Buying a system vs. coding yourself

[WilkensOne]

Hey folks,

Just wanted to get some opinions on buying a system with the code written compared to writing it yourself. I am by no means good with computer programming and learning amibroker whilst worthwhile I think will take me a long time. Just as an example as we would all know it I think, the chartist BBO code you can purchase for around $650 and are able to modify it yourself for any fine tuning you may want. I am leaning towards something like this at the moment, thoughts?

Wilkens

 

[DocK]

Re: Buying a system over coding yourself

Hey folks,

Just wanted to get some opinions on buying a system with the code written compared to writing it yourself. I am by no means good with computer programming and learning amibroker whilst worthwhile I think will take me a long time. Just as an example as we would all know it I think, the chartist BBO code you can purchase for around $650 and are able to modify it yourself for any fine tuning you may want. I am leaning towards something like this at the moment, thoughts?

Wilkens

That's a question I asked myself a couple of years ago.
1. I have limited time (self-employed with teenagers),
2. I've no programming knowledge or expertise,
3. I am still in the early stages of building my investing knowledge and confidence

I looked at the cost of amibroker and the like, data etc, combined with the hours it would take to learn how to work it, let alone develop and backtest a system. I decided it would be better for me to simply subscribe to the Chartist Growth Portfolio and follow that for a few years while building knowledge. Although the entire nuts and bolts of the system aren't made public (why would they be?) the basic methodology appealed to me and the past performance history is certainly made available for scrutiny and met my criteria. I've been able to replicate it's signals to a degree with very basic scanning on Incredible Charts and have a bit of a fiddle with different screening options to compare my thought processes with the signals the GP provides. I've had some winners, some losers and overall my "dummy system" has not outperformed the one I'm paying for, so clearly I'm better off paying someone with more skill to do my stock picking for me. I still do a little discretionary stock picking on a more fundamental analysis basis, and that's enough to satisfy my need for independence.

I can certainly appreciate though that those with an interest in system design, or of a more "I did it my way" nature, would prefer to do their own coding etc - and more power to them. If I should wake one morning with a stunning idea for an "edge" I may very well join them, but until then I'm happy to outsource.

 

[WilkensOne]

Re: Buying a system over coding yourself

Hey DocK thanks for the reply, it is useful that you have been in a similar situation!

Can I ask what type of results you have experienced from following the growth portfolio?

The Chartist also offer some of the systems from Unholy Grails for purchase that you can import to Amibroker yourself and then essentially follow the trades it produces which appeals to me. It seems this may be similar to the Growth portfolio except it may only follow once particular system, the benefit of this however is that the system is disclosed to you meaning in the future you are able to adapt or improve the system if required. This appeals to me as a somewhat middle ground between subscribing to a service and writing your own. I am tempted to pay for the system from Nick such as BBO and run that system as in the book the simulated CAGR results were ~27% which seems incredible to me.

Thanks,
Wilkens

 

[FlyingFox]

Re: Buying a system over coding yourself

Hey folks,

Just wanted to get some opinions on buying a system with the code written compared to writing it yourself. I am by no means good with computer programming and learning amibroker whilst worthwhile I think will take me a long time. Just as an example as we would all know it I think, the chartist BBO code you can purchase for around $650 and are able to modify it yourself for any fine tuning you may want. I am leaning towards something like this at the moment, thoughts?

Wilkens

If your programming skills are limited my advise would be to not go down that path unless you have the time and energy to up-skill yourself. I have no experience in regards to the trading aspects that you mentioned but as a former software engineer and still programmer, the most difficult part isn't necessarily the coding but making sure your code is doing the right thing and that you will understand it in 2 yrs time.

 

[WilkensOne]

Re: Buying a system over coding yourself

If your programming skills are limited my advise would be to not go down that path unless you have the time and energy to up-skill yourself. I have no experience in regards to the trading aspects that you mentioned but as a former software engineer and still programmer, the most difficult part isn't necessarily the coding but making sure your code is doing the right thing and that you will understand it in 2 yrs time.

Thanks FlyingFox, that is why I am thinking I may be a bit over my head as I don't currently have the time to dedicate to learning that much. Full time university and part time work take up enough time at the moment!

If The Chartist's growth portfolio costs $650p/a and the coding to his BBO code is also $650 (which apparently returns a compound annual growth rate of 27% or so, would it not be better in the longterm to buy the code outright and cut the ongoing costs as all you do it input it into amibroker and run it once a day!?

