Cheers Frank. All the best :-) C
It's a long road but got the time
{quote} In my years of experience I believe for a correct evaluation of a system in back testing, one must follow these rules: 1. Always use 3 different years for back testing. 1 year of downtrend, 1 year of uptrend, 1 year of sideways trend. 2. Check how the system copes with all these conditions. 3. Never use a trailing stop for back testing. MT4 does not process real ticks, so the trailing stops are always simulated, most of the time giving artificially positive results. The only trailing stop you can use would be large trailing stop, like 50...
Thanks for reacting! toke some notes because some of those are really solid tips
I've already tested on offline chart but because i'm using the 15 minute chart i could not go further back than a half year, so have to search for a solution for that one!
I never use trailing stops because it not works for me, one little stupid wrong movement and you get stopped out thats not what you want!
To talk further about moneymanagment, i already thought about that one, wanted a good base to build something around so 1% risk is the most i take with one trade!
Frank
{quote} Frank, In my trading place of work, I have a trading pin-board, which may even be similar to the one Copernicus has hanging on the wall of his trading place of workPinned to my board is a quote from Mark Douglas that along my own journey has become what I consider to be one of the most, if not the most important trading quotes of all. In the very first pages of Trading In The Zone, Mark Douglas says, "I assume that you already have your own system, your own edge." Personally, my own debt of gratitude to Mark Douglas lies in that sentence...
Personallly in my case i will read the book completley to say if i miss something, but i always set my profit in PIPS, to make sure that i'm looking at how much pips did i make not how much money, i'm sometimes (i know this from my self) a bit greedy and impatient but i can say enough is enough, so working with real will hopefully with the small account go right.
Talking about psychology, i'm an overthinker so again i was looking for something to avoid sleepless nights, so i only trade the hours between 9-5 and close all trades for sleep
Their is so much to talk about, to keep it short i wrote everything down i knew that was my weak point and from there i started...
Frank
- #31
- Edited 8:48pm Feb 27, 2016 7:38pm | Edited 8:48pm
- | Commercial User | Joined Apr 2013 | 4,366 Posts
I personally think that trading without extensively testing is a really unwise move. For example, as a matter of due diligence it is incumbent on a fund manager to fully understand the intricacies of their system and it's performance across as wide a variety of market conditions as possible, before taking the plunge and risking client money. Obviously as a retail trader however we can ignore this as it is our money, but it is usually at our peril.
It is no excuse to say that a back-test does not mean anything as live trading is different, as this means that the back-testing process has not been thorough enough or the tester has insufficient experience in how to back-test. That's why back-testers use in sample and out of sample data (demo, forward or live) to evaluate performance and compare and contrast results.
To also say that back-testing is no good as markets change all the time has no bearing on the usefulness of back-testing at all, as this is the prime directive of why we back-test in the first place. The back-test is a measure to assess how the strategy performs over varying market conditions. The volatility of returns is assessed through the back-test to give crucial information that is necessary to evaluate your live trading performance. You don't just stop back-testing when you go live trading. You continue to back-test in tandem with your live trading performance.
I tend to think that those who avoid back-testing are going to run into the trap of searching for profitable systems for the rest of their lives and never quite getting there as without a back-test for comparison, their is no way to quantify whether or not your strategy has got any legs.
If you want an equity curve that for the last 30 years has produced this......
.....then you need to backtest to understand that there will be periods where you get this.....
For most retail traders, they would bin a strategy if it did not produce a positive return over one week, let alone say 5 months of negative returns. The reason of course for this variability of returns is that markets are unpredictable and do change. It is only through applying the Law of Large Numbers that you have any chance of seeing through the variability.
Furthermore, it is pretty naive to not have a few different systems up your sleeves recognising that market conditions do change. It is a bit like this analogy. Do you go to war with only a knife, or do you go to war with a knife, a gun, a grenade, a rocket launcher, a mini-gun, a plasma rifle *snigger*? I tend to think you might be better prepared if you choose the latter.
PS Granted the ability to back-test comprehensively in MT4 is fairly limited and I suspect in a cynical way their is a very good reason for that since it is a platform recommended by the brokers, but this should not influence the decision on whether to back-test or not. You at least can start developing the skills required on MT4 and then move onto other more comprehensive platforms like RightEdge, Wealthlab, Tradingblox or even Amibroker/Ninjatrader etc., or better still, develop your own backtester.
