Developing a System to Trade Live

[RodC]

Re: Van Tharp- Trade Your Way to Financial Freedom

Moderator's Note: This thread is for general discussions on systems development and testing. Please start a new thread if you have a specific system in mind or search for an existing thread on systems testing.

Im happy to work with you and all here in developing another method which we can test and trade live like we have been doing with T/T for the last 3 yrs.

A larger interest gives better feedback and more minds participating.
If there are also other Amibroker or Metastock/Tradesim users around that will also be a great help.

Hi tech/a

Are you still interested in developing a system to test and trade live?

I'd like to have a go at developing a system, but to be honest I don't know where to start. I do however, have Amibroker so I guess I'll need to learn how to drive it properly.

regards,

Rod.

 

[tech/a]

Re: Van Tharp- Trade Your Way to Financial Freedom

Sure.

 

[mit]

Re: Van Tharp- Trade Your Way to Financial Freedom

If there's a vote, given that TT is out already out there as a working model of a Long Term System I would suggest a short term EOD system. With the possibility of a more volatile market a ST System may have a smoother equity curve.

However, a note for beginners a ST system is harder psychologically to trade. With a long term system if you are a bit wobbly with your entries and exits it doesn't matter as much.

MIT

 

[cabbie]

Re: Van Tharp- Trade Your Way to Financial Freedom

G'day All,

If you don't mind, you can count me in, I've been tinkering around with AB & would be interested in trying to develop a structured system.

I've read a bit of Van Tharp & some of Tech/a's threads (who hasn't! ), so I've got any idea of the concepts of expectancy, MM & risk management etc so I'm keen to put it all into some code.

I'm finding the hardest part is trying to develop a testing methology without being sidetracked on testing different combination of indicators various stops etc by using AB's optimiser & trying to interpret the results. Despite all the warnings of over optimising etc.....must be the programmer in me getting carried away!

What I tend to get is multi MB report files, making meaningful analyse difficult. So I think the problem is trying to most releant parameters for indicators to test etc.


I wrote a bit of code that will looks up the base index (or any index) of a stock & buys only if the index is trending (could be modified for any trend indicator). I'll tidy it up sometime & post it any one is interested, ( although there's probably other stuff around that does the same).


Also can anyone tell me how relevant the book "The Trading Game" is to shares, Guppy's review on Moneybags says it's only for futures is that correct? as the contents seems generally relevant.



Regards

 

[cabbie]

Re: Van Tharp- Trade Your Way to Financial Freedom

I forgot to ask, what are you calling ST? 1-5 days or < 30 days?

Regards

 

[mit]

Re: Van Tharp- Trade Your Way to Financial Freedom

I did follow one for awhile and it tended to be 1-5 days (longer if the trend is good).

MIT

 

[tech/a]

Re: Van Tharp- Trade Your Way to Financial Freedom

Well.

In the 12 yrs Ive been doing this I'm yet to find a short term system for stocks that is longterm profitable.
I've had a go at many but in the end they all tank.

Now I have seen good short term futures systems,what have you all got in mind.

MIT like to kick off with any suggestions.
What sort of system testing software do you have?

Do we have any proficient Amibroker users here?
I have Metastock and Tradesim enterprise.

Are people here trading anything short term profitably that we could tinker around with so that we arent using the exact method but have perhaps a starting template?

 

[mit]

Re: Van Tharp- Trade Your Way to Financial Freedom

Tech,

I'm an Amibroker user. I was using for awhile some systems that zipzap had mentioned on reefcap back in 2003. I thought shares as I thought it might be more accessible to more people (me included). You are right as it is tougher and started to become less worthwhile when CMC started charging commission and I started to get interested in your techtrader.

But I am happy with Futures. I have EOD spi data going back to '83.

MIT

 

[mit]

Tech,

I have finally read Larry Williams book on short term systems. This is probably what has got me fired up about short term systems. Although stuff from this book has been tried on forums before, I feel that people have only taken a single example and tried to make a system out of it rather than the book in total.

I am just throwing up some ideas but am happy even if it is longer term systems which is easier for beginners. Should we have a vote?

MIT

 

[tech/a]

How about we approach this a little differently than most.

Lets start with.

Trading universe/s.---What are we going to trade.Can be a future/s or a universe of stocks.---or both--(Think there would need to be different systems for stocks and futures.).

I presume for stocks we are trading LONG---Futures both Long and Short.

From here portfolio and Initial capital.

Suggestions?

 

[RodC]

I was thinking stocks, but that's probably mainly because that's what I'm most familiar with, never really thought about futures - maybe I should.

For stocks I was thinking of a medium/long term system, mainly because I like the idea of letting profits run. I also imagine transaction costs and sudden adverse movements would eat into the profits of a short term system.

What does everyone else think, do we have any futures traders here who are willing to participate, if that's the path we choose?

Rod.

 

[tech/a]

Just quickly I'm with Rod purely because I think that most here and many 'lurkers" are share traders.

As for timespan why dont we let the results tell us what timeframe its in.
The system will have an average length of trade determined in the testing.

Running profits maybe 20 ticks in a 4 tick risk---thats a good run---but to have a postive expectancy of 5:1 would require much larger wins every now and again.---this ofcourse is linked to strings of losses and number of winning trades to losers----But I'm getting ahead of this!!