 

[DocK]

Re: Buying a system over coding yourself

Hey DocK thanks for the reply, it is useful that you have been in a similar situation!

Can I ask what type of results you have experienced from following the growth portfolio?

The Chartist also offer some of the systems from Unholy Grails for purchase that you can import to Amibroker yourself and then essentially follow the trades it produces which appeals to me. It seems this may be similar to the Growth portfolio except it may only follow once particular system, the benefit of this however is that the system is disclosed to you meaning in the future you are able to adapt or improve the system if required. This appeals to me as a somewhat middle ground between subscribing to a service and writing your own. I am tempted to pay for the system from Nick such as BBO and run that system as in the book the simulated CAGR results were ~27% which seems incredible to me.

Thanks,
Wilkens

I started following it in my smsf in late 2011 and am presently up 19.4% on capital employed, inc divs - based on closed and open trades. In my personal portfolio the results are far less attractive as I missed the market run up from early 2009 and started in late 2009, just in time to see the market go sideways or down for quite some time. It's not a system that does well in a non-trending market. I was therefore out of the market for a good period of time after making several small losses, and have been making up ground ever since. In the beginning I also made a few of all the classic beginner cycle mistakes of second guessing the signals, selling too early and missing upside, selling too late and making a larger loss etc etc. The amount of capital employed has varied greatly as well so it's difficult to put a realistic number on it. I'm currently up around 16% since going back into the market in late 2011. The results are not spectacular, but the drawdowns aren't either and capital preservation is more important to me than outperforming the XJO by a huge margin. There are a lot of traders/investors on this forum who will be doing dramatically better, using both their own systems or fundamental judgements, but I'm happy enough with the results I'm getting atm, and confident that if we ever see a sustained uptrend I'll see some good profits. I've a vague memory that the possible drawdown on the BBO system as backtested was greater than the GP, as well as the potential returns? I do take your point though that if you buy the system you only pay once, rather than a yearly subscription, but I take the view that it's less than .5% cut from my return each year and saves me paying for data etc. I'd also be reluctant to start tinkering too much with a proven system without doing a lot of backtesting to be sure the changes were going to give superior results in various markets, and I don't have the skill to do that or the time and inclination to develop it. If that appeals to you though it's definately worth considering. The one absolute thing I've learnt about investing/trading is that there's no one "right way" of going about it! I also strongly agree with FlyingFox that it's imperative that you understand why the system works (or doesn't) and how it makes the money.

 

[Gringotts Bank]

Re: Buying a system over coding yourself

Hey folks,

Just wanted to get some opinions on buying a system with the code written compared to writing it yourself. I am by no means good with computer programming and learning amibroker whilst worthwhile I think will take me a long time. Just as an example as we would all know it I think, the chartist BBO code you can purchase for around $650 and are able to modify it yourself for any fine tuning you may want. I am leaning towards something like this at the moment, thoughts?

Wilkens

Unless, or even if, you're a good coder, Buy, even if it's black box. Make sure it's legit first. Ask for broker statements.

 

[tech/a]

Re: Buying a system over coding yourself

Firstly you need to know what sort of market you expect you'll be trading.

Currently Mean reversion does very well.
Trend following HAS done well but may not continue.
Do you want to trade stock or perhaps futures where
you can trade long AND short.

Personally I learnt more developing systems than by any other method.
I learnt WHAT WORKS and WHY IT WORKS.
I learnt why systems work and why discretionary trading is the most difficult of all.
I learnt and understood Portfolio management/heat and answered many questions that would have been
left un answered---like optimization/Leverage and compounding.
It improved my discretionary trading no end.
I actually understood through development of systems what needed to be
in a profitable PLAN.
Many plans were plain hopeless---I thought they were brilliant--in a system they failed dismally.
Logic seems to have a flaw!.

Its the longest road but in my view the most rewarding!

 

[rnr]

Re: Buying a system over coding yourself

Thanks FlyingFox, that is why I am thinking I may be a bit over my head as I don't currently have the time to dedicate to learning that much. Full time university and part time work take up enough time at the moment!

If The Chartist's growth portfolio costs $650p/a and the coding to his BBO code is also $650 (which apparently returns a compound annual growth rate of 27% or so, would it not be better in the longterm to buy the code outright and cut the ongoing costs as all you do it input it into amibroker and run it once a day!?

