Forward Testing Discussion Thread | Page 10 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

MT forward testing and going live 5 replies

forward testing flippy 42 replies

Question on forward testing 5 replies

NEW " SET & FORGET" Forward Testing for VP Day Trading System 16 replies

Forward Testing Results Only 33 replies

Attachments: Forward Testing Discussion Thread
Exit Attachments

Forward Testing Discussion Thread

  • #181
  • Feb 28, 2006 3:54pm Feb 28, 2006 3:54pm
Quoting fijitrader
Of course this makes some serious differences between system performance. Did anyone read the comparison spreadsheets and follow up posts from a couple weeks ago where I said we need to find the provider that most closely resembles the forex markets we trade. Does anyone recall my suggestion to contact their data providers for explanations?

I see a couple of possible reasons for the differences in data:

1. different time of day when data is collected.
2. different source.
3. both.

I know that such price disparities do not exist between the retail platforms I have compared. At least not that consistently and not such large differences. Otherwise the arbitrage opportunities between providers would be much more lucrative. In the data comparison spreadsheets there are differences that extend greater than 80 pips, several in the 50 pip range and way too many in the 20 to 35 pip range. I could make a decent amount of money in arbitrage just trading the end of day differential these figures seem to show. The reality is that such differences do not exist as frequently or with the magnitude we see between data provides.

I contacted VP for assistance and the guy I spoke to told me he'd follow it up. So far I've not heard back from him.

FT
When FT originally posted the suggestion that we contact our data providers, I did just that, and was handed what I will refer to as the 'standard answer' in this situation. I was told that each data provider collects information from a variety of sources and therefore there WILL be some difference in the final results from each. Sound familiar? Having said that, we can see how this difference affects the trading strategies we are testing. Personally I would like to know what VP has to say about this, and FT has indicated he is waiting for an answer. In the meantime, it appears that the data provider used will have a definiite impact on how we decide to structure our 'set & forget' trading system. At this time, I believe we need to use a provider that supplies information that allows us to trade profitably with the chosen platform (Interbank FX).

This week I am demo trading Strat #4 with CSI data as well as the other strats. It will be interesting to see at the end of the week how my trades compare with drjoe4x.
  • #182
  • Feb 28, 2006 4:21pm Feb 28, 2006 4:21pm
I think I have a solution. When I spoke with the guy from VP who is supposed to be looking into this but has not come back to me yet... I asked him if there was a one to one relationship between the previous days high and the predicted high etc. He said there is. Therefore I believe we can use any of the three providers, take the prediction that VP provides and adjust our predictions by the difference between the data we trade with our IB and the data we get from VP.

I'm going to follow up with our guy at VP and first get his comments before we act on this to confirm this strategy.

FT
  • #183
  • Feb 28, 2006 4:23pm Feb 28, 2006 4:23pm
Quoting BeachBum
I did just that, and was handed what I will refer to as the 'standard answer' in this situation. I was told that each data provider collects information from a variety of sources and therefore there WILL be some difference in the final results from each. Sound familiar?
Yes but did you or anyone else ask the exact time they collect their numbers. If the times are different then the numbers are going to be different. If we find out the precise time and it is different then it may assist in choosing the best data provider.

FT
  • #184
  • Feb 28, 2006 4:51pm Feb 28, 2006 4:51pm
Quoting fijitrader
Yes but did you or anyone else ask the exact time they collect their numbers. If the times are different then the numbers are going to be different. If we find out the precise time and it is different then it may assist in choosing the best data provider.

FT
OK............I just got off the phone with CSI tech department and was told that Forex data is released by CSI at 4:30 pm est, which means they finish collecting this data before that time. I asked about this and the person I spoke with could not get me an answer as to the exact time they stop collecting forex data, compile and prepare for release, but I would assume it would not be earlier than 4 pm est, giving them a half hour window to accomplish the process. FT ........would you agree?
  • #185
  • Feb 28, 2006 4:57pm Feb 28, 2006 4:57pm
I just received an email from Rocket telling me he cannot continue with the forward test at least for this week because of health issues. I wish him a speedy and favorable resolution of the problem!

Strat #1 has been a loser form the beginning.........I don't have the time to demo trade it because I am already doing three other strats plus my own real daily trades. I think we can carry on without it.
  • #186
  • Feb 28, 2006 5:00pm Feb 28, 2006 5:00pm
Quoting BeachBum
OK............I just got off the phone with CSI tech department and was told that Forex data is released by CSI at 4:30 pm est, which means they finish collecting this data before that time. I asked about this and the person I spoke with could not get me an answer as to the exact time they stop collecting forex data, compile and prepare for release, but I would assume it would not be earlier than 4 pm est, giving them a half hour window to accomplish the process. FT ........would you agree?
Thanks for making that call.

I think we need to find out exactly what time they use for the close of the day. How long they spend collecting data or when they start collecting or finish collecting means little or nothing. The time they use to determine the close is what will make the biggest difference between prices. The price changes from second to second and certainly from minute to minute. It is essential that we find out what the exact time of day is that they use for the numbers. Perhaps another call to find out who we need to speak to in order to get that information is in order.

