Advertisement
cAlgo backtesting
Sep 07, 2016 at 06:37
1,544浏览
13 Replies
Sep 07, 2016 at 06:37
(已编辑 Sep 06, 2016 at 10:29)
Hi,
I am optimizing my strategy on calgo (ctrader) with tick data precision from server and with 10,000 deposit I am getting results like 13 billion profit with 10k deposit.
How reliable is such backtesting, do you think real results could be at least close to it?
BR
Tomas
I am optimizing my strategy on calgo (ctrader) with tick data precision from server and with 10,000 deposit I am getting results like 13 billion profit with 10k deposit.
How reliable is such backtesting, do you think real results could be at least close to it?
BR
Tomas
附件:
Sep 24, 2016 at 03:59
I got no experience on cAlgo, but whats the avg trade duration?
if i were you id test it on a small account and see what happens..
worst case you loose a few hundred, but sure as hell worth the risk..
if i were you id test it on a small account and see what happens..
worst case you loose a few hundred, but sure as hell worth the risk..
if you follow the flock like sheep you always end up stepping in shit!
会员从Nov 12, 2010开始
160帖子
Sep 25, 2016 at 07:34
I think it all depends on the strategy, if this is a scalper, then it will be highly dependent on the spread and execution rate which can jeopardize the real results greatly.
What time span is that? 5 years?
Results look really good, but too good to be true for live testing, the market is always working against you, but then again, with 8000+ trades, it is highly possible to have good returns, but the DD is what I would be leary of. So low that it generally implies a scalper, which I have never seen a scalper that actually makes it without huge DD in real life. I don't trust scalpers, too many factors to make them profitable and reliable
What time span is that? 5 years?
Results look really good, but too good to be true for live testing, the market is always working against you, but then again, with 8000+ trades, it is highly possible to have good returns, but the DD is what I would be leary of. So low that it generally implies a scalper, which I have never seen a scalper that actually makes it without huge DD in real life. I don't trust scalpers, too many factors to make them profitable and reliable
See my profile or message me for my latest EA
Sep 26, 2016 at 12:30
stockea posted:togr posted:
No opinions?
First Why all your accounts are not verified?
First all of my ctrader accounts are verified. Second it is done automatically when you add ctrader account. Third how is that related to backtesting. Fourth can you read :)?
会员从Nov 21, 2011开始
1601帖子
Sep 26, 2016 at 13:42
Backtesting is a crap thing.
CTrader or MT4, I guess it the same crap thing. even with 99% Quality, backetesting means nothing.
I realizaed this by myself when I did perform some testings.
When you realize that you can get very different result from different brokers by using the same EA, you say "WTF"
Then when you backtest the day (current day) while you had your EA on live the same day... and realize that backtesting doesn't provide same trades... Then you understand backtesting is so rubbish.
The real use of backtesting, it to make sure you have well coded your strategy and EA opens/closes when it supposed to do it.
You test that EA behaves well as expected. That's the only point of backtesting.
Some scarmers then used backtesting to show unneatable performance... and then everybody started on the same way.
CTrader or MT4, I guess it the same crap thing. even with 99% Quality, backetesting means nothing.
I realizaed this by myself when I did perform some testings.
When you realize that you can get very different result from different brokers by using the same EA, you say "WTF"
Then when you backtest the day (current day) while you had your EA on live the same day... and realize that backtesting doesn't provide same trades... Then you understand backtesting is so rubbish.
The real use of backtesting, it to make sure you have well coded your strategy and EA opens/closes when it supposed to do it.
You test that EA behaves well as expected. That's the only point of backtesting.
Some scarmers then used backtesting to show unneatable performance... and then everybody started on the same way.
Sep 26, 2016 at 17:36
@CrazyTrader Some EA's are more reliable when it comes to backtesting then others, there are many factors..
If you know what your doing your backtest and live trading can be damn near exact within a few pips max..
If your backtesting isnt matching your live trading you really gotta ask yourself why, as its not the backtest its your algorithm especially if its a 99% tick data..
Main reasons peoples backtest dont match live trading:
- Their tp/sl is not more then 10x the avg spread..
- They curve fit periods without out of sample testing..
- EA has too many variables that allows it too much freedom on backtest makes it more of a curve fit..
- Trailing stops are used as takeprofit without taking into consideration spread widening on volatile markets..
- Pending stops are used, these create alot of slippage..
- Trade duration is too fast for retail trading..