Don't forget to factor in the cost of EOD ASX data at approx. $400 pa.

 

[waza1960]

Re: Buying a system over coding yourself

Unless, or even if, you're a good coder, Buy, even if it's black box. Make sure it's legit first. Ask for broker statements

Sorry but this statement is plain nuts.:screwy:
No system will work continually without adjustment to parameters.
Buying a Black Box is silly IMO. What happens when it enters drawdown and you have no idea what it does.
I'll tell you ,you will stop trading it and say it is broken.You won't know why or if it would have recovered .
Much better to buy a System like Radges' so that you know what it does and keep it in tune with the markets.
Because as Tech said you will learn lots by this approach

 

[Gringotts Bank]

Re: Buying a system over coding yourself

Sorry but this statement is plain nuts.:screwy:
No system will work continually without adjustment to parameters.
Buying a Black Box is silly IMO. What happens when it enters drawdown and you have no idea what it does.
I'll tell you ,you will stop trading it and say it is broken.You won't know why or if it would have recovered .
Much better to buy a System like Radges' so that you know what it does and keep it in tune with the markets.
Because as Tech said you will learn lots by this approach

Some systems have gone 8 years pr more without tweaking*. If it breaks the moment you buy it, too bad. Just don't pay too much for it in the first place... maybe 2% of your capital. Nothing wrong with that. If it's broken, buy a new one, a different one. Make sure it's been verified with some reputable 3rd party, that's all.

* if it's been successful over that long, then it will have built in adaptability.

 

[tech/a]

Re: Buying a system over coding yourself

It's ok GB got the " For " side of the debate.
Silly statements are allowed.

 

[WilkensOne]

Re: Buying a system over coding yourself

Firstly you need to know what sort of market you expect you'll be trading.

Currently Mean reversion does very well.
Trend following HAS done well but may not continue.
Do you want to trade stock or perhaps futures where
you can trade long AND short.

Personally I learnt more developing systems than by any other method.
I learnt WHAT WORKS and WHY IT WORKS.
I learnt why systems work and why discretionary trading is the most difficult of all.
I learnt and understood Portfolio management/heat and answered many questions that would have been
left un answered---like optimization/Leverage and compounding.
It improved my discretionary trading no end.
I actually understood through development of systems what needed to be
in a profitable PLAN.
Many plans were plain hopeless---I thought they were brilliant--in a system they failed dismally.
Logic seems to have a flaw!.

Its the longest road but in my view the most rewarding!

Thank you everyone for the replies, despite somewhat differing opinions they are all helpful.

Tech,
So far I have not considered other platforms as I feel the first hurdle should probably be stocks and as I gain more understanding I could look at moving to something else. I also agree with what you are saying that the most beneficial will always be learning yourself and gain the ins and outs. However at the moment I don't have the time to dedicate to learning it from scratch and would like a system I can learn from and about while I do start learning and look at ways I can modify it to see the difference it may make.

Brings me to my next point: radge's turnkey systems are fully disclosed when you purchase them and therefore would allow me to trade a known and tested system whilst learning how to create my own, any advise on the systems he offers or critiques? (Such as BBO or 20% flipper)

Thanks everyone for the input!

 

[merlinnn]

Re: Buying a system over coding yourself

Thank you everyone for the replies, despite somewhat differing opinions they are all helpful.

Tech,
So far I have not considered other platforms as I feel the first hurdle should probably be stocks and as I gain more understanding I could look at moving to something else. I also agree with what you are saying that the most beneficial will always be learning yourself and gain the ins and outs. However at the moment I don't have the time to dedicate to learning it from scratch and would like a system I can learn from and about while I do start learning and look at ways I can modify it to see the difference it may make.

Brings me to my next point: radge's turnkey systems are fully disclosed when you purchase them and therefore would allow me to trade a known and tested system whilst learning how to create my own, any advise on the systems he offers or critiques? (Such as BBO or 20% flipper)

Thanks everyone for the input!

I too am interested to hear any feedback on the turnkey systems, has anyone aside from purchasing actually used a system for 12 months or more?

 

[polpak]

Re: Buying a system over coding yourself

Don't forget to factor in the cost of EOD ASX data at approx. $400 pa.

EOD ASX end of day data is available free of charge from various sites various formats.

eg asxcode, date, open, high, low, close, volume.