Attachments: D1 - 20 EMA & 50 EMA

D1 - 20 EMA & 50 EMA

#141
May 11, 2013 1:07pm May 11, 2013 1:07pm

claudia1
Joined Apr 2013 | Status: Trader | 10,297 Posts

Quoting thehenryfam

Disliked

{quote} Thanks! Will check it out. Scott

Ignored

Scott/Ati with the greatest of respect guys all this back testing is a waste of time. What does it tell you? I traded for a tier 1 bank and now trade by myself for a living and would not dream of back testing. What's the point. Markets are dynamic tomorrow is another day get your money out and trade what you see. Believe me looking into history in markets like this is wasting your time. Good luck.

#142
May 11, 2013 1:35pm May 11, 2013 1:35pm

claudia1
Joined Apr 2013 | Status: Trader | 10,297 Posts

Quoting Ati Baia

Disliked

hi claudia1, Thanks for your comment. Ati in Blue: {quote} So, I don't see any way to get out of this principal - so way not do it as sorely as possible.

Ignored

Hey

Do what you do and I wish you well but the right hand side of the chart is where the money is.

Edit: getting off topic here sorry 100pips

#143
May 11, 2013 1:42pm May 11, 2013 1:42pm

Ati Baia
| Joined Nov 2011 | Status: Trader | 1,042 Posts

hi claudia1,
Thanks for your comment.

Ati in Blue:

claudia1:
Scott/Ati with the greatest of respect guys all this back testing is a waste of time. What does it tell you? I traded for a tier 1 bank and now trade by myself for a living and would not dream of back testing.

I can't say that I traded for a tier 1 bank so you surely have more experience.
Nevertheless it is my opinion that back testing is not a total waste of time.

What's the point. Markets are dynamic tomorrow is another day get your money out and trade what you see.

Well, I do not know many/anyone who does not make their trading decision on what they learned from past data. You say trade what you see:
What do you see? Past candles all the way through: (I admit that some might be more recent than 5 years depending on your time frame - but it's just this principal: we only have past data to base our decision on or flip a coin.

Believe me looking into history in markets like this is wasting your time. Good luck.

Maybe here is the clue: like this, there might be better ways to do it.
But the principal that traders base there decision on Past data still holds - so way not do it as sorely as one knows to do it.

but the right hand side of the chart is where the money is.

Exellent Ati

Money is important, Love is essential.

#144
May 11, 2013 1:49pm May 11, 2013 1:49pm

claudia1
Joined Apr 2013 | Status: Trader | 10,297 Posts

Quoting Ati Baia

Disliked

hi claudia1, Thanks for your comment. Ati in Blue: claudia1: Scott/Ati with the greatest of respect guys all this back testing is a waste of time. What does it tell you? I traded for a tier 1 bank and now trade by myself for a living and would not dream of back testing. I can't say that I traded for a tier 1 bank so you surely have more experience. Nevertheless it is my opinion that back testing is not a total waste of time. What's the point. Markets are dynamic tomorrow is another day get your money out and trade what you see. Well, I do not know...

Ignored

Good post. But remember that in currency markets (spot or futures) a lot is driven by economic data on a daily basis. How can the past account for tomorrows news?

#145
May 11, 2013 3:50pm May 11, 2013 3:50pm

Ati Baia
| Joined Nov 2011 | Status: Trader | 1,042 Posts

Quoting claudia1

Disliked

Good post. But remember that in currency markets (spot or futures) a lot is driven by economic data on a daily basis. How can the past account for tomorrows news?

Ignored

Of course I also think that 5 years back historical rates might not necessarily help with today’s news, but nevertheless if one thinks about it: the past events are quite important.

e.g. How do you decide what to do if tomorrows news is such and such..... Mainly one's decision will be based of what one has learned: how the market reacted to similar news in the past. (may it through personal experience like you who traded for a bank or through some ideas someone else posted - but it is most of the time based on past market reaction: It's the main thing one got even for fundamentals / news.

Anyway, this one I really liked: but the right hand side of the chart is where the money is. it sticks in the brain

CheersBetter not to spoil a nice thread

Money is important, Love is essential.

#146
May 11, 2013 9:29pm May 11, 2013 9:29pm

turtlefan
| Joined Apr 2008 | Status: Slow and steady wins the race... | 545 Posts

Quoting claudia1

Disliked

{quote} Scott/Ati with the greatest of respect guys all this back testing is a waste of time. What does it tell you? I traded for a tier 1 bank and now trade by myself for a living and would not dream of back testing. What's the point. Markets are dynamic tomorrow is another day get your money out and trade what you see. Believe me looking into history in markets like this is wasting your time. Good luck.

