Looking for trading partners

[Razorsharp]

Interesting thread and very encouraging to know that succesful systems are being developed . It is an area that appeals to me and i would like to explore it further. I have a few questions from some of the experienced guys :

I am looking to develop systems using indicators. I am familiar with the idea that any simple system using basic indicators will produce many signals, some false and some correct thus making the system unattractive.

Could these results differ or be improved if trades were taken only in the direction of a strong trend ? If during a strong downtrend only short trades were taken and the opposite for long .

As far as my understanding goes, systems go wrong when they are going sideways or changing direction. If we then waited for a definite trend to develop would the system results then be greatly improved?

I know that this in itself is a great task but IF one had a method they used to identify this, how would this afect the oucome of the results?

I am assuming that if, for example, an EMA crossover is picked during a strong trend the i is likely to be correct, if we can then avoid further trades during ny whipsaw action

Is there any sense in my logic or am i missing something ?

 

[bluewater18]

Could these results differ or be improved if trades were taken only in the direction of a strong trend ? If during a strong downtrend only short trades were taken and the opposite for long .

As far as my understanding goes, systems go wrong when they are going sideways or changing direction. If we then waited for a definite trend to develop would the system results then be greatly improved?



Hi Razor

First you would have to define:

- What is a strong trend?

and

- What is a definite trend?


Bluewater18

 

[bluewater18]

Hi SoP

Thanks for the feedback. I moved from IG to OandA because IG's spreads are criminal. Hence there are no rollovers, only interest payments (which go both ways) and you are also paid interest on your account balance.

In relation to your figures, what are your money management rules? ie how much are you trading per trade? And how much do you think IG will let you take before they close you down?

Bluewater18

 

[Razorsharp]

Thank you bluewater.

I have a method that i use to determine a strong trend. It may limit my entries but it is quite good to colect a substantial move when i am disciplined. There is currently some discretion involved but I am hoping there may be a way of setting up some clear rules that could be programmed.

I think as a start it would be good to have a strategy that may not be a big player but consistent in generating a small profit.

Once the bread and butter issue is resolved can always try to improve results later.

 

[v1per]

Hi,

I am very intrigued on how you actually manually backtest using spreadsheets. As a competent user of spreadsheets, It would be educational to see what format you use for the backtesting.

Please let me know if this is feasible?

 

[meanreversion]

I use ProRealTime to program and backtest and it's very powerful. Some limitations I have encountered are -

1. It struggles to do more than one action per bar. E.g. if I'm using dailies and the system gets stopped out on day x, it can't re-enter later on during the same day x (this is because the only data held for each bar is H, L, O and C - there is nothing to tell the machine what happened DURING the day).

2. I've not been able to use the %CAPITAL feature, so the best I can do is assume a fixed percentage bet everytime. With FX, I use shares, a starting capital of 100,000 and a bet size of 1,000 (1 pct).. so for example if a strategy goes from 100,000 to 150,000 it's easy to see it's made 50 pct.

3. I can't extract data to Excel, this is very annoying because the system won't run two strategies alongside each other.

I've mentioned ProRealTime on trade2win before, but I really believe it's the best system out there for the money, better than MetaStock for example.

 

[bluewater18]

I use ProRealTime to program and backtest and it's very powerful. Some limitations I have encountered are -

1. It struggles to do more than one action per bar. E.g. if I'm using dailies and the system gets stopped out on day x, it can't re-enter later on during the same day x (this is because the only data held for each bar is H, L, O and C - there is nothing to tell the machine what happened DURING the day).

2. I've not been able to use the %CAPITAL feature, so the best I can do is assume a fixed percentage bet everytime. With FX, I use shares, a starting capital of 100,000 and a bet size of 1,000 (1 pct).. so for example if a strategy goes from 100,000 to 150,000 it's easy to see it's made 50 pct.

3. I can't extract data to Excel, this is very annoying because the system won't run two strategies alongside each other.

I've mentioned ProRealTime on trade2win before, but I really believe it's the best system out there for the money, better than MetaStock for example.