It is no excuse to say that a back-test does not mean anything as live trading is different, as this means that the back-testing process has not been thorough enough or the tester has insufficient experience in how to back-test. That's why back-testers use in sample and out of sample data (demo, forward or live) to evaluate performance and compare and contrast results.
To also say that back-testing is no good as markets change all the time has no bearing on the usefulness of back-testing at all, as this is the prime directive of why we back-test in the first place. The back-test is a measure to assess how the strategy performs over varying market conditions. The volatility of returns is assessed through the back-test to give crucial information that is necessary to evaluate your live trading performance. You don't just stop back-testing when you go live trading. You continue to back-test in tandem with your live trading performance.
I tend to think that those who avoid back-testing are going to run into the trap of searching for profitable systems for the rest of their lives and never quite getting there as without a back-test for comparison, their is no way to quantify whether or not your strategy has got any legs.
If you want an equity curve that for the last 30 years has produced this......
.....then you need to backtest to understand that there will be periods where you get this.....
For most retail traders, they would bin a strategy if it did not produce a positive return over one week, let alone say 5 months of negative returns. The reason of course for this variability of returns is that markets are unpredictable and do change. It is only through applying the Law of Large Numbers that you have any chance of seeing through the variability.
Furthermore, it is pretty naive to not have a few different systems up your sleeves recognising that market conditions do change. It is a bit like this analogy. Do you go to war with only a knife, or do you go to war with a knife, a gun, a grenade, a rocket launcher, a mini-gun, a plasma rifle *snigger*? I tend to think you might be better prepared if you choose the latter.
PS Granted the ability to back-test comprehensively in MT4 is fairly limited and I suspect in a cynical way their is a very good reason for that since it is a platform recommended by the brokers, but this should not influence the decision on whether to back-test or not. You at least can start developing the skills required on MT4 and then move onto other more comprehensive platforms like RightEdge, Wealthlab, Tradingblox or even Amibroker/Ninjatrader etc., or better still, develop your own backtester.
- #33
- Feb 27, 2016 9:40pm Feb 27, 2016 9:40pm
- | Commercial User | Joined Apr 2013 | 4,366 Posts
{quote} Seconded. The hostile sentiment towards backtesting has always puzzled me on this forum when you really have nothing to lose by performing an extensive historical simulation. The common excuse is that future market conditions are always changing and that past results do not accurately indicate future performance. These posters then champion the idea of forward-testing instead of backtesting which is actually a logically inconsistent position to hold because the same limitation of historical testing also applies to your period of live forward...
But you could have added Ed's evil demonic image from the other night to add to the gravitas
- #34
- Edited Feb 28, 2016 12:04am Feb 27, 2016 11:46pm | Edited Feb 28, 2016 12:04am
- | Commercial User | Joined Apr 2013 | 4,366 Posts
- #38
- Feb 28, 2016 12:43am Feb 28, 2016 12:43am
- | Commercial User | Joined Apr 2013 | 4,366 Posts
{quote} Thanks. I appreciate your insights and sharing of your knowledge. And yes, you can take a rain check; hopefully the rain has set in for a long time, so no rants, just plain informative and lucid contributions as always! :-)
But you could have added Ed's evil demonic image from the other night to add to the gravitas
Now that we are on the bandwagon, to add to your rant A, the cynic in me says that those who deride the backtest over extended periods are those who find that when they test out their favourite whim of the day and see that it fails probably within 1 or 2 years duration assume that the back-test is wrong as opposed to taking the reality pill and recognizing that their strategy has been curve fitted for current conditions. No wonder that there is so much derision over back-testing in this forum. Reality is a hard pill to swallow. Here is a test. How...
For example, that guy who responded to you assumes that you are prescribing a 10 year test period as a hard and fast rule. Adopting such a rule would no doubt be helpful for most here but thats really not the meaning behind the statement. The preference for a longer test period is based on the understanding that the longer the test period that more significant the conclusions drawn from the result. Inferences made from 10 years of results are more plausibly true than those made from 10 weeks of results.
Systems don't need to be changed. If you end up changing systems frequently you probably should consider changing your research methodology as well.