Fully automated futures trading

[stt]

Right now I'm using on my 400K account (running at 25% annualised risk target a year) about £102K in margin; so about 25.5% cash.

However this will vary depending on how close I am to HWM and signal strength. Right now I'm running at about 73% of average risk. At average risk then my margin usage would be around 35%.

To take a few other days as a sample

April 1st: 29.5%
Jan 1st: 30%

2014:
Sep 1st: 29.4%
June 1st: 37.8%

The numbers are perhaps less than you might expect (I have never been about 50%), perhaps because I won't always have positions in all 40 markets unless I have sufficiently strong signals in all of them, and also because margins aren't as chunky in say Eurodollar as they are in Emini's.

But it's reasonable to say I'll use between 20 and 50% of account value as margin, averaging 35%.

Thanks a lot for this. It gives a good idea on what margin (and thus capital requirements) one can expect. if max DD is 25% and margin requirement is 20-50%, this still gives 25% extra cash in the account.

Have a good day!!

 

[globalarbtrader]

Thanks a lot for this. It gives a good idea on what margin (and thus capital requirements) one can expect. if max DD is 25% and margin requirement is 20-50%, this still gives 25% extra cash in the account.

Have a good day!!

I think in backtest my max DD is around 37%. This is probably optimistic, but on the other hand it doesn't allow for kelly reduction in capital as DD happens. Also reduction in risk as DD happens also reduces margin. So if you're paying 50% margin, and you lose half; by then you'd only need half your original margin (25% of capital).

 

[stt]

I think in backtest my max DD is around 37%. This is probably optimistic, but on the other hand it doesn't allow for kelly reduction in capital as DD happens. Also reduction in risk as DD happens also reduces margin. So if you're paying 50% margin, and you lose half; by then you'd only need half your original margin (25% of capital).

True. position sizing based on equity value (kelly's way) reduces the need. and in theory makes it impossible to go bust (asymptotically goes to 0 but never 0). Thanks for the insight into margin needs.

One more question - since you are doing it outside large organization, what data sources are you able to use in your strategies. At ManAHL, you probably had lot of other data in addition to price and volume (say weather, supply, demand metrics, forecasts etc in case of different commodities).

In your own trading, what data do you use apart from prices and volumes.

Thanks again.

 

[globalarbtrader]

True. position sizing based on equity value (kelly's way) reduces the need. and in theory makes it impossible to go bust (asymptotically goes to 0 but never 0). Thanks for the insight into margin needs.

One more question - since you are doing it outside large organization, what data sources are you able to use in your strategies. At ManAHL, you probably had lot of other data in addition to price and volume (say weather, supply, demand metrics, forecasts etc in case of different commodities).

In your own trading, what data do you use apart from prices and volumes.

Thanks again.

I don't. I just use prices.

This is a time related decision, rather than a 'I can't afford the data'. Theres plenty of free data out there.

I like having a system I spend only a few minutes a day looking after my system (on average; rolls can be a bit hectic, sometimes I have to spend half a day with a new feature or very occasional bug). Time spent on price data filtering related issues is maybe half of that.

With 45 markets, lets say I average 100 minutes a month doing on data filtering on my system, or 4 minutes per working day. If I added say 10 new data fields I'd be spending 40 minutes a day. Actually that might be an underestimate because system complexity also increases exponentially with the number of data fields, so I'd probably be spending an hour or more.

I'd rather not.

At AHL we had teams of people working 24 hour shifts across 2 sites worrying about this stuff. Now its just me.

 

[stt]

I don't. I just use prices.

This is a time related decision, rather than a 'I can't afford the data'. Theres plenty of free data out there.

I like having a system I spend only a few minutes a day looking after my system (on average; rolls can be a bit hectic, sometimes I have to spend half a day with a new feature or very occasional bug). Time spent on price data filtering related issues is maybe half of that.

With 45 markets, lets say I average 100 minutes a month doing on data filtering on my system, or 4 minutes per working day. If I added say 10 new data fields I'd be spending 40 minutes a day. Actually that might be an underestimate because system complexity also increases exponentially with the number of data fields, so I'd probably be spending an hour or more.

I'd rather not.

At AHL we had teams of people working 24 hour shifts across 2 sites worrying about this stuff. Now its just me.

That's a very good argument. With more data you certainly have to put more time in making sure it is clean/accurate/adjusted etc.

I guess if the extra data gives you benefit in overall sharpe/reducing drawdown etc, some may argue it is worth spending that time (or paying someone else to do it). But again that depends on what you want to do while trading this. If you want to minimize time spend in running the strategy, using just the price is best option.

Large organization have more people but at the same time inefficiency creeps in so what a team of 5 people doing there full time, you might be able to do it alone in few hours a day (specially since you are skilled in programming).

Thanks again for the insight. It is very useful to know from real trader what he or she thinks important.

 

[globalarbtrader]

New blog post:

http://qoppac.blogspot.co.uk/2015/06/systems-building-deciding-positions.html

 

[OddTrader]

Good to have the "hedging" stuff! Just don't know how much it would cost and how to do it.

