How to trade the Unholy Grails system?

[tech/a]

Craft.

If your placing any form of extra criteria and running it on the whole market then I would agree.

But I'm not.

Specifically to your question of how many prospects.
That varied day to day but was generally 4 - 10
There was only ever 10 in the portfolio so every time a position was closed there would be many to choose from

True random would be to just pick anyone as the next one.
I could have fluked everyone that fitted in my visual.
I'm personally happy that it makes no difference as my selections come from those satisfying the system rules,

As for Gurus mine is a Doctor of Physics. Currently working in Cancer research.
Happy for you to get Bandy involved.I'm surprised he hasn't jumped in already.
If this was about Amibroker he'd b all over it. Bandy I believe to be a Doctor of Mathamatics.

Must be money in books and lectures!

Monte Carlo analysis well that's a whole new topic.

 

[RandR]

I dont want to but in ... but I believe Howard Bandy has already left a comment about what is being discussed between you two (craft and tech/a) on this forum back in February in another thread.

Greetings --

Please permit me to raise a caution. This thread mentions both random selection and TradeSim. I do not have Nick's book or the code being discussed, so my comments may not be needed -- in which case ignore them or modify them as is appropriate.

Use of random numbers and Monte Carlo simulation is valuable under some circumstances. [One of my recent books (and forty-plus years of my professional experience) is devoted to that.] But that analysis usually begins with a set of trades that were chosen deterministically.

Before testing a system that has a random entry or random issue selection, ask yourself whether you would use a random process -- whether you would throw some dice -- to decide whether to trade or to choose among alternative trades. If you would, then continue. If you would have some preference, as I do and most people I work with do, that ranks those alternatives, include the code that describes that ranking -- perhaps as a positionscore -- and form a deterministic system. Then, following the validation process and walk forward runs, use the out-of-sample trades to work through the risk analysis, safe position size in keeping with your personal risk tolerance, and analyze profit potential. These are the steps where Monte Carlo analysis is most appropriately used.

Regards,
Howard

If I am reading this correctly, he is suggesting once you have back tested with monte carlo analysis if you are not selecting your trades randomly (throwing the dice to choose which of your candidates gets the money). Then you NEED to test your discretionary choosing to be able to appropriately posistion size and analyze true risk and profit potential. You cannot assume that by simply running monte carlo analysis you can then select your canditates from further discretionary criteria and have the results fall within the range of your monte carlo analysis. When using anything other then 'randomness' could indeed take you below the range or above. Am I interpreting that correctly?

 

[chops_a_must]

I dont want to but in ... but I believe Howard Bandy has already left a comment about what is being discussed between you two (craft and tech/a) on this forum back in February in another thread.



If I am reading this correctly, he is suggesting once you have back tested with monte carlo analysis if you are not selecting your trades randomly (throwing the dice to choose which of your candidates gets the money). Then you NEED to test your discretionary choosing to be able to appropriately posistion size and analyze true risk and profit potential. You cannot assume that by simply running monte carlo analysis you can then select your canditates from further discretionary criteria and have the results fall within the range of your monte carlo analysis. When using anything other then 'randomness' could indeed take you below the range or above. Am I interpreting that correctly?

Not really.

Because it's not truly random, and not truly discretionary.

It's just a choice between the options provided, which could be given a position score if code was provided.

 

[Wysiwyg]

Have now finished reading the Unholy Grail. Interesting reading about an Index filter. I remember discussing some form of Index filter on this forum after the GFC crash around 2010. Nothing new I am sure but my thoughts were a close below the 250 DEMA. Thread is here somewhere.
Anyways, the crux of the trend trader story is, when the bull runs come along you want to be ready and get on it. Longevity and intensity are the unknown factors but in ones mortal life there could be the mother-of-"all bull runs. You too could look back and say, "I was on it"

I actually have 2 visual filters in my Techtrader method.
(1) The stock cannot be in a clear long term range.
(2) The stock must be either breaking out of a longer term down trend or be In a clear up trend.

These filters may not be easily coded into the system or if hard coded , the system may be a failure. Experience does work better than hard coded ranking criteria and in this system it did. I like the discretionary final decision method.

 

[NewbieTrader1982]

Typically a software application such AmiBroker is used. You purchase the software for a couple hundred bucks, then subscribe to a data supplier such as PremiumData. You load the code up or try and program it yourself.

Every nite you run exploration to identify the trade candidates.

You put you orders in through a broker, such as Interactive Brokers.

Alternatively you might ask Nick how they could trade the strategy or something similar for you.

CanOz

Does this really work long term as its supposed to have a PE over time? Is his testing of the systems really accurate? I read somewhere that certain systems are best for bull markets and others for side trending. If Nick's tested systems are what they claim to be, couldn't anyone really just purchase a few systems for different markets and expect a PE overtime?

Forgive me if my questions are stupid, I am just trying to get my head around all this. Also, assuming Nick's stuff is accurate, I don't understand why so many ppl are trying to write their own code rather than just purchasing from him...what am I missing here.?

