I am in favour of using EA's when you have established consistent profitability, yet continue the process of learning within your method - so the EA, just helps with efficiency, speed of execution, and perhaps accuracy. But for God's sake (anyone's God that is - no I am not religious but I might be a dyslexic, agnostic insomniac who stays awake at night wondering if there really is a DOG), do not rob yourself of the opportunity of develop a truly profitable trading METHOD by falling for the "let's EA it" trap.


... if you are a new trader and have successfully reached the end of this post, then I (humbly) suggest that you are already a significant step ahead of all the folks who exited after reading the first line. And if you wish to disagree with me please do so, because I hope to never stop learning how to become a better trader.


Perhaps this thread will die a slow death, but perhaps it may start a useful discussion on the benefits of a METHOD vs SYSTEM and indeed what a method entails – something which has been probably covered before in this forum, but which I suggest is worthy of ongoing discussion.


One thing I am confident in saying though – the gradual evolution of your comprehensive trading METHOD has significantly more opportunity to embed learning into your trading approach than the adoption of any SYSTEM can ever hope to – and I am quite sure that my “holy grail” only came into being when I learnt how to learn.

Good trading to all.

I agree.

Hedgginghog,

Thank you for your interesting post, I love to fool around with systems or methods. I agree on the most parts of the post (so maybe this is an addition :-) ).

However i think that you can learn alot building EA's but not the ones you cannot backtest good.

I have a strong programming background and i have build my own programs to backtest my methods. This way i can measure every tiny bit of information.

Backtesting is very very important in my opinion, not for the possible profits but you must ask yourself "which results proof that my system works " ?

A few things that i want to see in my results

Alot of trades (the more the better for accuracy not spreads etc)
A climing equity curve without a lot of deviation
pips of profit must be at least 1/5 of the average tradesize (excluding spread)

the method must be simple, not more than 6 rules at most.

For me backtesting opened my eyes, first because you need to think wat good results are. second because you must do everything step by step.

When i start to investigate a certain system/idea i build it so that it is a always in the market system, without stoplosses en takeprofits. that way you can truly measure if the system works, when you get around the 1000 trades over a period of 3 years almost every curve-fitting-method will be evened out, offcourse when i stick to the max 6 rules per system

Sorry for my bad english and somehow unordered answer. cheers !

Hi there Kaas - thank you for your constructive comments, and based on what you say it would seem like you are using the EA's as part of a broader trading METHOD, which involves testing and re-testing, and the EA's are just one part of the method.. as opposed to the whole focus of what you are doing - therefore you are not "blind" to the realities of the market - I hope this continues to go well for you.

Interesting point you make about volume of trades - I suppose everyone finds what works well for them, and what they are comfortable with. Part of my realising CONSISTENT profitability in my own trading was removing around 80% of the trades in my 'trading plan' and leaving only those that both deliver high probability AND low risk trades - hence my doing just a handful of trades per month - when I started out trading, I always assumed I would be doing trades every day or at least every other day, but it was only when I removed this assumption from my trading method that I actually started to make consistent profits from the market.

Cheers.

Mr. Hedge (may I call you 'hedge'?)

I found your post enlightening and am heartened by your words. My battle has been just those things you talk about. I have studied the markets many years and feel I have learned much, but have always tried to automate my "system". I had to learn to write mq4 as I am not a programmer or anything close to it.

My lightbulb moment came after the 100th iteration of the 1,000th "system" that there is no program...none what so ever...that can instantly take into account all the variables present in a single chart and compute them with any accuracy for more than a few months.

After that the program needs to be "adjusted" and "optimized" and....the list goes on and on. The only software that does this is between your ears and does more in a single glance than 1,000 pages of code.

I am now back to learning again only this time learning markets and their behavior. My approach is simple. Be the "rat in the maze"...when you hit a wall, turn around and go the other way.

