Come Discuss Systematic Trend Following & Setting Up A Fund

globalarbtrader said:
This doesn't work
https://qoppac.blogspot.co.uk/2015/11/random-data-evaluating-trading-equity.html

GAT
More...
It's a variation of a standard CPPI approach, with it's positives and negatives. Obviously, the original strategy needs to have a positive drift and you are introducing a significant amount of path dependency. Like in any other hedging approach, you reduce variance of the strategy at the expense of the overall expectation (whichever way you like to measure it).

PS. I don't know what you mean by "doesn't work" and I am too lazy to read this guys musings.

Mtrader said:
Are drawdowns not always peak to valley? :confused:
More...
Nah, a lot of people talk in terms of HWM and back.
 
Stratos Capital said:
After doing some searching around for ways to reduce drawdowns, I am leaning towards methods that DONT require adding more rules to the system or "predictive" methods that try to time drawdowns.

Methods that seems to be more robust in my opinion include:

1) Studying correlation between markets. Selecting markets that generally do not correlate positively. I admit theres a risk that correlations may break down.

2)Diversifying into multiple strategies. Could be similar trend following strategies, different time frames, or entirely different strategies.

3)Simply diversifying into more markets. I read some CTAs trade more than 100 markets.

Using options strategies, hedging, refining signals, adding more complexity to a trading system, these are things that I try to avoid as much as possible.
More...

Those are all excellent ideas. You are on exactly the right track.

GAT
 
Stratos Capital said:
Hey GAT, could you maybe offer some advice on what are the most important elements of risk management at places like AHL?

I am also wondering what are the major pitfalls or difficulties in raising capital?
Because I read that most funds fail in this area. Seems like a good idea for me to start learning about the business part of trading.
More...

Here are some ideas, from easy to hard:

  1. Putting a floor on volatility estimations (to stop too much leverage being piled on when vol gets unusually low).
  2. Worrying about correlation shocks (what is my risk if all correlations go to +1 or -1 in the wrong direction?)
  3. Putting a ceiling on estimated risk, using current covariance matrix
  4. Worrying about other measures of risk i.e. putting a ceiling on my 1% VAR using current covariance matrix
  5. Scenario analysis (what happens if October 1987 happens now, with our current positions? What happens if 'made up bad event' happens now?)
  6. Using implied vol as an input into volatility estimation
I've implemented 1-3 in my own system. I personally think 4 is a waste of time. 5 requires a lot of coding. 6 needs you to pull in option surfaces and is a huge amount of work - but worth it; if realised vol is unusually low going into an expected shock event (economic number, or election).

Unfortunately raising capital isn't a speciality of mine. Plenty of people on this board who can hopefully help you. Might be worth creating a new thread, possibly in a different forum.

GAT
 
globalarbtrader said:
Here are some ideas, from easy to hard:

  1. Putting a floor on volatility estimations (to stop too much leverage being piled on when vol gets unusually low).
  2. Worrying about correlation shocks (what is my risk if all correlations go to +1 or -1 in the wrong direction?)
  3. Putting a ceiling on estimated risk, using current covariance matrix
  4. Worrying about other measures of risk i.e. putting a ceiling on my 1% VAR using current covariance matrix
  5. Scenario analysis (what happens if October 1987 happens now, with our current positions? What happens if 'made up bad event' happens now?)
  6. Using implied vol as an input into volatility estimation
I've implemented 1-3 in my own system. I personally think 4 is a waste of time. 5 requires a lot of coding. 6 needs you to pull in option surfaces and is a huge amount of work - but worth it; if realised vol is unusually low going into an expected shock event (economic number, or election).

Unfortunately raising capital isn't a speciality of mine. Plenty of people on this board who can hopefully help you. Might be worth creating a new thread, possibly in a different forum.

GAT
More...
Thanks a lot GAT. This is good stuff.

1. Limiting leverage when volatility gets unusually low seems like a really good idea.
My current system uses volatility based stops and my fixed position sizing rules automatically take on more risk when volatility is low. I should spend some time to look into this area.