Hi Meanreverison

Your points are well made. Prorealtime is good but not perfect for the reasons you state. It will, however, run two (or more) backtests simultaneously on the same data. The backtests will not know each other is running, though, and will produce seperate results.

Sorry if you knew that already.

I spent the whole of December placing results into Excel. It's slow progress but like you say, options are limited.

The main problem I have now is that I cannot TRADE realtime with prorealtime. I will have to write another code to do this, but it seems worth it.

Are you trading with any success based on your prorealtime code?

Bluewater

 

[meanreversion]

I'm trading fairly low frequency, medium term strategies, it's been fine so far. I don't need to deal directly from PRT as a result.

How did you get your data into Excel, was that done manually?

 

[bluewater18]

Yes, manually. My therapist has advised me not to think about those weeks.

Hey the future is totally in automatic systems trading themselves. Who needs to be looking at a screen 12 hours a day or even at a specific time each day?

 

[meanreversion]

Well, simply because you have an automated or pure rules-based approach, it won't stop you from taking a peek at rates during the day. I'm trying to limit myself to 5 peeks a day, in the absence of any of my orders being triggered.

 

[bluewater18]

Well, simply because you have an automated or pure rules-based approach, it won't stop you from taking a peek at rates during the day. I'm trying to limit myself to 5 peeks a day, in the absence of any of my orders being triggered.

Perhaps you should make it a rule to check your chart five times a day.

 

[Mahogany]

Yes, manually. My therapist has advised me not to think about those weeks..........

Hi Blue
This may help you (unless I misunderstood your process) to get your PRT results in to excell and avoid any keying errors.

Start with your results screen in PRT. Save this screen as a jpeg file. Run this jpeg file thru an OCR program (I use ReadIRIS Pro). Then you can copy the OCR result straight into Excel.

Hope that helps.

 

[bluewater18]

Thanks M

That could be good news.

By the way, do you know how I could move my stops directly into excel? They don't appear on the PRT results page unless they're hit.

They're important because they're the only way I can measure risk on any given trade.

Kind regards

Bluewater18

 

[The Artist]

I use ProRealTime to program and backtest and it's very powerful. Some limitations I have encountered are -

1. It struggles to do more than one action per bar. E.g. if I'm using dailies and the system gets stopped out on day x, it can't re-enter later on during the same day x (this is because the only data held for each bar is H, L, O and C - there is nothing to tell the machine what happened DURING the day).

2. I've not been able to use the %CAPITAL feature, so the best I can do is assume a fixed percentage bet everytime. With FX, I use shares, a starting capital of 100,000 and a bet size of 1,000 (1 pct).. so for example if a strategy goes from 100,000 to 150,000 it's easy to see it's made 50 pct.

3. I can't extract data to Excel, this is very annoying because the system won't run two strategies alongside each other.

I've mentioned ProRealTime on trade2win before, but I really believe it's the best system out there for the money, better than MetaStock for example.

Hi all,

I also started writing systems using PRT and have lauded this software on here before so allow me some input which may be of interest to you guys.
I still have an active PRT license actually but I have recently switched to MCFX for automatic execution purposes. In case you are not aware, MCFX handles drawbacks 1 & 2 meanreversion mentioned. Also, Probuilder/Probacktest language comes from Tradestation's EasyLanguage, as is the case of MCFX's language. This means that any PRT created systems can be copied and pasted into MCFX's editor and about 95% will be recognized. Insert a few ; here and there, change a couple commands, and you have it, ready to be executed.
So in theory it's all rosy. Now there is obviously a BUT, which is the fact that MCFX is more expensive than PRT, and you can only execute with a couple brokers. FXCM is one of them and does pay for MCFX once you trade above a certain volume though.
Anyways, I felt that we travel a similar path here so I hope this can be of value to you.

 

[Mahogany]

........ do you know how I could move my stops directly into excel? They don't appear on the PRT results page unless they're hit.
.........
Bluewater18

Hi Blue
No I don't. I had a similar rquirement but I can't work out a way to get that additional info into the PRT results !! .......