A word about "backtesting" from the EA stand point. While it can do 1,000's of trades and execute them "perfectly" according to its algorythm, it cannot take into account real time decisions and adapt to changing conditions. It will keep pounding away at that same wall until you run out of money. That's why backtesting historical data is always skewed. The same condition that would cause any sensible person to say, "perhaps we should look at another pair", or "lets leave this one be until later", would cause a ruination of your account....three in fact.

While I would certainly hesitate to call it the "holy grail", I will say that the "shut up and trade" approach has worked far better for me in the long run.

Hello gspajon - I enjoyed reading your post very much. In fact it is probably one of the most intelligent, useful and well written posts I have read on this whole site - your description of the EA vs the mind is brilliant, and frustrated traders would do well to take note if your words. Thank you and I wish you every success in your trading journey.

I disagree and agree with your post , i agree on the part that there are a lot (almost 95% in my view) of bad EA's, but it is possible to build a simple trading 'ea' that runs OK, but to do that we need to go back to the basics.

For example, a computer is able to define a current trend much better and more accurate than a non-computer with a certain math formula. This can be used in an EA.

I create my own EA's based on market basics and i will test other idea's.

For EA's you must stuck to the basics so not more than 3 or 4 rules to buy and to sell and not complex one's, for example : 'if the trend is up longer than 4 units of my tameframe BUY' etc.

EA's like buy if we have 6 bars that close higher then the open are not good or not working in my opinion, there is not enough math behind it, what if the bar is really small and the difference between close and open is 1 pip, does that count ? or what if the pair is really volatile and hoed from high/low/high/low to finally a close higher than an open.

Kaas...You are obviously a far better programmer than I...(which in truth so is pretty much anyone). However your point about getting back to basics is well taken. For in programing it is all about "definition". You see the computer needs an algebratic formula to define...well everthing. Its "logic" is simple and uncomplicated...but markets unfortunately are NOT.

Where a "non-computer"...human brain...would take one glance at a chart and say, "long trend", or "ranging"...a computer would need something like you described...if H1 > H2 && L1 > L2, then if MA1 > MA2...etc. While there are many "systems" that do just this the eternal flaw will come in just letting them go with no "non-computer" intervention.

While I still use EA's for scaling in and scaling out, they are just fancy calculators that save me the time and effort to calculate the "math" on any give position...yes I'm lazy too. But they are just a tool, like anything else. I am the master carpinter and I chose WHEN and what direction and for how long. Again, while I am no expert I believe that the "logic" my brain executes to make these decisions in the blink of an eye, would take 1,000's of pages to write assuming one could indeed account for all the variable that go into even a simple decision.

While I know we are NOT supposed to use emotions in trading (thus the big draw for automation), I find that emotions whether good or bad are very useful. If you've ever participated in any sport you will understand. We use our "fear" to alert us to danger...we do NOT allow it to overwhelm us but it is that emotion that allows us to be prudent...we take protective measures when appropriate. If you have ever ridden a bicycle did you where a helmut? If so why? If not why not? For those of us who have fallen and hit our head...or seen a friend fall we know it only makes good sense to wear a helmut...we don't stop riding, but we take protective measures. Likewise as we gain experience, we learn which roads are safer than others to ride and under what conditions they are safest. This does not GUARANTEE a safe ride but puts the odds in our favor...sound familiar? Emotions are not part of a computer program and therefore can never understand that little "tickle" in the back of your head that says..."get out now", or "tighten your stop", or "hhmmm, don't enter that". Granted these feelings need to be trained like any other profession, but when done will be far superior to any program whatsoever.

Which brings us back to the original point of this tread...learn to learn.

System is what i started with, entry on signal and exit on opposite signal... this left me with b/e trades most of the time. My system has now turned to method/stretegy which enables me to hold profitable ones longer thus minimizing risk. As previous mentioned trading is an ongoing challenge/battle not versus the market and other traders but against yourself.