5. Implied volatility could be something worth looking into as well. So far I have been relying on historical volatility.

Thanks again for sharing these.
 
Stratos Capital said:
It seems to me that doing this will inevitably reduce returns.
I must admit I am very unwilling to reduce drawdowns at the expense of returns.
We actually want to differentiate ourselves from the competition, which is so obsessed with volatility.
So the ability to generate strong returns is important to me (of course, without blowing up the account).

The better alternative for me would be to diversify into strategies that are not correlated to trend following (perhaps mean reversion). The idea is that if I can find such a strategy that can generate similar returns and drawdowns.
Then just allocating to both should help reduce drawdowns without sacrificing returns.

What does everyone think?
More...

If an account has a 47% draw down than I would say you are not trend following. volatility is a killer of all returns. If you take your strategy and compare it to a strategy with the same cagr but without the volatility you will make your investors more money at the end of the day all else being equal.
 
yobo said:
If an account has a 47% draw down than I would say you are not trend following. volatility is a killer of all returns. If you take your strategy and compare it to a strategy with the same cagr but without the volatility you will make your investors more money at the end of the day all else being equal.
More...
47% is the max drawdown ever recorded for a 30 year backtest.

If i look at hedge funds that can manage 25% CAGR but with a lower max drawdown, these are usually the ones who have been around for just a couple of years, so investors are not seeing the real risk with investing in these funds.

If I look at big CTAs and successful trend followers who have been around for more than a decade, the tradeoff is the same everywhere.
Those that have suffered only 30+% draw downs, they currently give a CAGR of sub 20%.
 
Stratos Capital said:
5. Implied volatility could be something worth looking into as well. So far I have been relying on historical volatility.
More...
You don't need to pull volatility surfaces, a vol index will give you plenty insight into the upcoming events (CBOE maintains a VIX-like fair variance index for almost every product out there, though unlike VIX you can't trade them).

HobbyTrading said:
In case you hadn't noticed, "this guy" and @globalarbtrader are the same person. He maintains his own blog at https://qoppac.blogspot.co.uk
More...
No, I have not noticed, so thanks.

Stratos Capital said:
47% is the max drawdown ever recorded for a 30 year backtest.
More...
In that case, let me ask you a few more questions that most FoFs would be asking
What is the strategy correlation with the equity market and VIX?
Is there a significant correlation to any macro fractors?
What is the worst case N-draw resampling drawdown in the modern markets (say since 2000 or so)?
What cases/market styles would be detrimental to the strategy? What are your contingency plans for those situations?
 
Last edited:
sle said:
In that case, let me ask you a few more questions that most FoFs would be asking
What is the strategy correlation with the equity market and VIX?
Is there a significant correlation to any macro fractors?
What is the worst case N-draw resampling drawdown in the modern markets (say since 2000 or so)?
What cases/market styles would be detrimental to the strategy? What are your contingency plans for those situations?
More...
The current strategy we have trades 33 markets including softs, metals, energy, currencies, rates, and equity indices. We will add more markets over time to further diversify.

As for correlation to the S&P 500 index, for the past 20 yrs the strategy ran a correlation coefficient of 0.54. During major bear markets, trend following strategies have even generated negative correlation.

Correlation to the VIX is higher at 0.76 as trend following strategies tend to produce outsized returns in volatile markets.

There is also a strong correlation to USD fluctuation, and this is expected because most commodities are priced in USD.

Trend following strategies generally tend to underperform in periods where volatility is low and markets are directionless or going through whipsaws. To be honest we currently do not have a contingency for this. But there are plans to explore combining with mean reversion strategies to help reduce drawdowns during these periods.

Just want to share some of my opinions on volatility. It seems the hedge fund industry is becoming obsessed with volatility in meeting the demands of institutional or pension funds. Very often I see traders trying to lower volatility at the expense of adding more complexity to their systems.
I would like to think that there are investors who are willing to accomodate higher volatility as long as the strategy is capable of generating outsized returns across market cycles. Investors like Warren Buffett run even higher drawdowns of 50+%.
 
You must log in or register to reply here.
Back
Top