FT
  • #187
  • Feb 28, 2006 5:06pm Feb 28, 2006 5:06pm
Quoting fijitrader
Thanks for making that call.

I think we need to find out exactly what time they use for the close of the day. How long they spend collecting data or when they start collecting or finish collecting means little or nothing. The time they use to determine the close is what will make the biggest difference between prices. The price changes from second to second and certainly from minute to minute. It is essential that we find out what the exact time of day is that they use for the numbers. Perhaps another call to find out who we need to speak to in order to get that information is in order.

FT
I will try again tomorrow during regular business hours.
  • #188
  • Feb 28, 2006 5:31pm Feb 28, 2006 5:31pm
Quoting fijitrader
Yes but did you or anyone else ask the exact time they collect their numbers. If the times are different then the numbers are going to be different. If we find out the precise time and it is different then it may assist in choosing the best data provider.

FT
FT/BB,

Sorry, I must have overlooked that in the previous posts...I will be interested to see the results this week using the two separate providers. I had the thought Fiji stated about adding to the PH/PL according to the difference our data provider had compared to the MT4 platform data as long as data collection time is the same, just forgot about it til he mentioned it earlier.

Hope all is well with everyone and I'm sure we will continue to see improved results with the amount of effort everyone is putting forward. It is all appreciated.

drjoe
  • #189
  • Mar 1, 2006 12:26am Mar 1, 2006 12:26am
Quoting BeachBum
I just received an email from Rocket telling me he cannot continue with the forward test at least for this week because of health issues. I wish him a speedy and favorable resolution of the problem!
Rocket thanks for your help! Hope your health issue is just temporary. No need to have anything permanent.

Get well,

FT
  • #190
  • Mar 1, 2006 2:43am Mar 1, 2006 2:43am
Quoting BeachBum
OK............I just got off the phone with CSI tech department and was told that Forex data is released by CSI at 4:30 pm est, which means they finish collecting this data before that time. I asked about this and the person I spoke with could not get me an answer as to the exact time they stop collecting forex data, compile and prepare for release, but I would assume it would not be earlier than 4 pm est, giving them a half hour window to accomplish the process. FT ........would you agree?
Hi Guys,

Just to let you know, I also contacted CSI to establish when they stopped collecting data each day and was told that the cut off point was 16:30 New York time, and the data starts to be collected at 17:00 New York time each day. Then they obviously release it at 18:30 once they have done whatever is necessary to compile the data for release. They didn't actually mention what time they began recording data on Sunday night or finished on Friday evening which would also need to be established.

By the way, did anyone else not recieve an update for the GBP/USD on the night 27/02/06? I have also been looking at the dates on alot of the data when it is exported to excel and have discovered that there is indeed alot of days where certain markets just don't get updated. Therefore if you are trying to establish correlation etc it is quite difficult because the dates get all mix up the further down you go, bringing the correlation into question.

I would just like to add to the debate about the forward testing that is going on at the moment. As you will have seen, my set & forget system was based around trading only when the PSTT and PMTT were both positive or negative. This is a good indicator and perhaps if you are looking to more methods of fine tuning the system I would suggest you try trading only on days that meet those parameters. In addition, I have also added the predicted neural index as an extra indicator with varied to negative success. I would conclude that the predicted neural index is not really a good enough indication to use on its own however, along with the PSTT & PMTT it does enhance it a little (although I remain sceptical about this indicator).

Anyway, great results for the forward testing especially strat 4, keep up the good work. My forward testing of my PSTT, PMTT & PNI strategy trading 7 pairs is currently down by 1 point for the two weeks, not really exciting but its more a long term thing and hopefully over the long term it will continue to build.

Good Trading everyone,

FoggyTrader
  • #191
  • Mar 1, 2006 7:06am Mar 1, 2006 7:06am
Quoting foggyTrader
Hi Guys,

Just to let you know, I also contacted CSI to establish when they stopped collecting data each day and was told that the cut off point was 16:30 New York time, and the data starts to be collected at 17:00 New York time each day. Then they obviously release it at 18:30 once they have done whatever is necessary to compile the data for release. They didn't actually mention what time they began recording data on Sunday night or finished on Friday evening which would also need to be established.

By the way, did anyone else not recieve an update for the GBP/USD on the night 27/02/06? I have also been looking at the dates on alot of the data when it is exported to excel and have discovered that there is indeed alot of days where certain markets just don't get updated. Therefore if you are trying to establish correlation etc it is quite difficult because the dates get all mix up the further down you go, bringing the correlation into question.

Good Trading everyone,

FoggyTrader
Hi Foggy............

I have observed the problems with the GBP/USD pair as well and when I asked about this was not given a real definitive answer. I think it is interesting to note that this one pair has had the worst performance during our forward test period. Could this be because of faulty data? In fact, looking back at Strat #5, the performance of this strat would have been significantly better if the GBP/USD had been left out.

