The only thing I have learn | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

I want to learn and learn 53 replies

Its better learn wisdom late, than never to learn it at all 8 replies

Only 1 good thing 39 replies

The Only Thing You Should Try and Pick is Your Nose 0 replies

Lessons you learn only when you blow up your account. 7 replies

The only thing I have learn

  • #1
  • First Post: Mar 18, 2010 6:29pm Mar 18, 2010 6:29pm
This is the only thing I have learn after 1 year of trading, and I'm sure that it's completely true:

- You have to follow your method, if you don't have a method or you don't follow it, you will fail.
- If you have a method, you can write it and backtest it from past 10 years. If your method can't be coded, you don't have method. If you test your method back 10 years and your method fails, you will fail.

If you don't follow literally the 2 points of above and you don't fail, that's only because 2 reasons:
- You have privileged information or
- You have too much lucky, but you will lose it soon

Wake up! Human intuition doesn't exist.
Admit the truth: YOU CAN'T PREDICT THE FUTURE.

I think this truth is unquestionable, but I have found it after a year trading, so my goal now is: if I can't write a profitable EA for the past 10 years, I will not ever trade.

Because the only reasonable thing in this market is to write an EA and put it to work, there is no reason to trade manually, that's because 95% of traders fail, the rest 5% are brokers, traders with good EAs or very very lucky traders (maybe the 0'1%). You can say to me: hey man, I only trade during news and there is no EA for that, well, YOU'RE WRONG, you can write an EA for trade during news times, in fact there are lots of EAs that take the hour of the new from a web page and trade it. YOU HAVE NO EXCUSE, if you don't write a method that works for the past 10 years, you always fail, doesn't matter that you trade automatically or manually.

You will not find a profitable commercial EA, if I would have a profitable EA, why to sell it? I could became rich by using it, not selling it. So YOU have to learn and write it, there is no other way.

This is the truth, admit it.
  • #2
  • Mar 18, 2010 6:43pm Mar 18, 2010 6:43pm
How on earth should I backtest an EA which uses 10 sec, 1 min, 5 min and daily timeframes, and which uses limit, breakeven, etc etc orders ?

If you have the answer don't hesitate to share it with me
  • #3
  • Mar 18, 2010 6:44pm Mar 18, 2010 6:44pm
The problem "I" have with this line of thought is that it's all probability based. The problem with building a method based on probabilities is that even with 10,000 trade examples, trade #10,001 could be a long tail event that makes your system a loser.

I would be interested in discovering how you convince yourself that 10 years of backtesting is sufficient to establish the probability of future earnings?
  • #4
  • Mar 18, 2010 6:58pm Mar 18, 2010 6:58pm
Quoting Trad3r
This is the only thing I have learn after 1 year of trading, and I'm sure that it's completely true:

- You have to follow your method, if you don't have a method or you don't follow it, you will fail.
- If you have a method, you can write it and backtest it from past 10 years. If your method can't be coded, you don't have method. If you test your method back 10 years and your method fails, you will fail.

If you don't follow literally the 2 points of above and you don't fail, that's only because 2 reasons:
- You have privileged...

................
  • #5
  • Mar 18, 2010 6:59pm Mar 18, 2010 6:59pm
Quoting Darkstar
The problem "I" have with this line of thought is that it's all probability based. The problem with building a method based on probabilities is that even with 10,000 trade examples, trade #10,001 could be a long tail event that makes your system a loser.
Daytrade and use stop losses and avoid unusually high spreads (news, etc).
  • #6
  • Mar 18, 2010 7:14pm Mar 18, 2010 7:14pm
Can you get codes for trendlines, flags, pennants etc....Im not a programmer

Or maybe Im reading it wrong?
  • #7
  • Mar 18, 2010 7:26pm Mar 18, 2010 7:26pm
Quoting Trad3r
This is the only thing I have learn after 1 year of trading, and I'm sure that it's completely true:
What I learned the first year was that "what I knew to be correct - may not be". Perhaps that will be what you learn in year 2.
In trading, there is no bullshit. You either make money or you don't.
  • #8
  • Mar 18, 2010 7:39pm Mar 18, 2010 7:39pm
Quoting Trad3r
This is the only thing I have learn after 1 year of trading, and I'm sure that it's completely true:

- You have to follow your method, if you don't have a method or you don't follow it, you will fail.
- If you have a method, you can write it and backtest it from past 10 years. If your method can't be coded, you don't have method. If you test your method back 10 years and your method fails, you will fail.