Ignored

Well, Claudia, it is not a complete waste of time because it is the only way of quickly discovering whether the method has a long-term "edge" and viability. If it is a consistent loser in the past, it is probably not worth even pursuing on a forward test. However, I would agree that it's usefulness is definitely limited, and outside of the general edge, not much else can be derived.

I will also say that i am especially suspicious of the results generated by the Strategy Tester. All one has to do to have their doubts is look at 100Pips' reports and see the irregularities. As I pointed out earlier, in many of the pairs the average profit and average loss were very close, and in a few cases, the average loss was actually quite a bit larger! Now of course that makes little to no sense given that this method always uses a larger profit target (350:250). So something is drastically wrong there, and thus I am pretty sure those results cannot be trusted. The modeling quality was also much lower than it should be (some were just over 50% - yikes!). I may try to backtest my EA, but I am hoping to use cleaner data and get results that appear more consistent with the method's rules.

If I get energetic this next week, I may embark upon a manual backtest of a couple of the currencies. I think it would be a more reliable method. In any event, the forward test is still going to be our best bet on seeing how successful this thing can be.

Scott

Prices are never too high to begin buying or too low to begin selling.

#147
May 11, 2013 9:47pm May 11, 2013 9:47pm

100PipsADay
| Commercial User | Joined May 2012 | 3,285 Posts

Quoting thehenryfam

Disliked

{quote} Well, Claudia, it is not a complete waste of time because it is the only way of quickly discovering whether the method has a long-term "edge" and viability. If it is a consistent loser in the past, it is probably not worth even pursuing on a forward test. However, I would agree that it's usefulness is definitely limited, and outside of the general edge, not much else can be derived. I will also say that i am especially suspicious of the results generated by the Strategy Tester. All one has to do to have their doubts is look at 100Pips'...

Ignored

Scott,

I agree about the lack of consistency with the back testing overall. Is there any other automated way (maybe a company?) to perform back testing for multiple currency pairs? I would actually like to try 4 different sets of SL & TP parameters for 6 hour delay and 4 different sets of SL & TP parameters for 12 hour delay as I am pretty sure that some parameters will work better for some pairs and will not be as effective on others that have a smaller average daily range.

If anybody knows of a company who can perform quality back testing I would be interested. During the past 24 hours I have back tested the same strategy but used a 275 SL and a 450 TP and have experienced even better results for most pairs. However, I am also going to be trying to other sets of parameters for pairs that are less volatile and have a smaller average range.

**I have also noticed that for some reason my SL's are trailing every once in a while on the back test and am not sure why... This doesn't happen all the time though so not sure what is going on. This is the biggest reason for the smaller average pips per win from what I can tell.

#148
May 11, 2013 9:51pm May 11, 2013 9:51pm

100PipsADay
| Commercial User | Joined May 2012 | 3,285 Posts

I stumbled across this company, anybody have any experiences with them in the past? http://www.onlinestrategytester.com/index.html

#149
May 11, 2013 9:55pm May 11, 2013 9:55pm

Dewey McG
| Joined Jun 2009 | Status: Trader | 285 Posts

Back testing is a waste of time if youthink a strategy that is successful in back testing automaticallywill be successful going forward. However, it does have several uses:

If a strategy performs poorly in backtesting you can be fairly certain it will not be successful for you.

Back testing can help you tweak astrategy. For example you might want to know how much a trailing stopwill affect draw down and overall profitability. It can help youdetermine other parameters as well.

Back testing is a waste of time if youthink a strategy that is successful in back testing automaticallywill be successful going forward. However, it does have several uses:

If a strategy performs poorly in backtesting you can be fairly certain it will not be successful for you.

Back testing can help you tweak astrategy. For example you might want to know how much a trailing stopwill affect draw down and overall profitability. It can help youdetermine other parameters as well.

One of the problems is that most peopledo not know how to back test or optimize. Back testing with poor datadoesn't help. Optimization without a walk forward analysis is justcurve fitting.

It is a guide and nothing more. It cansave you a lot of time telling you whether a forward test is worthpursuing.

One of the problems is that most peopledo not know how to back test or optimize. Back testing with poor datadoesn't help. Optimization without a walk forward analysis is justcurve fitting.

