Fxx's macro algo adventure

[FXX]

Introducing the USD-IDX and the AUD-IDX. The first image shows the USD-IDX overlayed with the USD and the second image is the USD-IDX overlayed with the AUD-IDX. Next steps are to work through each economy as I now have the tooling to rapidly build and assess the correlations. The image with the indexes overlayed is going to be the main view I will use for trading. I also need to now pay for fx data where i can overlay these with currency pairs. Right now I have a 20-year dump of fx data but i want to be able to chart it in realtime. Luckily I found a source for relatively cheap. Total costs for data is about £300 a month and when i setup the azure architecture we probably looking at 400 a month - It will no doubt pay for itself.

Still, early days but i am closer to having a solid tool for trading allowing me to ease off on the volume of research i do and move more towards automated trading.

 

[FXX]

small update

I have added Vix to my core dataset and busy working on incorporating it into each index. How is this going to be done is subject to some RnD work. my initial thoughts are to take the standard deviation and extract the difference with the latest daily print.

This will then serve as a weighting factor which will fill in the gap between economic data and market sentiment. My first pass will be to apply this weighting factor and negativity impact risk currencies like AUD, GBP, ZAR etc.

Now this does open up a potential problem as i will be assuming increased volatility will lead to dollar strength. If 20 years of testing shows this to be a problem, then I'll have to incorporate news flow (just headlines from sites like dailyfx or forex live) where I can correlated mentions of currencies\pairs in either a positive or negative way and use this to apply the weighting to relavant currencies.

I actually have a rock solid way of working textual data like this as I have spent many months working on doing exactly this over central bank minutes. So I am not overly worried about implementing this.

I do have something to show and you can visualise how this would work for the index.

here is AUDUSD overlayed with the vix. those areas of spikes are in fact areas where the algo index loses correlation. By incorporating this I can fill that gap.

 

[Trader333]

So strength in Vix appears to correlate with USD strength in the above case, so is that also the same for other currencies ?

 

[FXX]

So strength in Vix appears to correlate with USD strength in the above case, so is that also the same for other currencies ?

With dollar demand, any currency paired with the dollar will weaken. If you think about it, Vix measures high volatility in dollars because its scoped at price fluctuations against the s&p 500.

I recognised on my indices that I lose what is actually very good correlation for periods. I knew it was related to market sentiment. While the vix won't reflect strength sentiment in pairs like eurcad, it certainly covers all dollar pairs.

I will need to see how I can fill the gap with other pairs and I might have to access other volatility indices locally or build my own - lucky I have worked at an index provider so know all about the process of building these, just with my own flavour.

 

[Trader333]

While the vix won't reflect strength sentiment in pairs like eurcad, it certainly covers all dollar pairs.

Yes that makes perfect sense and if you are able to make use of it for USD related pairs then is there really a need to look into volatility of non USD pairs ? After all if your correlations work for the USD would that not give you more than enough profit without having to do the same for other currencies which may not have the data as readily available ?

 

[FXX]

Yes that makes perfect sense and if you are able to make use of it for USD related pairs then is there really a need to look into volatility of non USD pairs ? After all if your correlations work for the USD would that not give you more than enough profit without having to do the same for other currencies which may not have the data as readily available ?

Yes and no. Yes to provide enough trading opportunity with dollar related pairs. No because I am doing all this research for more than my own trading advantage. Having spent 4 years working at an index provider, I love the business model and believe I have a gap in the market because none of them give a crap about fx.

 

[FXX]

Well in less than 2 months over 41 trades with 1 loss on a test account dedicated to this work, I have made 103%. Still early days and lots of work ahead but not a bad result. Yes leveraged to the brim but I am not looking at that, I am focused on the win loss ratio and risk. All I have done is traded pullback opportunities (very simple) in the direction of the index.

 

[FXX]

I have started work on the Canada index and wanted to share one of the components to show to everyone the importance of paying attention to the fundamentals of an economy, and for Canada it includes oil.

 

[FXX]

FYI no updates on this, I have been working on a fixed income project for a client so everything is currently on hold for a bit. I have still been testing what I have built so far and the account is up over 200% ( 55 trades and 1 loser)

 

[Trader333]

..........the account is up over 200% ( 55 trades and 1 loser)

Impressive , well done.

