has anyone ever backtested a method and then traded it profitably?

[schweiz]

Like curve fitting?

More or less. But it can happen without them realizing they are just curvefitting. I saw many newbies testing all kind of indicators that are freely available on the internet. They just changed the parameters to see what setting would give the best return. Some of then even "optimized" these parameters on a daily or weekly basis. But in reality they have/had no clue at all about the market.

Question for you: Instead, I should start with incorporating/finding the fundamental factors that affect price?

The first thing you should do is think. What moves the markets, how does it move the markets, how can I manage these moves in such a way that it is profitable?
You should first understand the behavior of the market, which is in fact the behavior of the participants. After that you can try to find a solution; you should first understand the real problem. You cannot find a good solution if you understand the problem in a wrong way.

I never followed/bought any trading course, neither did I read books about it (except for Market Wizards as I received these books for free from my broker, but at that time I had already a profitable system). My brain was never spoiled by all the BS you find on internet. It help me also to think in a logical way that was not influenced by others. The real out-of-the-box thinking. It was a journey that lasted many years, so I was apparently too stupid to find in "no time" a good system like many do as it seems.

 

[Steve Alexander]

It seems like everyone on this forum is like "i backtested and then it didn't work" or they're like "i don't backtest."

Just postulating.

I am currently running 5 different strategies on four ETFs. All were thoroughly back tested and all were profitable (averaged about 25% annually). I expect the live trading to be approximately the same, provided I can my mischievous hands off the machine.

 

[MarkBrown]

wtf this thread - someone was just born - do you even toby crabel dude? hundreds of others open your eyes.

 

[Palindrome]

Yes, I have backtested then traded the plan.

But after you backtest, you have to first figure out if you can even EXECUTE. Learning to Execute takes time.

My system backtested does about 125% annually... when I executed it...it yielded 75% annually. So EXECUTION is a whole other world.

 

[qlai]

My system backtested does about 125% annually... when I executed it...it yielded 75% annually. So EXECUTION is a whole other world.

Don't you test with various degrees of sleepage, fill rates, etc? Let's say you run your system for a year, do you go back and re-test it with exact same parameters to see what caused divergence? Is there a name for this - Back to the Future testing or something

 

[schweiz]

So EXECUTION is a whole other world.

Or realistic backtesting...
Maybe the backtest was not correct.

 

[Palindrome]

Don't you test with various degrees of sleepage, fill rates, etc? Let's say you run your system for a year, do you go back and re-test it with exact same parameters to see what caused divergence? Is there a name for this - Back to the Future testing or something

Slippage was a very minor factor.

The divergence in returns was because of ME, mistakes I made, not holding long enough, not entering the right set up etc. I largely executed well though for not being a computer. I'm not sure a computer can be programmed like me though. I guess it can, but I'm about 20% discretionary and 80% systematic.

 

[Palindrome]

Or realistic backtesting...
Maybe the backtest was not correct.

It was just the psychological influences... like "I've made too much money the last three days, I should hold back on the next trade" OR "I've lost too much money this week, I should not take the next entrance" It's just me....but I am improving, that is all that matters.

Also my backtest is correct. I'm conservative with my backtesting, I don't like it to say I can do 15% per month when it's really 11.5%. I underestimate.

 

[southall]

It seems like everyone on this forum is like "i backtested and then it didn't work" or they're like "i don't backtest."

Typically traders backtest systems optimised for the last few years. These are unlikely to hold up well in live trading. The system might still be long term profitable if it is based on sound and robust trading concepts but no where near as profitable as the short term back test suggested.

Back testing is more likely to match live trading results if you use just a few variables in your system and don't over fit. Also the back test should include 1000s of trade samples and test over many different market conditions, e.g. bull, bear, side ways, volatile, quiet.

The system doesn't have to work well under all these market conditions but such a back test will give you an idea of how the system will behave when you get unfavourable conditions which might happen just after you start trading live.

 

[ElectricSavant]

for sure southall...

[quote snippet]when you get unfavourable conditions which might happen just after you start trading live.[end quote snippet]

also, I think a lot of traders do not realize how hard it is to endure "sideways" for 6 months or more.

ES

P.S. I wish to all of you... smoothly sloped upward equity curves with your forward tests in live accounts...hehe...in other words make money with low drawdowns