Dump it Here

[qldfrog]

@Joe90 in https://www.aussiestockforums.com/threads/dump-it-here.34425/page-168#post-1079495
put me onto http://blueowlpress.com/books/introduction-to-amibroker/ which I've found to be excellent.

P. 8-4 states:



I've not tested this, but it's a good assumption that Mr. Bandy is correct.

Hope this helps...

yes,definitively
just be aware I have had repainting happening while profile was not detecting any future code.
Another way creating differences is synchronization of data, you run explore thinking your data is up to scratch byt DJ for example is not finalised etc...anyway, so the reason I take a snapshot weekly of the explore results and confirm the following week that whatever was recommended in the past is still present exactly as it was then.

 

[qldfrog]

Did we miss the FLN bus?

I missed it too both Dump it and another system.

 

[martyjames]

Can you code into amibroker buying at last close plus 'x'percent, rather than buying on open and see what the difference is to backtested results?

 

[Warr87]

Can you code into amibroker buying at last close plus 'x'percent, rather than buying on open and see what the difference is to backtested results?

off the top of my head, that should be something like this:

BuyPrice = Ref(Close, -1) * 1.05;

Though if you have trade delay set, it would likely be even simpler

BuyPrice = Close * 1.05;

 

[qldfrog]

Can you code into amibroker buying at last close plus 'x'percent, rather than buying on open and see what the difference is to backtested results?

Yes you can but what you do is check in your buy code that price multiplied by percentage is higher than ref(L,+1) for your buy on open
So that your backtest reflect cases that actually miss buys
I did that then do not bother anyone
I prefer lighter simpler code even if backtest will vary a bit from real

 

[qldfrog]

Ideally you compare with open and adjust open or lower at your limit for purchase price
As you see a lot of complexity to be truly accurate but what for?
Will not actually change the strategy

 

[martyjames]

thanks for the replies guys, i'll have a play with it. Im interested to know how much a difference it would make with my trendfollowing system, i suspect not much

cheers

 

[qldfrog]

thanks for the replies guys, i'll have a play with it. Im interested to know how much a difference it would make with my trendfollowing system, i suspect not much

cheers

Please share your results, as i noted above, not really critical and is the aim of the gamevis not to have real equal backtest..once you can track and understand explain difference, my view is who cares.i use backtests as comparison tool, i do not expect to be Elon Musk after 20 y even if my backtest tells me so

 

[Skate]

Skate.

 

[qldfrog]

Did we miss the FLN bus?

jumped on FLN at asking price 0.67 later in the day, I may differ from Mr @Skate who from memory might close order if not passed in the morning, in my case, they remain open for the day

 

[frugal.rock]

FLN clearly in the open list.
The Action Strategetist bought at 0.67 QF.
Hopefully holders do ok. I don't hold.
My portfolio is behaving crabby- going sideways.
Guess what happens if I put my fingers near it? Crabby bite's.
Will chalk FBR up as a win though.

 

[Saqeeb]

@Trav. Thank you for your prompt. It has been a concern for me as well when there is a weekly buy signal after a significant gap up during the previous week. I have been burned previously. I am considering adding a rule where I do not enter a position if it has gapped up significantly during the past week.
However, At this stage of paper trading I am going to follow the system without any questions but will keep an eye on these two positions, FLN and WSP, to see how they perform.

@qldfrog and @Trav. , I will check to see if there is any repainting with the system. Thanks for the ideas guys. However, in case of PPG I do not think it is repainting as it was a trailing stop hit. Nevertheless, I will check to see if there is any repainting concern with my system.

 

[Trav.]

I am considering adding a rule where I do not enter a position if it has gapped up significantly during the past week.

That's the beauty of AmiBroker mate you can add a rule and backtest pretty easily. Your sleep at night factor (SNF) will be different to mine so if you are not sure then I would think that there is less chance of following the system correctly.

It appears that you are not having any issues finding signals so one option could be penalising a stock in your position score ranking if stock gapped up

some code that could be used to determine the size of a gap that you are not comfortable with.

BigGapUp = ( L > 1.02*Ref(H, -1) ); // today's Low 2% above yesterday's high
BigGapDn = ( H < 0.98*Ref(L, -1) ); // today's High 2% below yesterday's Low

Good on you for doing the paper trade mate, as I wish I had your patience but unfortunately I am a sucker to learn the hard way.

 

[Saqeeb]

Good on you for doing the paper trade mate, as I wish I had your patience but unfortunately I am a sucker to learn the hard way.

@Trav. LOL, I am the same as well. But in this instance I decided that I will paper trade first before I dive in deep.

Thank you for the code snippet. I will test this out and hopefully will help my SNF, thanks.

 

[Warr87]

amibroker has a built-in function for gap up/gap down.

http://www.amibroker.com/guide/afl/gapup.html

Something like:

Buy = buycondition1 AND (Gapup == False);

 

[Trav.]

amibroker has a built-in function for gap up/gap down.

Correct mate but that just returns a true or false as you have shown.

The code above allows you to define a % which is more useful in backtesting (optimize) as a small gapup might be ok, and a +10% might be no good.

 

[Linus van Pelt]

Good on you for doing the paper trade mate, as I wish I had your patience but unfortunately I am a sucker to learn the hard way.

Excerpts from http://blueowlpress.com/books/introduction-to-amibroker/:

P. 3-44: Example 9 - Perform Walk Forward Validation

The walk forward process consists of several iterations of:
1. Developing a system over a set of in-sample data.
2. Testing the profitability over a set of out-of-sample data.
3. Stepping forward in time to a new set of in-sample data and a new set of out-of-sample data.
The combined results of all the out-of-sample data are used to decide whether the system recognizes profit potential or not.

• The out-of-sample results from the walk forward run are the best estimate of future performance of the system. They provide the information used in analyzing the reward and risk profiles, and in determining the position size that will produce greatest account growth while holding drawdown to a level of tolerance determined by the trader.

P 7- 4: Perform Walk Forward Runs

While there is nothing special about optimizing, there is something very special about a walk forward run.

P 9-17+: Walk-Forward tab

(Details of Walk-Forward testing)

P 9-32: Walk Forward

While there is nothing special about the optimization process - it is just an organized way to look at a lot of alternatives - there is something very special about the walk forward process. The walk forward process is the best tool available to you for trading system validation. While you are developing your trading system, you have access to all the data. Using your best efforts, you try to develop your system using in-sample data and test it using out-of-sample data. But what will happen when you start actually trading the system? The walk forward process can give you an idea of what might happen.

In the development phase, you used the in-sample data and ran optimizations to search for patterns, then tested on out-of-sample data. The walk forward process automates this for you and provides several important things:
• It automates the process of optimizing over a set of in-sample data.
• It automates selection of the best set of values.
• It automates testing over the out-of-sample data.

Etc, etc, just search on "Walk Forward".

Sooooo, the $64,000 question: Is Amibroker Walk-Forward testing a suitable proxy for Paper Trading?

 

[qldfrog]

Interesting today is that a sea of red, fln is one of my few green
Who would have thought..one of the reason i actually keep gap up...not found rewarding to eliminate, statiscally in my backtests

 

[Warr87]

Interesting today is that a sea of red, fln is one of my few green
Who would have thought..one of the reason i actually keep gap up...not found rewarding to eliminate, statiscally in my backtests

just checked and was definitely a sea of red. i lost half of my open profits today. no matter though, which is early lol.

 

[over9k]

Really frog? You weren't expecting red today after red on wall st last night + all the virus numbers here?