 

[FXX]

Impressive , well done.

thanks, I am not going to get overly excited about it until I get a good sample size behind me. My next step is to get the other indices done and have all the updates automated so I can use this in trading analysis regularly without needing to worry about updating stuff. I reckon I'll make myself an azure website to host this all so I can access it on any device. Much still to do just time is a commodity i have little of.

 

[FXX]

update

While I haven't been working on the project, I have been poking around the data and specifically the data vendors I am using and the data quality. Unfortunately there are some rather dodgy chances selling data with absolutely no clue on how to manage their data. Having to constantly email them about data issues such as data series that haven't been updated, which in some cases are years out of date.

So I am tired of having to deal with morons and have decided to build an interface for Fred, World Bank, Other central banks, and whatever sources I can get. if I am to build these indices I must have a consistent reliable data source. Premium feeds consist of indicators from these sources - you are paying for convenience which unfortunately doesn't guarantee quality.

The very expensive feeds have proprietary data and these can be several thousand dollars a month - which is cheap compared to some of my clients who pay hundreds of thousands a year for feeds from refinitive for example. I have the skills to build this but it will set me back a few weeks from being able to carry on with the index work.

it sucks but I guess Rome wasn't built in a day and certainly wasn't without hurdles. It sucks even more that I am so busy that I haven't had time for this.

on a side note, and yes this is rather ridiculous. The test account which I have been using as part of this index work based on what I currently have, is rather surprising. It is up 320 percent and I am not sure how I feel about that. On one hand I know it is entirely possible that this is just a winning streak. On the other I have 20 years of close correlation evidence when viewing the index against price data. 64 trades with 1 loser... yeah its a bit ridiculous which is why I am not going to think much of it and focus on getting the work done so I can get back to building this out for about 8 economies which should give me more opportunities and focus on diverging indices.

 

[Trader333]

...It is up 320 percent and I am not sure how I feel about that. On one hand I know it is entirely possible that this is just a winning streak.

Unlikely in my view that this is just winning streak based on the thoroughness with which you have apporached the whole project. As you said, there are always hurdles to deal with but then if that wasn't the case everyone would be able to do the same which would probably result in having no competitive advanatge.

 

[FXX]

Unlikely in my view that this is just winning streak based on the thoroughness with which you have apporached the whole project. As you said, there are always hurdles to deal with but then if that wasn't the case everyone would be able to do the same which would probably result in having no competitive advanatge.

well the account is now on 700% gain and still only 1 loss. Not sure how this is even possible. I have a 75 percent win rate with the regular strategy. I have never experienced anything like this before and therefore I am still reluctant to give it any weight. What I can say is the account has 95 trades since 19th of October.

 

[FXX]

this is what i have been testing to give you an idea

1) trade in the direction of the index
2) trade the setup shown

 

[FXX]

here is the ausi index - still being worked on but so far doing a good job

 

[FXX]

Update

This is an intraday trade I closed at midday. The testing has involved intraday and swing trading. The setup is simple and the workflow is

1) check upcoming news
2) find divergences in indices
3) find setup for those crosses with diverging indices
4) trade

I am going to complete some more indices and begin adding this strategy to my daily trading on a live account.

View attachment trade.png

 

[Trader333]

There is a "general view" that intraday trading is less profitable than swing or position trading so I am curious to know if you are finding this to be the case in the approach your are using ?

 

[FXX]

There is a "general view" that intraday trading is less profitable than swing or position trading so I am curious to know if you are finding this to be the case in the approach your are using ?

I have been trading fundamentals intraday for a long time. I guess its more closely aligned to sentiment trading with economic data and macro news driving trades with ongoing analysis.

As to the profability difference, since the blonde bimbo was elected US president, lots of long term macro guys have been burned. Corvid19 is another issue for macro guys. I don't have any comparison long term these days in terms of profitability.

I haven't traded long term positions since about 2010, and then I found it slow and difficult with space you need to give trades. I like to have risk off the table over weekends and holidays.

are you a long term trader?