$20000 loss.
There is definitely something wrong, even when I run the code during the bull market, eg 03-07, I am only getting about a 6% p/a return. I cant identify what the problem is..
Simple Trading Systems
- Thread starter ARTHUR.S
- Start date
$20000 loss.
There is definitely something wrong, even when I run the code during the bull market, eg 03-07, I am only getting about a 6% p/a return. I cant identify what the problem is..
tech/a
No Ordinary Duck
- Joined
- 14 October 2004
- Posts
- 20,417
- Reactions
- 6,356
My my you still don't get it.
The average winning trade is over a year so testing it fir a few months is pointless.
You'll all get different results as the system spits out prospects which are different day after day so you'll be testing different portfolio constituents each run--- even with same start dates.
Yes the system was and is leveraged
The data base was and still is the BT MARGIN LIST because they could be leveraged.
10% of capital is position size and initial stop is 10 % of purchase price--- those who think about it a bit will see that that's only a 1 % risk on capital.
Those that also study the system will find that it closed ITSELF out of ALL open positions during the GFC.---you'd have been in cash.
Back to it
WYSY
Just checked your code.
The exit criteria is So far wrong it doesnt matter!
.
Who ever posted it up seems to have been stuffing around with the exit.
The exit is a Cross of the 180 day EMA of the LOW.
No wonder you get crap results!!!!
If your going to dissect it make sure your dissecting the right code.
The average winning trade is over a year so testing it fir a few months is pointless.
You'll all get different results as the system spits out prospects which are different day after day so you'll be testing different portfolio constituents each run--- even with same start dates.
Yes the system was and is leveraged
The data base was and still is the BT MARGIN LIST because they could be leveraged.
10% of capital is position size and initial stop is 10 % of purchase price--- those who think about it a bit will see that that's only a 1 % risk on capital.
Those that also study the system will find that it closed ITSELF out of ALL open positions during the GFC.---you'd have been in cash.
Back to it
WYSY
Just checked your code.
The exit criteria is So far wrong it doesnt matter!
.
Who ever posted it up seems to have been stuffing around with the exit.
The exit is a Cross of the 180 day EMA of the LOW.
No wonder you get crap results!!!!
If your going to dissect it make sure your dissecting the right code.
Yes Tech/A I already knew the exit criteria is wrong (I realised this after I went back and researched what TechTrader is) but that doesn't matter because I'm not trying to test the system, I'm trying to work out why results are differing so much. I already know your system worked very well for the market it was designed for.
This is not the reason tech/a, I've run the backtest enough times to cover for this inherent variation and it never produces results similar to AlterEgo - he has also run the test at least twice and the results are completely different to mine. I'm not sure what the reason is but that randomness is not enough to cause that big a difference.
This is not the reason tech/a, I've run the backtest enough times to cover for this inherent variation and it never produces results similar to AlterEgo - he has also run the test at least twice and the results are completely different to mine. I'm not sure what the reason is but that randomness is not enough to cause that big a difference.
tech/a
No Ordinary Duck
- Joined
- 14 October 2004
- Posts
- 20,417
- Reactions
- 6,356
Try running 50000 portfolios like I have.
When you've done that you'll know what I'm talking about.
Alter the exit signal to start with
Do it properly or don't do it at all.
I don't mind the method being scrutinized but I draw the line at butchered
Rubbish in rubbish out
When you've done that you'll know what I'm talking about.
Alter the exit signal to start with
Do it properly or don't do it at all.
I don't mind the method being scrutinized but I draw the line at butchered
Rubbish in rubbish out
You don't get it Tech/A,I'm not interested in the system, I'm not trying to test it at all . The system would probably lose money in the current market so why would I be testing it's performance right now? I'm interested in the software/data and why there are differences. If you have nothing to say except comments to protect your pride/ego and some historical trading system then you are not helping with this! I already know that the system works in the right conditions, I already said that, no butchering from me mate.
tech/a
No Ordinary Duck
- Joined
- 14 October 2004
- Posts
- 20,417
- Reactions
- 6,356
Get a grip.
I'm telling you why your getting a difference and your
Not even interested.
Check your inputs make sure they are identical then settings then data
Then run Montecarlo analysis to check 1000s portfolios.
Only then will you see correlations
Single runs or even multiple runs can and will show massive differences.
Particularly if the system tested has huge swings in loss to profit.
Knock yourself out
I'm telling you why your getting a difference and your
Not even interested.
Check your inputs make sure they are identical then settings then data
Then run Montecarlo analysis to check 1000s portfolios.
Only then will you see correlations
Single runs or even multiple runs can and will show massive differences.
Particularly if the system tested has huge swings in loss to profit.
Knock yourself out
Ok Thanks for the feedback.
Hi Billyb - I put a post for you on the system parameters link but I'm not sure it is what you are looking for or if it would be helpful.
Elizabeth