The purpose of developing and debugging an EA actually is testing the untested system. How else could you test a system without first writing a working EA for it?

Hi 7bit. Thanks for your input. I think if you put that quote back into the context of the overall post it, then your question is more easily answered.

Yes, there is a time and place for EA's, but as stated above, so many "systems" discussed on this website become totally molested to the point of uselessness, simply because every wants to "EA it." The majority of traders I know who are consistently successful in the market - whatever kind of trading market that is, learnt how to trade without the use of EA's, or their equivalent, and when appropriate added in automation later. EVERY single person I have spoken to who has started off trading down the automation path, without a proper method behind them, has lost money, and sometimes a lot of it - very few of them are trading anymore, and the one's who are still trading don't use much automation at all.

How to test a system without an EA? If you have not already, get out a pen, paper, calculator, and excel spreadsheet, and sit and watch the market for hours, days, weeks, and months, with as few indicators on your chart as possible. It takes longer than writing an EA, but my god it's amazing how much you can learn...

Brokers love giving their clients the ability to use EA's, and as many indicators as possible - they know that in most cases (not all of course), they will loose their money faster that way, and top up their account again, usually twice before totally quitting and moving on to something else.

I'm not trying to argue with you, because I would rather encourage healthy debate, and everyone's opinion can be respected. All the best.

Cheers.

Hi Hedginghog,
I've got to admit that what caught my eye is your
statement that you obtain a monthly return of 5%-15%, with only
3 to 8 trades in this time frame.

Correct - some time ago I used to trade every day (or at least every other day) and I had a trading plan a mile long, full of trades that made a bit of money over the course of a year, but that made my account suffer large drawdowns from time to time, even though it would be profitable "in the end" - I just couldn't handle that scenario. I learnt something very, very important about my risk profile here.

So I stripped away all the crappy trades, and ended up with just a handful of trades that trigger each month - each of them has the profile of being both high probability AND low risk. Of course you can still have a high probability trade eg 80% likely to win, but sometimes you might be screwed if it fails - as I found out where the risk is hidden away in your trades and you only find out the hard way..(what happened? the backtesting didn't show it? It can't be right? etc, etc) . Hence the required probability and risk profile of a single trade, and sometimes it is very difficult to determine that RISK from the backtesting (especially in a bloody EA!!) because that risk may lie in your head, not in the data...

But back to your original comment - not surprisingly the monthly profit is proportional to volume of trades (in my situation at least). So last month there were 11 trades (the most since Feb this year) but this equated to an above average 18% return. So far this month, just 2 trades and 3% return on capital for the month to date - if that's all I get for the month before Christmas, then so be it -I wont go searching for more trades that aren't there..

....Update after I finished the post - make that 3 trades and 2% for the month (yes, I just had another "enjoyable loss" which took 1% off my capital) - I know I am doing my job properly when I have those kinds of losses!!!

OK - let's get to the crux of the thread: A trading METHOD, and learning how to learn.

Snippets will not appear in any specific order - I'd like to briefly cover one part of the METHOD. From theory to live trading. The below flow should be familiar to most of us:

(not an exhaustive list, just quickly belted out..)
1) Notice something that seems to happen regularly
2) Question - I wonder if I can trade that?
3) Definition of a theory - eg when price does X then I do the following....
4) Set up rules for testing - entry, exits, etc
5) Run back-test
6) Check results - alter rules and re-test to see if results get better.. or you realise that this probably wont be profitable for you leave it alone.
7) If results are good - set up forward test with demo account
8) Am I profitable on demo? What do I need to tweak to make it more profitable? Or close the door on this 'system' if it wont be profitable for you
9) Re-run live demo trading with improved rules. Can I be profitable with this, and does it suit my "style" of trading and my comfort levels?
10) Go live, track results and observations
11) Continue to refine on DEMO whilst live trading (if profitable). If not profitable the STOP and go back...
... timeframe, for this single cycle? Maybe 4-6 months as a guideline?

