oildaytrader
Senior member
- Messages
- 2,806
- Likes
- 125
Anybody have tradestation and is willing to back tests an oil breakout strategy and a Dax breakout strategy?Both are daily breakout mechanical strategies.
Rhody Trader
Senior member
- Messages
- 2,620
- Likes
- 266
Excel
Rhody Trader
Senior member
- Messages
- 2,620
- Likes
- 266
I promise you that if you learn to test your EMAs and MACD in Excel you will learn a lot more about those indicators than you will running tests in any software platform. It's not that hard to do. The math involved (and thus the formulas required) is not complicated.
charliechan
Experienced member
- Messages
- 1,008
- Likes
- 119
by nature i think backtesting is looking in the past to find something that is repetitive. some unfound pattern or sequence no one else has found that will 'predict' future moves on a statistically significant basis.
but repetition is no guarantee of future market moves/events.
ideas based upon past behaviour can block new insights.
cycles change.
if we could build such a prediction algorithm based upon past events, trading would be easy. 10 years of experience informs me otherwise.
understanding how to manage probabilities rather than predict is a lot harder to learn, but much more rewarding.
but repetition is no guarantee of future market moves/events.
ideas based upon past behaviour can block new insights.
cycles change.
if we could build such a prediction algorithm based upon past events, trading would be easy. 10 years of experience informs me otherwise.
understanding how to manage probabilities rather than predict is a lot harder to learn, but much more rewarding.
oildaytrader
Senior member
- Messages
- 2,806
- Likes
- 125
People don't try to predict with knowledge gained from backtesting.They are trying to calculate probabilities.
If backtesting is done for the purpose of finding what works consistently, then what worked in the past will always work in the future.
The currency market has remained the same for 40 years.It still works the same.
charliechan
Experienced member
- Messages
- 1,008
- Likes
- 119
how did you reach that conclusion?
oildaytrader
Senior member
- Messages
- 2,806
- Likes
- 125
Logic and study of market participants and how they behave in certain situations.Human psychology applied to trading.
Human beings with emotions were and are still doing the same thing as they have always done.They will keep on doing the same thing.
There may be a change , but only for a short period until things return to norm
virtuos0
Well-known member
- Messages
- 423
- Likes
- 117
I'm starting to feel like a broken record with this, but I might as well say it again. Back-testing means jack ****, because people confuse correlation with causation. The only thing worth backtesting is a concept based on some kind of market anomaly, inefficiency, or pattern that is caused for a definite and knowable reason by the behavour of market participants. If you don't have this, you have nothing. Backtesting MA's and MACD is pointless, you might as well backtest statistics for how many goats got raped in Somalia.
That's the final time I'm ever going to say it..
Until the next time.
charliechan
Experienced member
- Messages
- 1,008
- Likes
- 119
so how would a backtest reliably inform you that the market has reached an emotional extreme?
a backtest will never (in my knowledge) offer market context on which to understand probability. you'd have to consider economic influences of the period of the back test, participants, perhaps try and obtain an idea of participant inventory, if the competition are in profit or underwater, recent news stories, etc.
all of these will affect how people feel about their position (emotion) at the moment, and that sort of information is quite hard to incorporate in a back test, or impossible if doing a straight ohlc/ta analysis with tradestation or the like.
eg the market may be in a range, and your testing breakouts. if there is much uncertainty then breakouts may be more likely to be false. next 'period/n tests' may be in a time of certainty and any break out is more likely to be valid, as the range is purely a result of trimming positions, waiting for the next release, inaction....etc....
jmo.
a backtest will never (in my knowledge) offer market context on which to understand probability. you'd have to consider economic influences of the period of the back test, participants, perhaps try and obtain an idea of participant inventory, if the competition are in profit or underwater, recent news stories, etc.
all of these will affect how people feel about their position (emotion) at the moment, and that sort of information is quite hard to incorporate in a back test, or impossible if doing a straight ohlc/ta analysis with tradestation or the like.
eg the market may be in a range, and your testing breakouts. if there is much uncertainty then breakouts may be more likely to be false. next 'period/n tests' may be in a time of certainty and any break out is more likely to be valid, as the range is purely a result of trimming positions, waiting for the next release, inaction....etc....
jmo.
morgan_edge
Well-known member
- Messages
- 373
- Likes
- 93
Is it worth backtesting a price action strategy? I've been looking into how to go about this after having my best ever months trading results this month(approx 16 trades per day, 55 pips profit per day), but with my total lack of programming knowledge i wouldt know where to start, but are you saying that I can save myself the time?
Most people are very pro backtesting.
charliechan
Experienced member
- Messages
- 1,008
- Likes
- 119
charliechan
Experienced member
- Messages
- 1,008
- Likes
- 119
thats called sitting on the fence son and is not allowed as it stops trolls.
(some would call it being open minded - but thats no fun on a chat board is it!!)
morgan_edge
Well-known member
- Messages
- 373
- Likes
- 93
is it best to trade with no evidence that the chosen trading method works then?
Note;: im not pro backtesting. Ive never done it, as i would not know how to program/backtest my strategies.
In fact, I know so little about backtesting, im not even sure what people mean when they say it doesn't mean anything? If I went over years of looking at old charts and realised that when price makes a certain pattern (lets say) it tends to go up more often than not. I would then expect it to keep doing so...?... What are the chances that as soon as i start using it, it suddenly stops working after 10 years? I dont really get it.
Having said that, ive never backtested anything, and despite not setting the world on fire, I am in the minority of profitable traders after 4 years.
Note;: im not pro backtesting. Ive never done it, as i would not know how to program/backtest my strategies.
In fact, I know so little about backtesting, im not even sure what people mean when they say it doesn't mean anything? If I went over years of looking at old charts and realised that when price makes a certain pattern (lets say) it tends to go up more often than not. I would then expect it to keep doing so...?... What are the chances that as soon as i start using it, it suddenly stops working after 10 years? I dont really get it.
Having said that, ive never backtested anything, and despite not setting the world on fire, I am in the minority of profitable traders after 4 years.
oildaytrader
Senior member
- Messages
- 2,806
- Likes
- 125
It does not always work out that way.What you have said is correct in theory,those backtests can be worthless even if carried out for the reasons stated.The timing of the trends may be lagging or leading the causes of anomaly.
I backtest my strategies in adverse market conditions to see if the system survives.
O D T
