Systems On NZX

[Huskeez]

Any one on these NZX forums test trading strategies?

I have put mine through 20 years of DJIA data and after some tweaking got these results back, but still have lots of questions.


Cheers

[winner69]

Impressive

What's the system?

[bull....]

test stuff all the time just remeber its slightly different on paper to real money on the line

[CJ]

Any one on these NZX forums test trading strategies?

I have put mine through 20 years of DJIA data and after some tweaking got these results back, but still have lots of questions.


CheersWere they any big down years where you would have doubted the system.

Have you gone 'live' with your system.

[Huskeez]

Impressive

What's the system?

Hey winner,

Just a TA based system using 2 indicators along with a trailing stoploss, Just feeling around in the dark till I come up with some high probability situations

[Huskeez]

test stuff all the time just remeber its slightly different on paper to real money on the line

Yeah that's the whole point for my coming up with a system, which I hope to automate. Take the emotion out of trading.

[Huskeez]

CJ,


Here is the equity curve , there were a few big draw downs, but all in all I think rather successful.

[CJ]

CJ,


Here is the equity curve , there were a few big draw downs, but all in all I think rather successful.Sharp fall in 2007 but then everyone suffered that

[peat]

Were they any big down years where you would have doubted the system.

Have you gone 'live' with your system.
There was a 45% drawdown at one point !

[Brut]

Very Impressive,

What 2 indicators did you use for your TA (or is that a secret?)

[blackcap]

Does your system include brokerage fees?

[bmrm]

CJ,


Here is the equity curve , there were a few big draw downs, but all in all I think rather successful.

Very cool, and I'm a big fan of this sort of discussion.

Constructive criticism (and I may be dead wrong this is just based on the two images you've posted): The post 2008 return appears a bit too impressive vs the rest of the series (esp. considering the recovery), I would be a little worried about 'over-fitting' to that particular period.

[Huskeez]

There was a 45% drawdown at one point !

As with all systems .. human intervention is required.

This system was a just a scan with a "Long Biased" component.
In the areas of a GFC or a major stock market correction. It is highly probable this system would not be allowed to run , and in most cases the "Short Biased" system would take its place.

[Huskeez]

Does your system include brokerage fees?



Yes in the first image I posted "Total commissions" is in the bottom right corner. $15 dollar brokerage either way (BellDirect) as I live in Australia now. I didn't allow for the .1% increases in brokerage as capital gets bigger due to the complexity , (but I don't believe it would take away from the end result all the much)

[Huskeez]

Very cool, and I'm a big fan of this sort of discussion.

Constructive criticism (and I may be dead wrong this is just based on the two images you've posted): The post 2008 return appears a bit too impressive vs the rest of the series (esp. considering the recovery), I would be a little worried about 'over-fitting' to that particular period.


Yes I see what you mean , Sharetrader wont allow me to upload my Excel spreadsheet which has every buy and sell for the 20 year period. Also the equity curve image I posted is rather condensed (Highly Zoomed out 20 year time frame), and looks a lot different once spread out.

[bull....]

whats the trade for tonight?

[Silverlight]

Huskeez,

Very interesting returns, how did you develop the strategy (read about a book about it and modified it) or did you data mine? Is it index based or does the strategy stock pick? 562 trades in 20 years, is about a trade every two weeks, is that the length of the trade or is it just day trading, and the signals only trigger on average every two weeks?

Additionally you should compare the returns to index, and work out the risk and calculate the information ratio of the strategy, to see if the risk you are taking is worth the returns you are getting.

SL

[andysh]

Huskeez,

Very interesting returns, how did you develop the strategy (read about a book about it and modified it) or did you data mine? Is it index based or does the strategy stock pick? 562 trades in 20 years, is about a trade every two weeks, is that the length of the trade or is it just day trading, and the signals only trigger on average every two weeks?

Additionally you should compare the returns to index, and work out the risk and calculate the information ratio of the strategy, to see if the risk you are taking is worth the returns you are getting.

SL

I had a quick look at the index during that time, was 400% increase I think, so vs the 14040% increase, looks good.

[Silverlight]

I had a quick look at the index during that time, was 400% increase I think, so vs the 14040% increase, looks good.

That's a total return, most highly volatile strategies are not successfully commercially as the majority of investors can't handle the volatility. The investor putting in $1m right before the 45% loss period finds little comfort in the fact they now have $550k, despite "the strategy" historically returning 27% p.a, and even though the future returns bear this out, the majority of investors can't stomach that kind of loss.

Even if the IR came back at +1%, like CJ said earlier, in those years everyone "doubts" the system, and bails. You need steel balls to handle it.

Still I would be interested in knowing the finer details, if you are willing to share them Huskeez.