If you don't follow literally the 2 points of above and you don't fail, that's only because 2 reasons:
- You have privileged information...
long way to go.
  • #9
  • Mar 18, 2010 7:43pm Mar 18, 2010 7:43pm
Quoting Trad3r
This is the only thing I have learn after 1 year of trading, and I'm sure that it's completely true:

- You have to follow your method, if you don't have a method or you don't follow it, you will fail.
- If you have a method, you can write it and backtest it from past 10 years. If your method can't be coded, you don't have method. If you test your method back 10 years and your method fails, you will fail.

If you don't follow literally the 2 points of above and you don't fail, that's only because 2 reasons:
- You have privileged...
Hello Trad3r - please don't take any of my statements in this post as being arrogant, because that is not my intention, but your first post seems quite indicative of someone who has been trading for about 1 year.. probably some useful realisations have been made (ie essential to have a method) but still quite some learning to go...

After having been trading for 8+ years (ok not a lifetime but getting there), and trading full time for a couple of years with consistent profitability, I have developed quite a different view to you on the EA and coding statement:

My own experience, and that of every other full time and professional trader who I know, suggests that most of the best trading methods cannot be coded to realise consistent profitability.. and that most 'systems' that can be coded will eventually fail..

There will be many who will disagree with me (and that's fine by me, and I don't want to get into any heated debates on this issue) but this is just what I have experienced and therefore one of my 'truths' in trading.. if you find consistent, lifelong profitability in EA's, then good for you, and your 'truth' will be different to mine.. but in a forum like this, and with an original post like yours, it sometimes pays to be open to fresh perspectives, especially when you are still very early on in the trading journey...

In think "mrgreen" sums it up better than I did here..

Wishing you ever success in your trading. Cheers.
  • #10
  • Mar 18, 2010 7:54pm Mar 18, 2010 7:54pm
Quoting Darkstar
The problem "I" have with this line of thought is that it's all probability based. The problem with building a method based on probabilities is that even with 10,000 trade examples, trade #10,001 could be a long tail event that makes your system a loser.
As another trader has told you to solve that problem you have to use SLs, so trade 10,001 will be only a lose trade, you will not lose everything in just 1 trade if you use money management.

Quoting Darkstar
I would be interested in discovering how you convince yourself that 10 years of backtesting is sufficient to establish the probability of future earnings?
Because the last 10 years have been the most volatile and amazing years in financial markets: we have had 2 bubbles, 2 wars, the worst terrorist attact to USA, we have been closer to a financial and system default... If your EA can survive to all of those things, you have a high probabilities that in next 10 years you will get similar earnings, because people react equal to the same things, and your EA will be prepared to almost all possible situations. I don't know how markets will react if an asteroid impacts in Wall Street or if we are invaded by extraterrestrial, in these situations your EA could fail because they have not tested
  • #11
  • Mar 18, 2010 8:01pm Mar 18, 2010 8:01pm
Quoting Hedginghog
My own experience, and that of every other full time and professional trader who I know, suggests that most of the best trading methods cannot be coded to realise consistent profitability.. and that most 'systems' that can be coded will eventually fail..
Tell me what you can't code, I'm sure I can find a code to write it, maybe a long and complicated code, but there must be.
  • #12
  • Mar 18, 2010 8:06pm Mar 18, 2010 8:06pm
Quoting Trad3r
Because the last 10 years have been the most volatile and amazing years in financial markets: we have had 2 bubbles, 2 wars, the worst terrorist attact to USA, we have been closer to a financial and system default... If your EA can survive to all of those things, you have a high probabilities that in next 10 years you will get similar earnings, because people react equal to the same things, and your EA will be prepared to almost all possible situations. I don't know how markets will react if an asteroid impacts in Wall Street or if we are invaded...
all those thing is good for manual trading.
i can give u a simple moving avarage system that will show you it did make money in last 10 years but in real time it will empty ur account
ea is for backtest.not for real time trading.imho

