Diversified Trend Trading Approach

But you and I know that we are very special.....pffffttt.....those trend following nancy-boys......../tic .

Have a look at the YTD performance of the short term systems versus the longer term systems. This volatile up and down market filled with sumos is currently taking its toll on the longer term guys.

Be aware of those biases that are so prevalent in an evolved brain. Payman Taei suggests that "seeing is believing" might be a misnomer in some circumstances if left to our heuristic brains. Does this have implications with discretionary trading? I will let you be the judge.

Thanks C. I am only about 90% satisfied with what I have so far and it is too early still in infancy stage so trying not to get carried up. Last year was all about manual exits with intense watching around the clock on my mobile but using the same underlying principles of universal time action. This year, most of it is automated to bring further consistency with execution. I perform better when I sleep during London when most of the action is. Now I just got to worry about having someone monitor my portfolio on my behalf :- The remaining 10% is still manual requiring code tweaks several times which require 'human intelligence' to set the right value of constants.

Anyways. Thanks for your time. I will stand aside to stop the clutter and share where I am at month end to do a quick compare with your portfolio to assess my 'relative' performance if you are open to it (and if I have not blown my account lol)

Adios

This is how intense focus looks like:

Original



and instrumental cover:



And Down to the Bone groove:

Great shot R. :-)

Nice mate. You have made my night chilled and melancholy.

Hillarious chat between Elon Musk and shortsellers of Tesla.

https://twitter.com/elonmusk/status/...=244+272699400

Could you please comment on the "Risk to Ruin" screenshot I shared? What numbers or % I should be targeting for to prevent account blowup? Can it happen with what you are seeing based on the volume of potential consecutive losses? I have several creative features to prevent volatility based account blowups but they all are at the mercy of broker's execution and liquidity. I have a good handle on liquidity and volatility but no control over freezing, network latency, connectivity etc. If my automation stops more than one minute while I am sleeping, I am f*cked as there is a breakdown in Time Action which affects Frequency Action which then results in discrepancy in Risk, Exposure, Margin, Drawdown, Entry/Exit management functions. If I allocate 5mins to "gap exposure", it creates spikes in my equity/balance curves which I don't like to see. Perhaps, I just have to live with this.

Elon Musk is getting bored of his success, and pushing the limits. he is the victim of his own success.

if I were him; wtf I need to manufacture 1000 of cars per week?. just give them what can be done; if the factory can do 10 a week; so be it.

the factory and its employees has a capacity (i.e. daily range). you can only milk it what it can give you.

sleeping in the factory to get x number of cars out; will be seen as a weakness by his employees and by the market.

He could perhaps promise his employees extra bonus if they fix the problem; and spend some quality time with his ladies

Simply being with employees is not enough for them to get motivated. monetary benefits all that matters at the end of the day.

Check out this movie "Margin Call", it shows this exact point.

he is my #1
when I was kid, never loved Apollo rockets as they were not reusable, always dreamed of reusables n he did it
I was always dreaming electric cars n making my own battery toy cars while dreaming d real ones, he did it too


in fact I love him

he is over promising, I understand d reason which is d finance n cash flow, but over promise creates more cash issues.
n there r a lot stupid idiots non sense ppl that want to see him bankrupt where this guy has invested only in USA, not understandable

The hyenas are circling. It would be a damn shame to see vision bested by material greed and envy....buy hey....that's life....well at least that's what people say. :-)

Mate....Risk of ruin relates to market risk only. Chuck your stats into the risk of ruin calculator here from Perry Kauffman and then have a play with a few simulations (eg. up to 30 times). I prefer this method as it bases risk of ruin on your max drawdown tolerance. Also remember that it is a probability estimate only (like how we determine an edge)....not a guarantee and the confidence estimate improves as your trade sample size increases.

Here is a post on market risk that might help.

In regards to the other risks you have mentioned...these fall outside the scope of market risk but dependent on likelihood and consequence sound pretty serious to me given your margin exposure. With these risks on the table...that would certainly change the game for me. If you can't mitigate against them then you have to make a decision whether to accept these risks or not. If you were trading a regulated fund these would need to be conveyed to your investors as they are material in nature.

It sounds like they are more extreme risks than say counterparty risk (eg. broker liquidity) and your dependency on 100% connectivity is such a great issue unless your VPS provider (if you have one which I would assume is mandatory in your case) is willing to insure against these events.

Thanks for sharing. I would have a read the link on market risk few times to get it LOL. Further reading (prior to your post) on Risk of Ruin reveals yet another obstacle causing meaningless stat but haven't made a firm mind yet.

In coming months, I hope to share my definition of portfolio diversification. I think you are going to like it.... just don't want to get too much ahead of myself without enough stats to back by ongoing research :-)

Cheers mate:-) Risk is always present....and can't be avoided....simply transformed. One of those constants that lie in complex systems due to non linear power laws that actually define complex systems. The best we can do is to be aware of them so they don't blindside us....and hopefully transfer those that we are aware of to other participants. That is where the edge lies.....aka those that get nailed at your expense.

Old phantom saying on risk management "if you want to destroy a rainforest understand the system. Don't clear fell it is overkill...simply kill the few trees that provide the canopy"

Risk Management is a fascinating area of research. It's mind boggling. My focus was controlling Max DD first and foremost and to remain above the Total deposits in the account. I want to now move to the area of measuring Portfolio performance as a measure of volatility. As I have conveyed before, I make more money when there is volatility in the market conditions regardless of direction. Could you please help me assess how I am doing in regards to volatility measures. i.e. standard deviation, sharpe, etc.? where should I prioritize my focus on and in what sequence (granted that I don't have too long a duration but I don't want to wait for a year... it needs to be a progressive and continuous improvement journey :-)

I would appreciate your thoughts as you have dealt with this a whole lot longer than I have. SnP volatility for past 5 years is about 17% (i read somewhere and I have no idea how to measure lol) so max loss can be twice the historical volatility (per experts). Is that the right way to measure portfolio performance based on volatility measures and calculations?