It is a guide and nothing more. It cansave you a lot of time telling you whether a forward test is worthpursuing.

#150
May 11, 2013 10:22pm May 11, 2013 10:22pm

turtlefan
| Joined Apr 2008 | Status: Slow and steady wins the race... | 545 Posts

Quoting 100PipsADay

Disliked

I stumbled across this company, anybody have any experiences with them in the past? http://www.onlinestrategytester.com/index.html

Ignored

No, I have not ever looked for companies that do testing for you. I guess at $15 for a month a testing, you wouldn't be out too much if it wasn't what you expected. Might be worth a try.

Does your EA have code in it somewhere for trailing stops? I can't imagine why your EA would be doing that in tests if it wasn't somewhere in the code. Did you create it yourself or did you have it created for you?

Scott

Prices are never too high to begin buying or too low to begin selling.

#151
May 11, 2013 10:27pm May 11, 2013 10:27pm

100PipsADay
| Commercial User | Joined May 2012 | 3,285 Posts

Quoting thehenryfam

Disliked

{quote} No, I have not ever looked for companies that do testing for you. I guess at $15 for a month a testing, you wouldn't be out too much if it wasn't what you expected. Might be worth a try. Does your EA have code in it somewhere for trailing stops? I can't imagine why your EA would be doing that in tests if it wasn't somewhere in the code. Did you create it yourself or did you have it created for you? Scott

Ignored

I had it coded for me by a friend and it does have trailing stop settings but I don't think they are "activated" for back testing. I know that my forward tests are not trailing at all so I'm not sure.

And like I said, it doesn't do it all the time with the back testing... LOL

As far as I can see though it looks as though this system is solid even if left on auto pilot but with a little manual intervention every so often this could be a very profitable way to trade multiple pairs.

I'll post results for the new SL and TP parameters in a few more hours for multiple pairs for anybody who might be interested and I am seriously considering trying out the strategy testing company to see what they have to offer, it's only $15 anyways.

#152
May 11, 2013 10:46pm May 11, 2013 10:46pm

Dewey McG
| Joined Jun 2009 | Status: Trader | 285 Posts

One of the things I was going to try is writing the same EA in Zorro. It is another platform that is much better than MT4 for back testing optimization and walk forward analyses. Also--it is free! Once I convert the code, it will take no more than a couple hours what would likely take a few weeks to do in Mt4

#153
May 11, 2013 10:47pm May 11, 2013 10:47pm

turtlefan
| Joined Apr 2008 | Status: Slow and steady wins the race... | 545 Posts

Quoting 100PipsADay

Disliked

{quote}As far as I can see though it looks as though this system is solid even if left on auto pilot but with a little manual intervention every so often this could be a very profitable way to trade multiple pairs.

Ignored

Agreed. It seems to have an edge and a good R:R ratio. Add in good money management, and it's likely to be a winning system over the long haul.

Again, thanks for sharing your ideas and your backtest results.

Scott

Prices are never too high to begin buying or too low to begin selling.

#154
May 11, 2013 10:53pm May 11, 2013 10:53pm

100PipsADay
| Commercial User | Joined May 2012 | 3,285 Posts

The following tests were done using the exact same strategy and rules with the exception of Stop Loss and Take Profit parameters. The following results were done using 275 SL and 450 TP parameters. This is a risk:reward ratio of about 1.6. The previous SL and TP parameters were 250 SL and 350 TP which is a 1:1.4 risk:reward ratio.

I'm going to try:

SL 275 TP 450 (risk:reward - 1.63)
SL 250 TP 350 (risk:reward - 1.4)
SL 200 TP 400 (risk: reward - 2)
SL 160 TP 320 (risk: reward - 2)

I'm going to look into which pairs work best for each set of parameters as I'm thinking that some of the less volatile pairs will probably work better with the 160 SL TP 320 parameters.

This is obviously not going to happen over night but I'll do my best in the coming weeks to see which set of parameters works best for each pair. By the way, I figured out what was wrong with the "trailing stops" but unfortunately this was after the fact so I may do these tests over with correct settings as this could have played a major role in regards to average pips per win for some of the tests...

#155
May 11, 2013 10:55pm May 11, 2013 10:55pm

100PipsADay
| Commercial User | Joined May 2012 | 3,285 Posts

USD/JPY (June 6, 2012 - May 9, 2013)

Total Trades: 58

Total Net Pips: 10,124

Profit Factor: 4.14

Win %: 82.76%

Stop Loss: 275

Take Profit: 450