I should have more info from CSI later today.
  • #192
  • Mar 1, 2006 10:19am Mar 1, 2006 10:19am
Spending some time on the phone this am with CSI got me this info from the "data Base Department":

Forex data collection begins at 5pm est

Collection ends the following day at 4:30pm est......this is their EOD, their CLOSE

Data is released at 4:50pm est

this agrees with what Foggy wrote earlier.
  • #193
  • Mar 1, 2006 3:14pm Mar 1, 2006 3:14pm
Quoting BeachBum
Spending some time on the phone this am with CSI got me this info from the "data Base Department":

Forex data collection begins at 5pm est

Collection ends the following day at 4:30pm est......this is their EOD, their CLOSE

Data is released at 4:50pm est

this agrees with what Foggy wrote earlier.
Thanks Beachbum,

I have had more problems with CSI today, my download kept coming up with an error so I couldn't even trade today as I got no data. Several emails later and the problem is still not fixed. Think I might have to ditch CSI...

Very unhappy trader

Goodluck and happy trading,

FoggyTrader
  • #194
  • Mar 1, 2006 3:43pm Mar 1, 2006 3:43pm
Quoting foggyTrader
Thanks Beachbum,

I have had more problems with CSI today, my download kept coming up with an error so I couldn't even trade today as I got no data. Several emails later and the problem is still not fixed. Think I might have to ditch CSI...

Very unhappy trader

Goodluck and happy trading,

FoggyTrader
Interesting...............after my test this week with strat #4 and CSI, I will make a decision about switching providers. I have a CD offering two weeks of free data from Genesis, and may opt to try this in the near future.
  • #195
  • Mar 1, 2006 6:33pm Mar 1, 2006 6:33pm
Quoting BeachBum
Interesting...............after my test this week with strat #4 and CSI, I will make a decision about switching providers. I have a CD offering two weeks of free data from Genesis, and may opt to try this in the near future.
With as much difficulty as I had with Genesis the problems were mostly with support issues. Once I finally figured out, on my own, how to get it running I had no further data problems with Genesis. Looking at the data issues only it appears so far that perhaps Genesis is the better provider in that their data seems to more closely resemble what we are getting from most retail market makers. Any thoughts?

FT
  • #196
  • Mar 1, 2006 10:31pm Mar 1, 2006 10:31pm
Quoting fijitrader
With as much difficulty as I had with Genesis the problems were mostly with support issues. Once I finally figured out, on my own, how to get it running I had no further data problems with Genesis. Looking at the data issues only it appears so far that perhaps Genesis is the better provider in that their data seems to more closely resemble what we are getting from most retail market makers. Any thoughts?

FT
I'd like to know about this as well..............

drjoe4x...........we'd like your input
  • #197
  • Mar 1, 2006 11:07pm Mar 1, 2006 11:07pm
Quoting BeachBum
I'd like to know about this as well..............

drjoe4x...........we'd like your input
I have not had any problems with Genesis in the slightest. I initially called them for a 14 day trial and they said the trial was actually 30 days...was up and running within minutes of requesting the trial. When I upgraded to paid service, again, gave them my credit card information, opened Trade Navigator minutes later and everything has went very smoothly since. You simply click on Trade Nav icon, a window appears when it opens to perform update, it updates the data quickly and automatically sends the info to VP software. No hitches whatsoever at any time.

Additionally, I tried CSI, and the Genesis platform seems far superior in organization and ease of use.

As far as being the most accurate data provider, I'm not sure, were we able to establish that with the historical data comparison of MT4 to the providers?

Anyways, very pleased with Genesis and I would personally recommend the service for use with VP based on my experience.

drjoe
  • #198
  • Mar 2, 2006 12:23am Mar 2, 2006 12:23am
Quoting drjoe4x
As far as being the most accurate data provider, I'm not sure, were we able to establish that with the historical data comparison of MT4 to the providers?
drjoe
Does MT4 allow EOD data to be exported in ASCII?

FT
  • #199
  • Mar 2, 2006 7:48pm Mar 2, 2006 7:48pm
Quoting fijitrader
Does MT4 allow EOD data to be exported in ASCII?

FT
It can be exported as ASCII (*.csv) or ASCII (*.pm)

hu9iiyuiyui8 --(That's from my oldest boy, Jonah, he's 4, sorry, but he just had to type something)

drjoe
  • #200
  • Mar 2, 2006 11:28pm Mar 2, 2006 11:28pm
Quoting drjoe4x
It can be exported as ASCII (*.csv) or ASCII (*.pm)

hu9iiyuiyui8 --(That's from my oldest boy, Jonah, he's 4, sorry, but he just had to type something)

drjoe
Why don't you do a dump from January 1 to Dec 31st of 2005 and post it in both ascii formats. Then dump the actuals from the three data providers for the same period. If we have a volunteer from each provider I'll happily compile another spreadsheet. Also when dumping the data from MT4 please state what mm you are using. Most will be ibfx but let's be sure.

Thanks,

FT
2 traders viewing now