Can I be confident with this strategy?

Outlander said:
Hi all,

I have been struggling with a question, hope someone can help here.

I have a strategy (let's say strategy A) that worked very well on first half year of 2016, yield a sharpe ratio of about 4.0+. It's an intra-day strategy trading futures. But after Aug, 2016, it suddenly stopped working, I lost half of my profit by the end of 2016.

I am working on another similar strategy (B), it works well through out the data of whole year 2016 (I guess I am not over-fitting), but it's not working for data of 2015, big difference.

The question is, should I be confident enough to continue to trade A or trade B with real money in 2017?

Thanks.
More...


u seem to be getting good answers here, one thing i can add is i go through the same problem however the approach i use is different on how i solve it,

i use what i call "bunch-ing"
first of all before i explain keep in mind few things about my strategies which might make my stuff not applicable to yours.

A.the losers are much smaller than the winners,
B.my trading style is swing 1 week to 2 months a trade
C. My backtesting requires that i take every signal either short or long
D. my position size is always the same in USD notional value whether commodities, stocks or futures.

i used to trade strictly forex and specific pairs,
recently i wanted to expand as i found better back testing results in commodities and some stocks,

now iam faced with same issue is how do i know if the back testing results will work the same in the future and how iam confident with it and what if it does work now but stops later? when do i dump it and how would i know if its short term draw down versus it stopped working?

or what if backtesting on one thing showed its losing but then starts working?

for me to hedge against that i add instruments in "bunches" i do that to hedge against possibly picking the worst instrument of what i back tested or just in case the best instrument becomes the worst as soon as i start trading it, (something that happens often to me)

so the answer if u can is to add more instruments all at once and diversify so that its like a basket, the more ur able to add the less chances what u picked is a loser or becomes a loser

for example, i recently back tested my system on crude, silver, palladium, Macys, nordstroms, Apple, and cocoa, all proved profitable, but that doesnt mean all WILL BE profitable? how do i know which one will and wont be??? i dont know so far so i simply add all of them to the list of instruments being traded to hedge against picking the worst one or what becomes the worst
 
I'll try to help you, not with theory but with some graphical examples.

Here's the out of sample profits of one of my trading strategies that's about to go live.

upload_2017-2-6_10-30-25.png


It spans from 2009 to 2016 and it looks great!
But if you isolate some years, months or weeks you can see that it had some very difficult periods.
Here's 2014:

upload_2017-2-6_10-36-20.png


The strategy had a drawdown of 40 trades. If in this situation you stop trading, you will lose the next 100 trades that will give you great rewards.

A great advice I listened once is that you have to memorize the drawdowns of your backtests. Memorize the annual drawdowns, the losing months, the horrible weeks. In that way you won't lose your confidence when you are struggling with the bad trades.

So, in your case, I would continue trading the strategy A if it is well tested. And if the strategy B is also good, start with few contracts at the beginning. It is totally good to add as many strategies as you can. It diversifies your portfolio of strategies as long as they aren't correlated.

If there is anything else I can help you, just let me know.

Cheers!
 
systematictrader said:
now iam faced with same issue is how do i know if the back testing results will work the same in the future and how iam confident with it and what if it does work now but stops later? when do i dump it and how would i know if its short term draw down versus it stopped working?
More...

Hello systematictrader,

Very good post and this is a good conversation/discussion.

I am confused bit when it comes to automated strategies and the back testing.

Forgive my ignorance or lack knowledge regrading automated systems.

Question:

If a trader back tested (automated or manually back tested) an automated trading systems for say +8 years and the results were profitable, why does a trader need to Stop the system when the system is producing losses?

If I spend hours/days/weeks/months/years developing trading systems and back testing a trading system and the results are profitable, I do not understand why I should never need to click Stop or manually stop trading the system.

Why spend all that time developing something automated, to stop it trading it when it starts losing a certain amount of money? For example, I am invested in Vanguard Total Stock Market Index, I will never take my money out of that index because I am losing money. However, I do understand the need to be diverse in developing automated system for different market, say one for ES and one for CL.

Someone please explain to me.

Thanks
 
SimpleMeLike said:
Hello systematictrader,

Very good post and this is a good conversation/discussion.

I am confused bit when it comes to automated strategies and the back testing.

Forgive my ignorance or lack knowledge regrading automated systems.

Question:

If a trader back tested (automated or manually back tested) an automated trading systems for say +8 years and the results were profitable, why does a trader need to Stop the system when the system is producing losses?

If I spend hours/days/weeks/months/years developing trading systems and back testing a trading system and the results are profitable, I do not understand why I should never need to click Stop or manually stop trading the system.