If you know what your doing your backtest and live trading can be damn near exact within a few pips max..
If your backtesting isnt matching your live trading you really gotta ask yourself why, as its not the backtest its your algorithm especially if its a 99% tick data..
Main reasons peoples backtest dont match live trading:
- Their tp/sl is not more then 10x the avg spread..
- They curve fit periods without out of sample testing..
- EA has too many variables that allows it too much freedom on backtest makes it more of a curve fit..
- Trailing stops are used as takeprofit without taking into consideration spread widening on volatile markets..
- Pending stops are used, these create alot of slippage..
- Trade duration is too fast for retail trading..
if you follow the flock like sheep you always end up stepping in shit!
Sep 27, 2016 at 07:36
CrazyTrader posted:
Backtesting is a crap thing.
CTrader or MT4, I guess it the same crap thing. even with 99% Quality, backetesting means nothing.
I realizaed this by myself when I did perform some testings.
When you realize that you can get very different result from different brokers by using the same EA, you say "WTF"
Then when you backtest the day (current day) while you had your EA on live the same day... and realize that backtesting doesn't provide same trades... Then you understand backtesting is so rubbish.
The real use of backtesting, it to make sure you have well coded your strategy and EA opens/closes when it supposed to do it.
You test that EA behaves well as expected. That's the only point of backtesting.
Some scarmers then used backtesting to show unneatable performance... and then everybody started on the same way.
@CrazyTrader
Perhaps if you spend more time backtesting you could have at least one profitable account. You dont. Point.
Sep 27, 2016 at 07:52
sirius1fx posted:
@CrazyTrader Some EA's are more reliable when it comes to backtesting then others, there are many factors..
If you know what your doing your backtest and live trading can be damn near exact within a few pips max..
If your backtesting isnt matching your live trading you really gotta ask yourself why, as its not the backtest its your algorithm especially if its a 99% tick data..
Main reasons peoples backtest dont match live trading:
- Their tp/sl is not more then 10x the avg spread..
- They curve fit periods without out of sample testing..
- EA has too many variables that allows it too much freedom on backtest makes it more of a curve fit..
- Trailing stops are used as takeprofit without taking into consideration spread widening on volatile markets..
- Pending stops are used, these create alot of slippage..
- Trade duration is too fast for retail trading..
@sirius1fx
exactly
BT is first step
forward testing second
cent is next
会员从Nov 12, 2010开始
160帖子
Oct 03, 2016 at 08:16
I would say it depends a lot of the EA of whether or not backtesting will yield results close to what the Real Testing will yield. In the case of Scalpers, I never trust them in backtests for the reasons CrazyTrader mentioned. When you deal with only a few pips being won or lost, the backtest is very faulty because the spread can easily eat the winning trade in those cases.
The Strategy Tester is best for EAs that win 10+ pips but it still is no where near perfect. Vontogr is correct that you then need to forward test and eventually live test it on a real account (at lowest risk settings) to really see if it profitable.
A lot of work, if the backtest fails, then that is an early sign that an EA is no good, but a good backtest is no indication that it will profit in real world trading.
The Strategy Tester is best for EAs that win 10+ pips but it still is no where near perfect. Vontogr is correct that you then need to forward test and eventually live test it on a real account (at lowest risk settings) to really see if it profitable.
A lot of work, if the backtest fails, then that is an early sign that an EA is no good, but a good backtest is no indication that it will profit in real world trading.
See my profile or message me for my latest EA
Oct 04, 2016 at 07:46
bestdarngood posted:
I would say it depends a lot of the EA of whether or not backtesting will yield results close to what the Real Testing will yield. In the case of Scalpers, I never trust them in backtests for the reasons CrazyTrader mentioned. When you deal with only a few pips being won or lost, the backtest is very faulty because the spread can easily eat the winning trade in those cases.
The Strategy Tester is best for EAs that win 10+ pips but it still is no where near perfect. Vontogr is correct that you then need to forward test and eventually live test it on a real account (at lowest risk settings) to really see if it profitable.
A lot of work, if the backtest fails, then that is an early sign that an EA is no good, but a good backtest is no indication that it will profit in real world trading.
Its about ctrader - calgo backtesting. Do you have any experience with it?
*商业用途和垃圾邮件将不被容忍，并可能导致账户终止。
提示：发布图片/YouTube网址会自动嵌入到您的帖子中！
提示：键入@符号，自动完成参与此讨论的用户名。