Thanks

 

[Trembling Hand]

Does this really work long term as its supposed to have a PE over time? Is his testing of the systems really accurate? I read somewhere that certain systems are best for bull markets and others for side trending. If Nick's tested systems are what they claim to be, couldn't anyone really just purchase a few systems for different markets and expect a PE overtime?

Forgive me if my questions are stupid, I am just trying to get my head around all this. Also, assuming Nick's stuff is accurate, I don't understand why so many ppl are trying to write their own code rather than just purchasing from him...what am I missing here.?

Thanks

NT whats PE?

As to your questions - the testing he has done is robust. The problem with all systems is stationarity. You design and test on past data that validates your signals so you deploy it in real time. The market is always evolving from bull to bear to just out right depressed and going nowhere. Volatility increases and decreases then the stocks that are trending greatly change. For a few years its small caps then they just bog down for years and the Large caps start running. Just at the time you notice they are the hot thing they stop. When someone starts a system like Nick's inevitably its after a hot period and running into a 6-12 month period of drawdown. They trade it for 8-12 months and it goes nowhere so move on. Nicks systems are certainly a longer term commitment.

People code their own systems looking for better performance. Either different instruments, timeframes, Max drawdown stats, etc or just outright better returns.

 

[pixel]

I don't understand why so many ppl are trying to write their own code rather than just purchasing from him...what am I missing here.

People write systems for two reasons:

1. to make money from selling it
   
2. to gain an edge over other traders

As soon as a significant number of traders follow the same "system", they won't have an edge over others (each other) any more. From a certain threshold level, a smart algo (bot) will detect the setup required to lead to a particular mob reaction, where a large number of system traders will either buy or sell. ... and the "Shooting the fish in a barrel" fun is about to begin.

 

[NewbieTrader1982]

NT whats PE?

As to your questions - the testing he has done is robust. The problem with all systems is stationarity. You design and test on past data that validates your signals so you deploy it in real time. The market is always evolving from bull to bear to just out right depressed and going nowhere. Volatility increases and decreases then the stocks that are trending greatly change. For a few years its small caps then they just bog down for years and the Large caps start running. Just at the time you notice they are the hot thing they stop. When someone starts a system like Nick's inevitably its after a hot period and running into a 6-12 month period of drawdown. They trade it for 8-12 months and it goes nowhere so move on. Nicks systems are certainly a longer term commitment.

People code their own systems looking for better performance. Either different instruments, timeframes, Max drawdown stats, etc or just outright better returns.

Thanks for your feedback. When I said PE I meant Positive Expectancy. So is it safe to say that with any system, eye balling or certain criteria is applied first? I think @tech/a said somewhere his two visual criteria first is "1) The stock cannot be in a clearly long term range. (2) The stock must be either breaking out of a longer term down trend or be In a clear up trend."

Does it then mean that Systems Trading could see you go 24 months with no profit? I like the mathematical edge (basically means you are becoming the casino) with a Systems approach...just still trying to get my head around it.

Are you a Systems trader?

Thanks again.

 

[NewbieTrader1982]

As soon as a significant number of traders follow the same "system", they won't have an edge over others (each other) any more.

Interesting point. I wonder how much this actually would be an issue since my strong hunch is most people wouldn't have the correct mindset from the outset to implement the system anyway, Am I wrong to assume this?

Thanks

 

[NewbieTrader1982]

People code their own systems looking for better performance. Either different instruments, timeframes, Max drawdown stats, etc or just outright better returns.

For the average person (that's myself), I'm guessing they would have buckleys chance of coding something better than Nick?

 

[NewbieTrader1982]

The problem with all systems is stationarity. You design and test on past data that validates your signals so you deploy it in real time. The market is always evolving from bull to bear to just out right depressed and going nowhere. Volatility increases and decreases then the stocks that are trending greatly change. For a few years its small caps then they just bog down for years and the Large caps start running.

Ok I think I understand this. So wouldn't that mean that if someone had an endless supply of capital then over the long term they would have to perform very well, since human psychology never changes and eventually it will come back to the signals your systems works on? If that assumption is correct then I guess the issue is twofold, 1. No one has unlimited funds and 2. No one really wants to wait X amount of time to eventually get their projected PE return if it ends up being 5-10 years.

Or am I completely off-track?>

Thanks

 

[Roller_1]

Thanks for your feedback. When I said PE I meant Positive Expectancy. So is it safe to say that with any system, eye balling or certain criteria is applied first? I think @tech/a said somewhere his two visual criteria first is "1) The stock cannot be in a clearly long term range. (2) The stock must be either breaking out of a longer term down trend or be In a clear up trend."

Does it then mean that Systems Trading could see you go 24 months with no profit? I like the mathematical edge (basically means you are becoming the casino) with a Systems approach...just still trying to get my head around it.

Are you a Systems trader?

Thanks again.

Remember that there are a infinite number of ways to approach the market and the same goes for systems trading. You can trade a monthly, weekly, daily or intra day. It depends of your goals/ beliefs.

I trade using a systematic approach but use Mean Reversion strategies that a trade may last 1-7 days. This style 'usually' creates a higher win % than some of the longer term systems in unholy grails. Long term systems can have long periods of Drawdown so can be hard to trade psychologically, as TH said its a long term commitment (as any trading should be).

