Julian’s Journal

[JulianCohen]

Mar ’18

US MR: -0.3%
US MOC: 0.9%

S&P 500 Momentum: -4.2%
NASDAQ Momentum: 2.7%

ETF Momentum: -0.01%

Total Account: -2.8%

[JulianCohen]

Apr ’18

US MR: 1.2%
US MOC: 2.4%

S&P 500 Momentum: 0.4%
NASDAQ Momentum: 1.4%
ASX Momentum -5.4%

Total Account: 0.01%

[JulianCohen]

I have been thinking about institutional risk over the last few weeks. Basically I don’t like having a lot of money with one broker or bank, I like to spread the risk. So I looked into TDAmeritrade and I’m setting up an account with them to hold my US rotational and trend following systems. They have a cheap commission rate and I’ll feel better about having some of my funds with a different broker.

I am based in Singapore so I can access them but I have a feeling that Australian residents can’t.

[SaidBitar]

Honestly it is not bad idea, I use two brokers as well.

[Nick Radge]

I have two brokers; Macquarie in Australia where I keep my SMSF and all other trading at IB.

I did move the funds from IB to Macquarie during the GFC for that exact reason.

[RobGiles]

I have all my ASX exposure with Commsec (online broker in Australia) and IB for US rotational and MOC systems.

[JulianCohen]

I just tested an idea I had and wanted some thoughts.

I wanted to see how my MOC system performed under different volatility situations, so I took the high and low of the VIX for the last 100 days and split that into 4 sections, then tested my system against each section to see how it performed.

The highest volatility quartile in my case, produced 9% CAGR and -20% MDD so I tested the whole system without that volatility environment and there was a slightly lower CAGR but a great improvement in MDD, as you would expect.

I was pretty happy with myself until I tested against this YTD, and then I found that the system would be flat for this year, whereas my existing system is up 25%.

I’m not quite sure what to take away from this. Long term the “improved volatility system” has better metrics. My guess though would be many of us might have the same result for this year if the highest amount of volatility was removed.

I also tested against ATR of individual stocks as opposed to the market environment as a whole, but strangely or not, that had little effect. Each quartile of volatility had good results. Of course my code could be wrong….there’s always that

[ScottMcNab]

One way to approach it is to consider whether the maxDD was a concern for you prior to this study? It not then you may have added a layer of complexity to the system when it may not really have be necessary (Occam’s razor etc).

On another note, I think the difference in returns for 2018 is very interesting and your research may form the basis of another system that has low correlation to your current one.

[JulianCohen]

Scott McNab wrote:

One way to approach it is to consider whether the maxDD was a concern for you prior to this study? It not then you may have added a layer of complexity to the system when it may not really have be necessary (Occam’s razor etc).

No it wasn’t particularly but I was very happy to reduce it with no cost to the CAGR, but I agree with you completely about Occam’s Razor. Sometimes, make that most times, we are trying to perfect something that doesn’t need fiddling with.

Scott McNab wrote:

On another note, I think the difference in returns for 2018 is very interesting and your research may form the basis of another system that has low correlation to your current one.

Another bloody rabbit warren to run down!

[JulianCohen]

Scott McNab wrote:

One way to approach it is to consider whether the maxDD was a concern for you prior to this study? It not then you may have added a layer of complexity to the system when it may not really have be necessary (Occam’s razor etc).

On another note, I think the difference in returns for 2018 is very interesting and your research may form the basis of another system that has low correlation to your current one.

So I’ve been thinking about this some more.

I believe trying to cut back on the MDD can have negative and unforeseen implications. Having the movement in Feb in the backtests is very important going forwards as we had a day where there was a big drawdown, followed by an even bigger recovery day. If you try to prevent the big drawdowns by writing a system that avoids them, then you also skip the big recoveries and that is the reason there is a big difference in the two systems performances.

In other words, you have to have the drawdowns in order to get the returns. I believe that is what Jerry Parker means when he says “you have to learn to love your drawdowns”. I finally think I understand that now.

That’s the thing about reading about this business. You can read something, and think you understand it, and then ten years later you discover something that makes you understand it again to a level you realise you hadn’t before.

Sorry that’s a bit profound.

[Nick Radge]

Its important to not confuse volatility with risk. The Feb move wasn’t risk per se – it was volatility.

Volatility is required to generate positive returns. Think about a bank account; no volatility, no return.

Obviously the far end of the equation is risk, specifically risk of ruin usually from volatility coupled with extreme leverage.

[TrentRothall]

I agree about the loving DD Juian, especially like Nick said When volatility is present and causes ‘short term DD’ or account volatility. Those periods are often followed by large gains in the system. I am talking from a MR perspective mainly.

[JulianCohen]

Apr ’18

US MR: -0.2%
US MOC: 4.6%

S&P 500 Momentum: 0.2%
NASDAQ Momentum: -2.12%
ASX Momentum 0.74%
ASX Growth 2.01%

Total Account: 2.32%

I’m happy with this month. Sorted the SB problem now so I have added to my MR and MOC systems. I also switched all my Momentum systems to a different broker, or I will by close of business tonight I hope

[Anonymous]

Julian,

Would you be prepared to share your findings with moving to the other broker?

I am also in the same boat that I am feeling a little uncomfortable in keeping all the eggs in one basket at one broker and for the systems that have relatively low frequency and can otherwise be managed through a simple web based portal then this might be a good option (E.g. 10 to 20 trades a month on my trend following system can easily enough be done by hand in probably 20 minutes tops once a month through something simple web based).

I think on a previous post you said TD Ameritrade. Is this the direction you ended up going? Any other findings/musings to share?

Thanks,
Matt

[JulianCohen]

Matthew O’Keefe wrote:

Julian,

Would you be prepared to share your findings with moving to the other broker?

Yes Matt I went with TD Ameritrade. I am based in Singapore so I chose them, but they don’t operate in Australia. I chose them to hold any long term assets, such as my WTT system and Momentum systems. The first two months of trades were brokerage free, so it allowed me to set up all the systems and place all the trades for free. As you know these systems are relatively slow on turnover so although their brokerage is twice IB, I don’t mind holding the stocks there on slow turnover systems as a form of diversification.

That’s about it really. Quite easy to set up and run. Although I am using it for slow turnover stocks, brokerage charges were a consideration, as anything more than .001 will start to eat into your performance over the years. My bank tried to convince me to hold stocks with them but they charge 0.5 and I told them that takes my performance from 23% CAGR to 4% CAGR. I might as well buy a bond and not bother trading for that result. Of course they said they could help me with that solution too!!

[Author]

Posts