use this. but Start_Time=1 is no longer needed....
- #681
- Apr 27, 2010 3:17am Apr 27, 2010 3:17am
- Joined Apr 2009 | Status: Trader | 2,035 Posts
- #682
- Apr 27, 2010 3:34am Apr 27, 2010 3:34am
- Joined Apr 2009 | Status: Trader | 2,035 Posts
Geoff, thanks for the great work, here are a couple of programming ideas:
-- to be able extern bool value 'Show_Settings' work, we should extend CheckForCross() function as follows:...
If you look at the code you will see that there is a Magicnumber setting which contains a "base" number that can be changed manually, the currency pair and time scale are then appended to this base number to generate the actual magic number.
Hope this helps!
It is optimised for the past 12 months but how far back should we go?
I am not sure how valid long term back testing is anyway, the market changes so rapidly!
What we need is a "crystal ball" forward tester, that would solve everything!
im very new to EA so i still trying to find out how to use it long term. Thinking that if it can survive years of data then probably it can last longer.. but is it possible to like optimise it every mth with like a year of data and use to for the next mth to increase the EA validity? Cuz its really a good EA and hope to use it long term
- #685
- Apr 27, 2010 4:05am Apr 27, 2010 4:05am
Show Settings is there to be switched to false when back testing, otherwise it takes forever.
If you look at the code you will see that there is a Magicnumber setting which contains a "base" number that can be changed manually, the currency pair and time scale are then appended to this base number to generate the actual magic number.
Hope this helps!
I think there was many testing going on the last few days on various platforms. All people who read your thread 2 times over (or more) probably are convinced that you are the master of the universe(!!!!) because test trades are firing off with such precision that it is almost astonishing.
With regards to the previous idea to increase the lot size after a s/l I would like to raise my concern. If a s/l is hit in o/n session with 140 pips only because of a 12.00 am GMT news (like a great nationwide or RICS) then ok.
However a combined second and third Greece would let everything tank agst the USD and will cause probably more than just 1 session of volatily. Perhaps we need to think very flexible how we react to this case.
After all, apologies. I am falling short of the promise to provide you with implied vols data to be implemented in your ea but work on it.
Question: Can we have a closer look into the impact between demo and life acct? It seems to be of great importance with regards to the modelling quality. any clue?
- #686
- Apr 27, 2010 4:16am Apr 27, 2010 4:16am
- Joined Apr 2009 | Status: Trader | 2,035 Posts
Hi Geoff,
Can you please remove Start_Time from the next release as an extern variable so that we cannot change it. (hard code it as 23?)
We should only be able to change GMToffset - this will help solve all of the confusion.
Cheers, Andrew.
Thank you for all your help!
I thought I was the only one who was confused!
Yes, I think you are correct on this, it was only there in case people wanted to test the EA on other currency pairs in which case they may need a different start time.
- #687
- Apr 27, 2010 4:30am Apr 27, 2010 4:30am
- Joined Apr 2009 | Status: Trader | 2,035 Posts
Geoff,
I think there was many testing going on the last few days on various platforms. All people who read your thread 2 times over (or more) probably are convinced that you are the master of the universe(!!!!) because test trades are firing off with such precision that it is almost astonishing.
With regards to the previous idea to increase the lot size after a s/l I would like to raise my concern. If a s/l is hit in o/n session with 140 pips only because of a 12.00 am GMT news (like a great nationwide or RICS) then ok.
However a combined second...
I take on board your reservations about an "aggressive" response to a loss, trends we can live with but short rapid ranging can easily catch us on the wrong foot!
The difference between live and demo accounts is not only confined to the modelling quality in back testing unfortunately, it is also true on forward testing.
I have v5a on a live micro account that did the 2 short trades last night but I also have it running on a demo account and this did not take the trades last night.
Both these accounts are with Alpari, so it is easy to see that there will be even bigger differences from one broker to another.
I am confident that our EA is basically profitable, it all stacks up so far, so I will just use it and monitor its performance for as long as this remains the case!
By the way, thank you for all your input, it is appreciated!
- #688
- Apr 27, 2010 4:35am Apr 27, 2010 4:35am
- Joined Apr 2009 | Status: Trader | 2,035 Posts
im very new to EA so i still trying to find out how to use it long term. Thinking that if it can survive years of data then probably it can last longer.. but is it possible to like optimise it every mth with like a year of data and use to for the next mth to increase the EA validity? Cuz its really a good EA and hope to use it long term
Thank you for your interest!
- #690
- Apr 27, 2010 4:49am Apr 27, 2010 4:49am
I agree. A EA that can survive years of data would be a dream. But I don´t know if its possible.
If You start this EA 2000 and trade it until today it blows Your account several times. Maybee its because of bad data? Maybee its impossible with the same setting for 10 years?
Maybee its a good idea to optimize it one time/month?
Best wishes
Elovv
consistency wins
- #691
- Apr 27, 2010 4:51am Apr 27, 2010 4:51am
It is very rewarding to see so much interest!
I take on board your reservations about an "aggressive" response to a loss, trends we can live with but short rapid ranging can easily catch us on the wrong foot!
The difference between live and demo accounts is not only confined to the modelling quality in back testing unfortunately, it is also true on forward testing.
I have v5a on a live micro account that did the 2 short trades last night but I also have it running on a demo account and this did not take the trades last night.
Both...
now you made me concerned.
I am testing V5a on Alpari demo account and everything went fine tonite. Opened 2 trades 1.25 Alpari time and close with 22 and 49 pips tp. Now you tell me that your live micro account took the trades and the same demo version that I use did not take the trades?
That reminds me of my OTC brokers of the past. I think we all need to check our brokers in demo and live and hope that reality does not proof what I suspect.
- #692
- Apr 27, 2010 5:20am Apr 27, 2010 5:20am
- Joined Apr 2009 | Status: Trader | 2,035 Posts
Geoff,
now you made me concerned.
I am testing V5a on Alpari demo account and everything went fine tonite. Opened 2 trades 1.25 Alpari time and close with 22 and 49 pips tp. Now you tell me that your live micro account took the trades and the same demo version that I use did not take the trades?
That reminds me of my OTC brokers of the past. I think we all need to check our brokers in demo and live and hope that reality does not proof what I suspect.
When I asked Alpari what the difference was they told me that their demo accounts are based on ECN feeds whereas their live accounts are not.
I have an EA that on back testing gives several losses in the year, yet on a live account using real money it has not made one loss in the past 12 months, now that I cannot explain!
- #695
- Apr 27, 2010 6:27am Apr 27, 2010 6:27am
- Joined Apr 2009 | Status: Trader | 2,035 Posts
Geoff
Does it trade on IBFX mini live account. 0,01 lot=1 cent
Elovv
The EA adjusts itself to 1 standard lot ($10), 0.1 mini lot ($1) or 0.01 micro lot (10 cents) size.
This will be a useful test, many thanks!
- #697
- Apr 27, 2010 6:43am Apr 27, 2010 6:43am
- Joined Apr 2009 | Status: Trader | 2,035 Posts
On IBFX standardaccount 0,01 lot=10 cent/pip
On IBFX mini account 0,01 lot = 1 cent/pip
Elovv