Why spend all that time developing something automated, to stop it trading it when it starts losing a certain amount of money? For example, I am invested in Vanguard Total Stock Market Index, I will never take my money out of that index because I am losing money. However, I do understand the need to be diverse in developing automated system for different market, say one for ES and one for CL.

Someone please explain to me.

Thanks
More...

This is a great question. If I backtested a strategy and it is profitable, why should I stop trading it?

Here's the simpler answer. Because markets change, so eventually ALL profitable strategies will stop being profitable.

So, how could I know if my strategy doesn't work anymore? There's a lot of complex ways to determine that. But the simpler I know is this one. Take the worst drawdown of your strategy and multiplicate it by 2. If your live trading exceeds that drawdown, then stop trading that strategy. That means of course that your money management has to be prepared to deal with that kind of drawdown. And, why do you have to multiplicate the drawdown by 2 instead of stopping at the moment that the live trading reaches the worst drawdown of the backtest? Because live trading yields worse profits than a backtest due to the natural decay of all strategies.

If there is anything else I can help, let me know.
 
pauljherrera said:
This is a great question. If I backtested a strategy and it is profitable, why should I stop trading it?

Here's the simpler answer. Because markets change, so eventually ALL profitable strategies will stop being profitable.
More...

Good post pauljherrera. Thanks for reply.

So why spend time developing an automated trading system that was back testing for many years, if eventually it will stop being profitable over the long run?

If I back test a system for 10 years, should it not work the next 10 years? Maybe its trader full time job to develop many system.

I am not understanding the purpose of developing something that will eventually stop working. Maybe because the profit returns the system runs will beat the stock market index year after year.

Thanks,
 
Well, that's one of the hard realities of trading. Nothing last forever. That's why all profitable traders, hedge funds, etc, are constantly developing new systems.

Nevertheless, there are statistical studies that say that some trading strategies last longer than others. For example, price patterns based strategies can last for decades. They are very hard to code, and that's why this kind of strategies suit the manual trader very well. Instead, quantitative strategies have a shorter life but are simpler to code. And that's why the algorithmic trader has to develope tons of these strategies, because they may die in a shorter time.
 
pauljherrera said:
Well, that's one of the hard realities of trading. Nothing last forever. That's why all profitable traders, hedge funds, etc, are constantly developing new systems.

Nevertheless, there are statistical studies that say that some trading strategies last longer than others. For example, price patterns based strategies can last for decades. They are very hard to code, and that's why this kind of strategies suit the manual trader very well. Instead, quantitative strategies have a shorter life but are simpler to code. And that's why the algorithmic trader has to develope tons of these strategies, because they may die in a shorter time.
More...
Thanks pauljherrera,

ok, its making some sense to me now.

What developing automated systems to trade intraday resistance and supports?

I like trading supports and resistance, but i do this manually now. Is it difficult/challenging to code, backtest, and overall automate trading supports and resistance going forward.

Sometimes I can not be next to the computer during market hours.

Thanks
 
Last edited:
systematictrader said:
u seem to be getting good answers here, one thing i can add is i go through the same problem however the approach i use is different on how i solve it,

i use what i call "bunch-ing"
first of all before i explain keep in mind few things about my strategies which might make my stuff not applicable to yours.

A.the losers are much smaller than the winners,
B.my trading style is swing 1 week to 2 months a trade
C. My backtesting requires that i take every signal either short or long
D. my position size is always the same in USD notional value whether commodities, stocks or futures.

i used to trade strictly forex and specific pairs,
recently i wanted to expand as i found better back testing results in commodities and some stocks,

now iam faced with same issue is how do i know if the back testing results will work the same in the future and how iam confident with it and what if it does work now but stops later? when do i dump it and how would i know if its short term draw down versus it stopped working?

or what if backtesting on one thing showed its losing but then starts working?

for me to hedge against that i add instruments in "bunches" i do that to hedge against possibly picking the worst instrument of what i back tested or just in case the best instrument becomes the worst as soon as i start trading it, (something that happens often to me)

so the answer if u can is to add more instruments all at once and diversify so that its like a basket, the more ur able to add the less chances what u picked is a loser or becomes a loser

for example, i recently back tested my system on crude, silver, palladium, Macys, nordstroms, Apple, and cocoa, all proved profitable, but that doesnt mean all WILL BE profitable? how do i know which one will and wont be??? i dont know so far so i simply add all of them to the list of instruments being traded to hedge against picking the worst one or what becomes the worst
More...
I see, the keyword here is portfolio.
 
You must log in or register to reply here.
Back
Top