There are arguments within arguments about all of the above (eg the real value of backtesting?), but I'd like to briefly focus on where I think many people go wrong in point #6, as indeed I had done many times:

NOT REALISING WHAT THE DATA IS REALLY TELLING YOU AND FOOLING YOURSELF INTO THINKING THAT YOU ARE LOOKING AT A "GOOD SYSTEM".

How many bad trades did you ignore and remove from your results because it was an "anomaly?"
Did you look only at the trades that succeeded and see the $ signs in your eyes?
Why did you not look carefully at the trades that failed and perform a comprehensive risk assessment of what this will mean in live trading?
Is your sample large enough? How would it perform in markert conditions that are different to when the sample was taken?
WHAT ARE THE ASSUMPTIONS IN THESE TEST RESULTS THAT HAVE NOT BEEN PROPERLY TESTED?
Why did you ignore the heard of elephants in the room?

Did you say "I must succeed as a trader, and therefore I will make this system work," or did you realise that this system was not likely to succeed, and instead document what you had learned with the hope of applying it to your next test with something else?

Did you find a way to enjoy the learning process (if for no other reason then it might save you going broke), or is that kind of analysis, critical thinking, and self examination just getting in the way of you wanting to make money.. fast?

I recently saw in a posting some words from one of the most prolific EA contributors on this site (who is no doubt also a very well respected and well meaning person), words to the effect of:

"I'm great at writing EA's, but I'm not a very good trader.."

In a different context, another well-meaning trader had identified a way to identify strength in one pair by looking at two other pairs that are both related to that pair.... (er..um.. instead of just looking at a simple chart for the identified pair to identify underlying strength and weakness?)

Our famous EA writer then jumps in to help everyone out and ends up writing an EA to try to "simplify" a trading execution process for a very simple trade, but which has been complicated by the incessant desire to complicate something that is otherwise very simple... and on... and on we go.

This is all well meaning stuff going on here, so I should be careful not to be too harsh on the respective contributors, but really... are we talking about help, or hinderance???

On the other hand I just saw an excellent thread which is seeking to talk about the value of PATIENCE in one's trading method, and by contrast it seems to have comparitively little interest... perhaps because it is not a "system" that can be EA'ed... hmmmm.

Interesting thread, compounding a lot of experience. Maybe you could share a chart or two, I mean your high probability index setups? After all this is your thread, and if we can learn a thing or two, why not?

Hi Soekitox, Thanks for your suggestion. Let me briefly outline the background of some of my set ups - but at this stage I don't intend to share the exact mechanics of my trades for several reasons:

1) This thread is intended to be about the learning process of trading, the benefits of manual trading creating the essential learning that automation is unlikely to provide, and some aspects of a comprehensive "method." I don't want to change the purpose of the thread. This is all about learning how to learn.

2) You can already guess the other reason. The simplicity of my set-ups are easily open to commercialisation/'theft'

3) I am bound by a confidentiality agreement for a small part of my trading process, which I would never want to come even close to breaching, even within the comparitive anonimity of such a forum, as it is a matter of personal integrity

3) These are not fx trades, and any detailed discussion of my method should belong in a different part of the forums, though based on the above points I do believe the "learning process" associated with different instruments to be the same.

Anyway - two major trading types I can share:

1) Trading the seasonal moves of the S&P e-mini and correlated world indices, which most often occur around US public holidays - just before and just after. Your public holiday patterns in US are easy to trade because the holidays fall on "2nd Monday of month" (for example) and therefore the day of the week and time of the month are consistent, whereas in other parts of the world, we need to contend with "June 15th" as a holiday, which might fall on a different day each year, and is therefore open to more variation. The public holiday effect on the S&P sometimes flows to other world indices also. Sometimes oil is good to trade for the same reasons, and these are well known "mini-cycles" for want of a better term.. If anyone is interested check out your S&P500 chart before and after US holidays and look for occurances with 70% or greater probability (80%+ is better). You need to look at many years of data.