I'd suggest going to thechartist.com.au and read some of the articles on there. I think there's a fair bit of info you can access

 

[NewbieTrader1982]

Remember that there are a infinite number of ways to approach the market and the same goes for systems trading. You can trade a monthly, weekly, daily or intra day. It depends of your goals/ beliefs.

I trade using a systematic approach but use Mean Reversion strategies that a trade may last 1-7 days. This style 'usually' creates a higher win % than some of the longer term systems in unholy grails. Long term systems can have long periods of Drawdown so can be hard to trade psychologically, as TH said its a long term commitment (as any trading should be).

I'd suggest going to thechartist.com.au and read some of the articles on there. I think there's a fair bit of info you can access

Really appreciate your feedback. Actually I am purposely not going to thechartist website yet as I am hoping to get unbiased information/feedback/bearings here first before I go there..call it part of my Due Diligence process

Very interesting point you make re timeframes. I know for my own goals, I am more interested in shorter term cash flow/profits (say 1 week to 2 months) rather than longer term profits which I have covered by other asset streams. I like your timeframes and using a Systems approach. I guess my question is, if you have a higher win % with less frequent DD, does that then mean when you do have DD it eats up weeks or months of your profit in one go?

Really appreciate your input mate.

NT

 

[pixel]

Interesting point. I wonder how much this actually would be an issue since my strong hunch is most people wouldn't have the correct mindset from the outset to implement the system anyway, Am I wrong to assume this?
Thanks

Generally speaking, your hunch is probably spot-on. However, systems that are advertised and promise a decent ROI will tend to attract a number of "serious" investors/ traders that have the inclination and discipline to follow the rules quite meticulously. It's those systems that a speedy algo can "learn" and take advantage of.
You may have heard of HFT boxes sitting right next to the main trading computers in Stock Exchanges, squeezing in their own buy and sell orders nanoseconds before yours? That may not immediately drive a system trader into ruin, but it certainly blunts the edge by clipping a tick or two off every trade.

 

[NewbieTrader1982]

.
You may have heard of HFT boxes sitting right next to the main trading computers in Stock Exchanges, squeezing in their own buy and sell orders nanoseconds before yours? That may not immediately drive a system trader into ruin, but it certainly blunts the edge by clipping a tick or two off every trade.

Nope complete newbie in all this, but I've heard of it now Very interesting. In that case, I guess private traders might want to edit the code of something to try to get that edge of other traders using the same code. Again, I wonder how much of a difference it really would make? Since I'm assuming the money System traders are making is from the masses and not from their peers?

 

[pixel]

The problem with all systems is stationarity. You design and test on past data that validates your signals so you deploy it in real time.

+1
That's why Systems should have a use-by date, just like other perishable goods.

 

[NewbieTrader1982]

+1
That's why Systems should have a use-by date, just like other perishable goods.

Would you then not go back to the System when the market goes back to similar signals that the System is designed to do well in?

Cheers

 

[pixel]

Nope complete newbie in all this, but I've heard of it now Very interesting. In that case, I guess private traders might want to edit the code of something to try to get that edge of other traders using the same code. Again, I wonder how much of a difference it really would make? Since I'm assuming the money System traders are making is from the masses and not from their peers?

If you can afford to place a HFT box into a Stock Exchange, you're no longer a System Trader of the kind we're talking about here. Those boxes are owned by instos and make up over 50% of the entire turnover. Yes, they take it from the masses by "learning" when a sufficient number of us are likely to buy or sell - and then they jump in and clip our wings. The greater the number of individuals trading in a similar pattern, the more clearly it will show, even if it's only a small fraction, percentage-wise.

 

[Trembling Hand]

Would you then not go back to the System when the market goes back to similar signals that the System is designed to do well in?

Cheers

At what level do you decide that the market has changed and the other system needs to be turned on? Probably after its clear and again about to change! We are talking about making predictions about future events.

 

[Roller_1]

Really appreciate your feedback. Actually I am purposely not going to thechartist website yet as I am hoping to get unbiased information/feedback/bearings here first before I go there..call it part of my Due Diligence process

Very interesting point you make re timeframes. I know for my own goals, I am more interested in shorter term cash flow/profits (say 1 week to 2 months) rather than longer term profits which I have covered by other asset streams. I like your timeframes and using a Systems approach. I guess my question is, if you have a higher win % with less frequent DD, does that then mean when you do have DD it eats up weeks or months of your profit in one go?

Really appreciate your input mate.

NT

It's not neccessairy less frequent DD but i've found DD to be shorter in duration because you are trading a lot more frequently and can dig out of drawdown quicker. If you have a longterm system that uses a index filter for example you might not place a trade for 6-12 months, while it may reduce the MAX Dd of that system it's hard to make money if you aren't trading.

does that then mean when you do have DD it eats up weeks or months of your profit in one go?

Not really but anything can happen. Sure you can go into a 10-15% DD and you don't make that over night or in a week clearly, but the PE comes from long term application of any system you have to stick at it. It comes down to your risk tolerance too.

Unless you have a 95% win rate (even then) there will be DD