sho
  • #13
  • Mar 18, 2010 8:09pm Mar 18, 2010 8:09pm
Quoting Trad3r
Tell me what you can't code, I'm sure I can find a code to write it, maybe a long and complicated code, but there must be.
if you can code come and see my thread and code them.you will see all the rules and how i trade.
i am sure you wont be able to
  • #14
  • Edited 8:27pm Mar 18, 2010 8:12pm | Edited 8:27pm
So what have you learned again.... btw that's a crapload of IFs your EA will have to survive. Lookup Ronald Raygun on the forums he uses EAs but has no foolhardy belief that an EA can run constantly and adapt to all market conditions. He uses his own judgment to turn them on and off.... Btw who are you trying to get to admit something??? I hate posts with crap like that in it... "ADMIT IT"
  • #15
  • Mar 18, 2010 8:17pm Mar 18, 2010 8:17pm
Quoting trade price
all those thing is good for manual trading.
i can give u a simple moving avarage system that will show you it did make money in last 10 years but in real time it will empty ur account
ea is for backtest.not for real time trading.imho

sho
Please, tell me in what that MA consists, I'm very interest in know why fail in real time but not in last 10 years, excuse me but I think that's impossible.

Quoting trade price
if you can code come and see my thread and code them.you will see all the rules and how i trade.
i am sure you wont be able to
I'm reading a lot of systems and testing them, I will try to code your system when I can, thanks for post it.
  • #16
  • Mar 18, 2010 8:30pm Mar 18, 2010 8:30pm
Quoting Trad3r
As another trader has told you to solve that problem you have to use SLs, so trade 10,001 will be only a lose trade, you will not lose everything in just 1 trade if you use money management.
I guess I posed the question wrong. It's not so much blowing your account on trade 10,001 that I was concerned about, its that trade 10,001 through X could be a long string of losses.

The point I was trying to make was that the only reason you have any confidence in your EA is that it has performed consistently over 10 years of data. The problem is that even 10 years of data isn't statistically significant in determining the "true" probability of a system.

You can see what I mean by running large numbers of montecarlo simulations. On a run of 10,000 equity curve simulations you may see a max drawdown of 25%. Running that same simulation again, you may see max drawdown of 35% or 50% or 150%. The larger your number of simulations, the deeper into the long tail event risk you can see and at some point you WILL find that your account CAN blow up.

So I guess the real question is, how can you put money behind a system based purely on probabilities calculated from a finite dataset?

PS- Were not even going to get into the risk of curve fitting data, but it's worth thinking about in your future pursuit of trading success...
  • #17
  • Mar 18, 2010 8:30pm Mar 18, 2010 8:30pm
Quoting Trad3r
Tell me what you can't code, I'm sure I can find a code to write it, maybe a long and complicated code, but there must be.
Our realities are different - you might find mine an interestring read as some of your current statements on trading are mentioned in my thread in one way or another:

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

I'm am not suggesting you are "wrong" (if that is the case you will work out that for yourself) but it pays for your thinking to be challenged when you are still early on in the journey.. happy trading.
  • #18
  • Mar 18, 2010 8:45pm Mar 18, 2010 8:45pm
Quoting Acedup
Can you get codes for trendlines, flags, pennants etc....Im not a programmer

Or maybe Im reading it wrong?
Of course, here you have a little example of an indicator that draws automatically trendlines:

http://codebase.mql4.com/source/6184

There is a code for everything, the key is to find it.
  • #19
  • Mar 18, 2010 9:12pm Mar 18, 2010 9:12pm
Quoting Darkstar
You can see what I mean by running large numbers of montecarlo simulations. On a run of 10,000 equity curve simulations you may see a max drawdown of 25%. Running that same simulation again, you may see max drawdown of 35% or 50% or 150%. The larger your number of simulations, the deeper into the long tail event risk you can see and at some point you WILL find that your account CAN blow up.
You are assuming that forex is a totally randomly market with no patterns, I don't say that you are wrong, I will try to find some pattern that allways repeats, but if there is no pattern... There wouldn't be no difference between trade forex and play into a casino: if you win it's only because you have luck.
  • #20
  • Mar 18, 2010 9:32pm Mar 18, 2010 9:32pm
By luck...

By Skill...

Or by running an EA based on a system with "true" positive expectancy.
0 traders viewing now