2) The S&P500 is obviously a major world index, and other indices like to "follow" what the S&P is doing, but at other times, they like to follow the events of their local market. It gets interesting because obviously the index futures markets are almost 24 hours running, not unlike fx - though we have more breaks in the index futures. So the day/cash market in US might correspond with (Part of) the night futures market in UK or the entire night futures market in parts of Asia. But there are times when the related index might want to do it's own thing for a while, before it then 'remembers' to follow the S&P lead, or it 'remembers' that it is meant to follow it's local market, not the Americans, and this presents some good trading opportunities at some times of the daily trading cycle and at some times of the week.

Trade #1 can be forecasted and planned for. Trade #2, while it falls within clearly defineable trading ranges, requires a lot of patience.. sitting.. watching.. waiting... and learning.

I rarely use any indicators - just plain charts - just price action, and highly disciplined yet very simple money management.. Not a negative month in the last 10; around 5-15% per month on average (not spectacular, but more than enough for me). PATIENCE is the key for me - the first 2 weeks of December yielded little over 1%, but this last week has popped up several trades and given 8% after another trade closed this morning. No bragging going on here - just talking about the realities of what has worked for me and why a comprehensive method, with ongoing learning, has been crucial to my realising consistent success. This only came about after I stopped trying to copy other systems and I learnt how to learn - like an undergrad, then post grad degree I suppose. It feels like I am now I am onto my masters, and all going well hope to reach PHD level (but no books, thank you!).

Anyway, that's all I intend to cover for now - I'd rather stick to the 'learning process' if I may. Cheers.

I hope that my current index futures trading methods will serve me very well for years to come, but I have decided to try to add another string to my bow - just for fun & general learning at first, but perhaps it may deliver something more substantial? I intend to do the MAX lite course in January. (see thread in commercial section).

I have recently been in contact with Jerry (aka Beach Bum) and despite the couple of very critical comments from people who have not done the course in the afrementioned thread, on the face of it, the learning method for the complete courses (Lite course being just an intro I think) appears to be very comprehensive and almost unique in fact, in a trading education market that has otherwise become quite commoditised and generally deserving of the scathing reviews by many.

I will reserve final judgement for later (ie when trading the Lite 'method') and look forward to posting a fair and honest review at the right time, but the goal is to prove the basic method by trading $1 'Lite' pips until the complete course has been paid for, should I chose to do it later... should be a fun challenge I have set myself anyway..

Hello Mr Hedging
Learning how to learn is really do most difficult to master. Ive manually backtested a lot of systems with all kind of indicators and found that it is very easy to have very profitable system , optimized on historical data. Unfortunately this somehow doesn't work in real market. Its like CCI forgets that during last 2 years price was following every hook, and stochastic suddenly doesn't care its overbought. And how many losers are just not count because in reality I wouldn't take it (for whatever reason).
I have no idea how to make backtest to have at least some value and it doesnt matter if its based on indicators or pure price.
Ok , I haven't been using indicators for last 2 years but again, in backtest u dont know what is market sentiment, what fundamentals are pushing the price, when are main news released...A lot of times a see people writing...price action doesn't look good so I dont take it, momentum is strong so be careful, PA is telling me to do this and that ....and so on.
And at the end most of these people end up with selling their courses, signals or chat room fees (which is not prohibited anymore).
This is very good thread , it is all what I have already thought about but hard work (not even 10 000 hours) is not enough if u are not lucky to put everything together before u wipe out your account.

Hello alter - you point out many hard 'thruths' in your posting. Thank you.

I have started up a new thread in the rookie section (less 'noise' over there I think and some practical discussions that focus on the basics which many people seem to forget about!) with the link below. Based on your comments above, you may find it interesting. All the best.

http://www.forexfactory.com/showthread.php?t=222471