Noosapalooza 2019 & System idea generation

[Anonymous]

Hi all, with Noosapalooza 2019 just around the corner, I thought I would put my hand up and raise a question/thought/idea that one particular thing I am hoping to get from it is generating ideas and thoughts on systems I can use in times of volatility, sideways, choppy markets etc.

I am currently running a reversion strategy using Bollinger bands for a break below the bottom band and an ATR stretch for the buy combined with a MOC exit. It is an intraday system. So far it is going ‘OK’ but only OK, not amazing our ground breaking. A challenge I am having with this system is that luckily I have a decent sized account and am also able to access margin, and in combination with the number of maximum positions that works well with this system I am having to lay some orders with massive volumes that just cannot be executed within seconds (I’m talking needing to order 40 to 50,000 units in some cases) and this just can’t be done in seconds, it usually takes a few minutes at best and therefore the backtesting/optimisation results in Amibroker are not exactly accurate because if prices touch a low even for one second in the day it assumes you get the complete fill. My Trade Size Limit is set to 1% of daily volume, so I am not concerned about swamping the days’ volume with my order.

So if we cross paths at Noosapalooza this is the first topic I’ll probably raise as an idea and pick your brains for any ways to tackle it, or, of course, come up with some completely different ideas to run in such market conditions.

Oh and also for the same conditions in the Australian markets too. Right now my intraday reversion system I use when the market is sideways on the US markets only. It would be nice to try and get a second use of the cash in every 24 hour period by doing something intraday on the Australian market so I can hopefully make a few dollars in the overnight session US and then again on the Australian market.

My Momentum/Trend systems I use for when the market is trending/bulling etc is fine, I don’t think I need any help there but am always up for suggestions.

Do any of you have any ideas to share? Will any of you be attending Noosapalooza this year for us to share ideas?

See you there!,
Matt

[JulianCohen]

I’m not going to Noosa unfortunately but I had exactly the same ideas as you with utilisation of funds during the Aussie day. I couldn’t make an MOC work on the Aussie markets in anything like the size you are thinking of as the fills were never there. You just cannot get set at your price. I gave up with it a long time ago, as did Nick I believe. I tried HK, Tokyo and London as well but for various reasons those wouldn’t work for me either. You are in China I believe, is that correct? If so if you have access to people that know about the HK market then it might be possible to do something there.

My issue with HK was the stock market is set up in lot sizes and I couldn’t find access to someone with the information I needed to code up a system correctly. I’d be happy to talk to you to discuss it further, as would Scott I would expect. But if you are looking to “swing a big line” then Australia is not the place.

[Anonymous]

Thanks for the tips Julian.

Actually I am living back in Australia now. I was able to move back here in October (but yes, was previously in China).

If I can’t make the Australian idea work intraday, I guess I’ll have to stick to US.

How about trading currencies during the Aus day?

[JulianCohen]

That’s a whole different can of worms A huge rabbit hole to get lost down for a while….

One thing I didn’t try on the Aussie was just sticking to the ASX 100, or maybe ASX 50; but that was only because I couldn’t make a system work on that universe.

The main issue I had with the Aussie, that some others have had as well, is getting the fills. The lack of fills degraded the system performance, plus the brokerage is higher. Now I know that if you have a backtest that produces a comfortable level of return with the brokerage at the higher level, then that should be fine, but when you start getting missed fills everyday you end up begrudging the brokerage on the ones that do get filled. I found I was paying the same in bro as I was making in profit after a couple of months, so I packed it in. If I found a system that worked on the absolute most liquid stocks, then it might be worth trying.

[Nick Radge]

Matt,
Have you explored the Accumulate/Distribute order?

https://www.interactivebrokers.com/en/index.php?f=1223

[Anonymous]

Nick, no I haven’t tried this order type. I will have a look at it.

My primary issue is the speed at which a large block of volume can be executed to catch the prevailing price, and have that match well with the backtesting results I am seeing in Amibroker. The unfortunate limitation of dealing with just daily data in Amibroker is that if parameters/optimisation are set where a price may have hit my targets/limits for just a very small period of time (even fractions of a second) then Amibroker will assume the fill is had in full. It would certainly be nice to be able to backtest with minute data with Amibroker time compression into daily data yet still have a minute based filter of sorts such as price needing to be below the limit price for at least say, 2 or 3 continuous minutes on the minute data time scale and only then would conditions be met for a fill. Still, that would require minute data and the last batch of minute data I paid $1100 USD for from QuantQuote was, quite frankly, complete garbage full of errors. I know Amibroker can do the time compression thing from minute to day and vice versa so I am not worried about Amibroker having the skills to pay the bills. It’s just the data source which I would need to find.

I am not that concerned about the size of my order affecting the market. I am using a volume filter and also set my trade size limit to 1%. I wouldn’t assume therefore that my trades are affecting the price too much.

With the Accumulate/Distribute order I get the feeling, if anything, that the execution time to complete the order will be longer than shorter and it’s primary focus is on not affecting the market price with large orders by distributing them into other random packets/quantities over a period of time. Still, I don’t know if I don’t try, so thanks for the suggestion!

[Nick Radge]

The best intra day data is via https://www.tickdata.com/ They do ASX as well.

[Stephen James]

My understanding is that AB can only compress time frames from lower to higher (e.g. daily to weekly), not higher to lower.

[Anonymous]

I always thought as long as you have a database in the lowest order time frame you want to use, you can use the TimeFrameCompress and TimeFrameExpand functions to pop between the time frames you like for any relevant calculations/uses. E.g. if your database is stored as minute data then you can use minute, hourly, daily, weekly as much as you like. But sure, agree it would be impossible to jump from daily to minute if you are only running on a daily database.

Thanks for the tip on Tickdata Nick. Will check it out.

[Anonymous]

Hi Nick, thanks for the suggestion for the accumulation/distribution order. Do you know if this can be used the Smart API or would this mess with the functionality of the API? Some interesting features to think about with the execution algo. Thanks!

[Nick Radge]

Dustin,
I’ve not investigated the order type specifically. Depending on its format it may be easily used with the API in its current form. I’m neck deep into getting ready for Noosapalooza so can’t investigate at the moment.

[Author]

Posts