Here's how my numbers look for May thus far. (Starting balance is $1K for the first 6 account and $6K for the last account in the list)

EDIT: I guess what I am trying to figure out is this... if my profit taking (and loss if 10% maxdd from peak equity is triggered) happens in market volatility zones, does that always equate to increase (or decrease) in volatility based portfolio risk measures? Example: if equity explodes to 10% from the previous Peak Equity levels, I do a close-all of all open positions. Consequently, if the equity drops from previous peak equity, maxDD is triggered which initiates partial closing of open positions for as long as equity remains below the maxDD threshold. This means standard deviation (in this example) will be in the range of 5% above and 5% below the midpoint of equity curve. This could happen all day in calmer or paused market conditions although I have a technique to address stalled markets to minimize the effects without using any any technical indicators like ATR/ADR etc

Spot on; the hayenas are definitely circling. no more Tesla stocks to short anymore.

they can smell blood miles away, watch out Elon; no sleeping in factory couch is gonna help. that will only attract more hayenas!

V

I am not sure where to start other than state a very simple physical principle in relation to risk management......that the summation of discreet equity curves with unique volatility signatures reflected by simple statistics such as Drawdown, sharpe, standard deviation or sortino ration etc. all improve with diversification. You may be familiar with Markowitz and the efficient frontier. The theory behind this is sound but its practical application can be applied simply by superimposing return streams on top of each other.

Assume you invest 100% into a particular strategy to produce a single equity curve. That equity curve has it's own unique signature which is a reflection of how that system fared under market conditions.

There is the price action of the market itself and there is also a derivative consequence reflected by that system in how it responds to those market conditions.

If the market was purely random....then the derivative consequence for all systems that attempt to trade it is a random equity curve. Now to have any hope of having an edge in any form of market we must assume that real signals exist among the noise. We won't be able to actually 'see' them or be able to discriminate between real signals and noise using traditional methods....however just as a sound engineer can isolate signals from noise through the superposition of sound waves on top of each other.....we can achieve the same result through diversification. Namely the superposition of equity curves on top of each other to boost the signal to noise ratio.

Go to the physics lab or to your bathtub and start splashing around...and you will see the principle in action. Because the peaks and troughs of random noise are just that....'random'....then if you superimpose two return streams on top of each other you get amplification of real signals and dampening of noise. What diversification does is it increases the signal to noise ratio just by the simple principle of superposition.

The same with drawdown. Assume we have a strategy that achieves a max drawdown of 20% at some point in time and we have another strategy that achieves a max drawdown of say 20% at another point in time. The result of superposition where you summate the result to produce an equity curve of the portfolio results in a max drawdown that will always be less than the individual constituents if you allocate 50% to each system. The reason for this is the difference in the timing of when these events occur. The result is a higher risk-weighted return as you can scale up the portfolio result as it has a lower volatility risk profile and achieve a higher total return than the sum of its parts in isolation. That is where the free lunch resides in diversification ....and that is where your edge arises simply through the more efficient application of finite capital towards volatile return streams.

Now provided that each singular return stream in the composition has finite 'capped risk' which ensures that the failure of one component won't bring the entire portfolio down you have a potential contributor with a unique equity curve that can assist in your total blend.

Each strategy has different return profiles....some much more volatile than others....but that does not mean that they are more 'risky' strategies. Convergent strategies have negative skew and those supposedly fantastic equity curves with limited volatility....that explode at the end of the day as the risk was not capped. The worst of these blighters is the martingale progression in which many are duped about the low vol nature of these strategies and fail to understand the intrinsic risk that hasn't been released by the system. I try to avoid these time bombs like the plague and can achieve great *relatively' linear equity curves with full risk release by diversification provided I use divergent techniques with full risk release.

As you say under full transparency using unrealised losses, you would be able to see the building risk inherent in these systems.

What this should be telling you V is that sharpe, drawdown, sortino etc...are useless if you do not understand the underlying assumptions inherent in them. Because I use divergent systems which have full risk release at the time of exit....then these stats do have some meaning to my strategies in terms of overall risk profile...but they have no meaning for those that make convergent plays.....and are really more dangerous tools of mass destruction.

In a nutshell V....be less concerned with the statistics themselves and more concerned with the nature of the equity curves which is a reflection of how your system fared under given market conditions. Some people say that the correlation statistic is an important measure....however that single number does not tell me much about the nature of the strategy. I can see correlation and how it changes under varying market conditions just by a visual examination of the equity curve against the market. It tells me so much more in terms of under what conditions the strategy is correlated versus under what conditions it isn't.

Statistics is a tool only and has very limited use if you use these tools blindly.

You say that backtests are not important....well I beg to differ. They are everything. They provide an objective road map for the future. Without them you are rudderless. They are the basis for which you determine the characteristics of your strategy and decide what future objective reasons you are going to use to guide you in uncertainty.

V...in relation to your performance stats....it's a good start but you need to build yourself a model to start tweaking the allocations and position sizing and more importantly deriving an equity curve of the entire portfolio. At the moment....they are too leveraged for my tastes but that's where additional diversification in terms of systems, timeframes and instruments can assist you. Also if you feel that backtests are useless then it is going to take quite a while before your results have any meaning. It is your call mate....but I know you understand the principle of the Law of Large Numbers. The only way I am going to detect the likelihood of a biased coin is through sample size mate.

Cheers

C