[tosspot]

ok wow. if you could pick any 2 stocks that you would resell in the next month or short term what would they be.

[Huskeez]

Hey Silver,

The system pretty much scanned the DJIA looking for the BUY and SELL signals I had entered. Along with my stoploss.

I had previously been using the system on the NZX and had pretty good results (but over a short period of time) which wasn't enough data to provide to me if over time it would remain successful, So I just did a mass scan over 20 years of data throughout stocks in the DOW looking for the set ups which I had identified and how I would have went if I had acted (Even through the tech bubble burst, and the sub prime bust in 08/09)

Buy And Sells were not limited to anytime frame at all, just when ever the buy or sell signal was activated a trade would be made.




Huskeez,

Very interesting returns, how did you develop the strategy (read about a book about it and modified it) or did you data mine? Is it index based or does the strategy stock pick? 562 trades in 20 years, is about a trade every two weeks, is that the length of the trade or is it just day trading, and the signals only trigger on average every two weeks?

Additionally you should compare the returns to index, and work out the risk and calculate the information ratio of the strategy, to see if the risk you are taking is worth the returns you are getting.

SL

[janner]

Sharp fall in 2007 but then everyone suffered that

Agree CJ.. Every one did suffer that.. BUT .. everyone has had the opportunity to go with the flow afterwards..

Not saying that the return is not impressive.. Just that the return was not impressive until after 2007/8.

[Huskeez]

Agree CJ.. Every one did suffer that.. BUT .. everyone has had the opportunity to go with the flow afterwards..

Not saying that the return is not impressive.. Just that the return was not impressive until after 2007/8.

This is the same strategy but from Jan 1990 - Jun 2006

[blobbles]

Hey Silver,

The system pretty much scanned the DJIA looking for the BUY and SELL signals I had entered. Along with my stoploss.

I had previously been using the system on the NZX and had pretty good results (but over a short period of time) which wasn't enough data to provide to me if over time it would remain successful, So I just did a mass scan over 20 years of data throughout stocks in the DOW looking for the set ups which I had identified and how I would have went if I had acted (Even through the tech bubble burst, and the sub prime bust in 08/09)

Buy And Sells were not limited to anytime frame at all, just when ever the buy or sell signal was activated a trade would be made.

Where do you get the data from to do this type of mass scan Huskeez? I assume this is some modelling done on your own PC? Using what software?

[blobbles]

I am a database/software guy and wouldn't mind looking into this in a bit more depth in the future... having data available to do it would be key.

I am thinking of creating a system whereby it would take into the three main TA factors (MACD, Volume and Stochastic), plus probably a manually controlled future potential value (e.g. for PEB it would be high but for RAK it would be low). Using this information and a whole lot of calculations, I think I could easily do some modelling using MS Access/Excel to create a trading system. It would be a fascinating exercise anyway...

But of course the biggest thing is getting the data, I believe!

[janner]

This is the same strategy but from Jan 1990 - Jun 2006

Thank you Huskeez.. I have never been a believer until I can put my fingers into the nail holes.

More power to your elbow..

[Huskeez]

Hey blobbles,

Data is simple to find. Go to Yahoo Finance, type in the symbol of what stock you would like historical data on and the select "Historical Data" on the left hand side. Enter the Dates you would like the data from , and wooolllaahh! At the bottom of "Prices" there is an icon "Download to spreadsheet" and there you go!

[stoploss]

Have you tried the system with the SPI, ASX futures contract data ?

[blobbles]

Oh right, that's gold dust! Especially because it has a "Download to Spreadsheet" link at the bottom... but my question is can you get a dump of all the data from the entire exchange? Going through each security would be rather time consuming, but entirely possible.Then the next question is how to get a daily feed, but I am sure that would be available somewhere. Hmmm... looks more and more possible...

[Silverlight]

Hey blobbles,

Data is simple to find. Go to Yahoo Finance, type in the symbol of what stock you would like historical data on and the select "Historical Data" on the left hand side. Enter the Dates you would like the data from , and wooolllaahh! At the bottom of "Prices" there is an icon "Download to spreadsheet" and there you go!

I used to toil with Yahoo, set up spreadsheets that extract data live from the website, etc etc, would take up 10 hours a week, had a massive database, with historical figures, prices dividends etc, but then I decided, our much do I value my own time?

The outcome for me was to get a Bloomberg terminal, most people baulk at the cost, but once you decide to upgrade, the cost becomes so minimal. I have any and all data series just a formula away, it has saved me years in time of back testing and monitoring.

[Huskeez]

ok wow. if you could pick any 2 stocks that you would resell in the next month or short term what would they be.

I did a quick scan and as of right now MFT,CEN,ARG are showing buys.