Back to universe----I think we should do some tests on best perfroming universes.
IE
Go back 5 yrs or more(I use 8)Filter stocks with particular characteristics and then work forward to see howmany have out performed--the index--the ASX 200 or whatever.Have in our pool of candidates those that are going to be likely out performers.

Like any Margin list this list should be pruned and honed and added to each 6 mths or so according to our criteria to be in the universe of tradable stocks.

 

[RodC]

I agree, I think most are certainly share traders as shares are seen as more accessible (especially to those starting out) than derivatives. Many will never go further (or wish to) than shares.

Letting the system determine the timeframe is a good idea, that way we aren't compromising the returns by imposing an arbitrary timeframe onto the trades.

Do we want this to be a margin loan system? If so then obviously any stocks in the "universe" must also be on a margin list.

Rod.

 

[GreatPig]

I have AmiBroker and am reasonably proficient at programming it - plus I'm a programmer by profession. However, I only chart stocks, not futures or derivates (although I do have options and warrants data), and don't have a huge amount of spare time.

Still, I'm happy to help where I can.

Cheers,
GP

 

[tech/a]

Thanks GP you will be able to do things M/S and Tradesim cant. Will be very handy.

Rod.
Other than stock selection dont think it matters wether margin or not.
You can trade 100K or 30K on margin.
Better leverage for money invested when profitable.
But that can be upto the trader.

I maybe wrong but people may well adopt some of those idea we come up with into their own model and produce something not exactly like the one we come up with.
After all what we come up with will be only "AN" idea not "THE IDEA of IDEAS".

Even a margin list can be further filtered.

 

[mit]

As for timespan why dont we let the results tell us what timeframe its in.
The system will have an average length of trade determined in the testing.

While this would be true for medium to long term systems short term systems are markedly different to longer term systems. So it probably depends on how we determine the initial candidates for a system.


MIT

 

[Milk Man]

A couple of thoughts-

Would the smaller stocks out of the asx300 or margin lists be a good idea- I know techtrader kind of does this but maybe theres a better way of sorting out the least volatile. Maybe via market cap rather than a dollar value or does this not make much difference.

I also thought a short and long system could be a good idea. You mentioned, Tech, that you had a signal to exit all positions that tested well. Maybe a signal on the XJO/XAO to short would be good too.

A little askew: Is there an uptick law for shorting in Oz?

Anyway, I thought since stocks tend to move down faster than up maybe this would be a good idea.

Has Techtrader been tested in reverse: new low then cross above 180 day EMA?

Hope this gets the wheels turning- probably won't be any new ideas but.....

 

[tech/a]

While this would be true for medium to long term systems short term systems are markedly different to longer term systems. So it probably depends on how we determine the initial candidates for a system.


MIT

Hmm I'm interested in your explaination of "markedly different"

The only things I find different are
(1) length of trade governed by tightness of exit.
(2) Generally more trades---cause your in a trade less time.
(3) With winning systems (I've only seen winning futures short term methods) More winners.
(4) Less Reward to Risk.
(5) Less runs of consecutive losers.

Loakglen
Some ideas here.
Personally I'd be looking for stocks which have the potential to outperform---how do we govern this.----technically we can set parameters--Fundamentally we can choose those seen as best value.Margin lenders sort of do this for you when only allowing certain stocks to trade.We can combine all of the above and anything else.

Personally I dont think trading short is practical on a portfolio basis.If trading futures then its a must---but we arent going to trade futures this time round.

Techtrader---dont want this to be anything like T/T.We can learn from lessons found during its developement though.

 

[Milk Man]

Tech, do you still like the buy high sell higher approach?

If so a Darvus style entry might be good.

I still think a short/long system would be worth testing- potentially riskier and harder to implement but the numbers would show if its viable. Definitely for some kind of portfolio exit indicator or similar though.

Spose we should put the horse before the cart though: I reckon the smaller caps out of the best performing margin list for the 'universe'. Might seem like deja vu to you Tech but it seems sound logic.

 

[tech/a]

Entry I'm not fussed---as long as the numbers add up then to me thats all that matters---how we get to those numbers is only important as to how good those numbers can be.

Yes I dont mind the lower valued Margin stocks but if someone would like to give me the list rated by market cap then that would make some purists happy.

I think the list could be further honed.
Id even thought a while ago eyeballing EVERY stock in the margin list I use 278 stocks---I might just do that---but those I keep and those I discard would be my calls and some may not agree with what I look for or see.

But its easy enough to test different universes---infact its preferable.

Ill get back with the BT list all stocks under $10 and all filtered by eyeball to those in an obvious uptrend OR Not making new lows.(I'll post examples,infact the whole list).
This can be but one list.The Full list can be another.

I'll leave it to you to play with a short side to a portfolio---I've spent many hrs without success (My definition) so please excuse me for not wanting to invest anymore time.Short and long sides dance to different drums.
Its not a matter of simply reversing an exit and going short---Pure disaster.
Hint---Youll be short in trades where youll still be long.(Well should be).

Have we agreed on Daily? Rather than Weekly?