?

http://www.elitetrader.com/et/index.php?threads/ideas-for-a-conditional-hedge.289494/page-2

Sounds like a few things being confounded here. What do you want? What problem are you trying to fix? If I understand correctly you've got something like a 130:30 portfolio, which will be making its money from relative value, a long bias to equities and a variable beta to equities.

a) If I have a strategy that is mostly long, but I don't want any equity exposure on average (maybe you've too much equity exposure elsewhere), then I should sell a constant hedge equal to my average exposure.

b) If I don't think the overall long/short timing of the portfolio is a profitable signal (you can check this) then I should sell a hedge which varies according to that exposure, or tweak the strategy so its always equity neutral.

c) I could add a separate signal to time the overall long/short of the index, to replace the natural timing of the underlying strategy. For example I could use a trend following signal on the index.

Now how do we hedge in a and b? We can do a linear hedge (sell a future), or a non linear hedge: an options hedge or a synthetic options hedge. A short only trend following signal will give you a synthetic long put (a symmetric signal is a synthetic long straddle). If implied volatility is low then it makes more sense to buy options. If implied volatility is high then it might make more sense to do the hedge with the synthetic option. Note that the trend following signal can be both an additional strategy, and a way of hedging.

 

[globalarbtrader]

?

http://www.elitetrader.com/et/index.php?threads/ideas-for-a-conditional-hedge.289494/page-2

The hedge I have in this account is much simpler than in that thread.

Basically I started my account up with 300K at risk; but I didn't put 300K of cash into the account, as to realise that would have involved selling some stock and paying capital gains tax. Instead I funded the account partly with stock and partly with cash.

Let's say I started with 200K in stock and 100K of cash, and I needed 150K to cover margin (exact figures not important). So I'm borrowing 50K against my stock to cover margin. If I make money (as I have done) my borrowing would go down.

But ideally I want the p&l coming from my account to be purely from futures trading, rather than beta exposure.

I don't expect the hedge to make money, though in the long run it should make some if the hedge was perfect. This is because:

- I'm receiving say 4.5% yield on my UK stocks
- Through the short FTSE futures hedge effectively paying the yield on the index which is lower
- I'm also effectively receiving LIBOR through the future
- I'll pay a little LIBOR on my margin borrowing; though that arguably belongs in the pure futures trading mental account

There is also currency risk, since some of my stocks are European. This is effectively unhedged - the future only hedges the delta. If say the euro depreciates 10% on a £200K european equity; then I lose £20K. If at the same time the index doesn't move then I will make nothing on my hedge.

As well as alpha risk and currency risk the beta hedge won't be perfect due to the large size of the future meaning I can't get the hedge ratio spot on.

So part of my portfolio is like an unleveraged market neutral hedge fund.

FYI last fiscal year I lost 2.3% on my hedging
http://qoppac.blogspot.co.uk/2015/04/futures-trading-performance-year-one.html

GAT

 

[OddTrader]

The hedge I have in this account is much simpler than in that thread.

Basically I started my account up with 300K at risk; but I didn't put 300K of cash into the account, as to realise that would have involved selling some stock and paying capital gains tax. Instead I funded the account partly with stock and partly with cash.

Let's say I started with 200K in stock and 100K of cash, and I needed 150K to cover margin (exact figures not important). So I'm borrowing 50K against my stock to cover margin. If I make money (as I have done) my borrowing would go down.

But ideally I want the p&l coming from my account to be purely from futures trading, rather than beta exposure.

I don't expect the hedge to make money, though in the long run it should make some if the hedge was perfect. This is because:

- I'm receiving say 4.5% yield on my UK stocks
- Through the short FTSE futures hedge effectively paying the yield on the index which is lower
- I'm also effectively receiving LIBOR through the future
- I'll pay a little LIBOR on my margin borrowing; though that arguably belongs in the pure futures trading mental account

There is also currency risk, since some of my stocks are European. This is effectively unhedged - the future only hedges the delta. If say the euro depreciates 10% on a £200K european equity; then I lose £20K. If at the same time the index doesn't move then I will make nothing on my hedge.

As well as alpha risk and currency risk the beta hedge won't be perfect due to the large size of the future meaning I can't get the hedge ratio spot on.

So part of my portfolio is like an unleveraged market neutral hedge fund.

FYI last fiscal year I lost 2.3% on my hedging
http://qoppac.blogspot.co.uk/2015/04/futures-trading-performance-year-one.html

GAT

Thanks GAT for your response. Here are my feedback.

I think you have an excellent blog, a profile of outstanding background, a range of diversified knowledge and skills, and a keenness wanting to share with us on ET, which is much appreciated.

However, your current threads suddenly offer us too much information covering too many important topics in a short time, that would easily cause some unwanted arguments with us on ET, perhaps a culture issue on ET.

My suggestion is perhaps you would focus on just a few most important topics to start with. Then people on ET could comprehend and discuss effectively the topics with you in a more productive manner.

Just my 2 cents!

 

[globalarbtrader]

Thanks GAT for your response. Here are my feedback.

I think you have an excellent blog, a profile of outstanding background, a range of diversified knowledge and skills, and a keenness wanting to share with us on ET, which is much appreciated.

However, your current threads suddenly offer us too much information covering too many important topics in a short time, that would easily cause some unwanted arguments with us on ET, perhaps a culture issue on ET.

My suggestion is perhaps you would focus on just a few most important topics to start with. Then people on ET could comprehend and discuss effectively the topics with you in a more productive manner.

Just my 2 cents!

Okay pick a topic....