Anyone know any good recources for doing all this in C++. I use a very specifc data feed in a rather unconventional way and have to code in C++. Im stacking way too much math in such a complex manor…i have around 25 calcuations on 5 different time horizons. Standard devtaion/slope rate coefients coverted to %/Stocahstic modeling with slope rate on 3 time horizons cross treaded to GDP/CPI in corss differetials/ A complex order flow agasint liquidity and volitilty optimization threasholds…on 3 time horizons side by side…But also I work from a replay envooncment rather then a static data set casting thread…its just more accurate to matching live makret conditions i have found over casting threads on aggregated data…Hence Python doesn work at least not in how i am dealing with data as I work with deltix and sierra charts both of which have subalyers of C++ that make it much easier…as well as i have to be able to control and adjust all the math anyway and rebuild all my calcuations to very specific calcuations a bit outside the norm. Learning to code indicators in C++ will certainly not get me there but it will be faster then coding them in python and figureing out C++ from there? Just wondering if anyone knows of any resources as Im 15 years into trading…5 years into algo but olny 2 months into engineerring and having trouble steering this to C++??
Hello Philip,
Since you want to learn to code in c++ for your trading system, you might find the following resources useful.
- Introduction to C++ for Financial Engineers: An Object-Oriented Approach - Daniel J. Duffy link
- C++ for Python Programmers link
- Financial Numerical Recipes in C++ link
- Building Automated Trading Systems link
- An Example Of A Trading Strategy Coded In C++ link
- Thinking In C++ link
Hope this helps!
Thank you
These are great resources thank you? Do you know of any people teaching this or places to interact with humans? Or a bootcamp specific to this field…in C++? I work for an investment firm who is paying for me to learn the engineering specific to my aglo. The books are super slow! I have a time limit and budget and it would go much faster with some human interaction? Everyone hangs out in python in this arena and I know a bit of python and feel I need more complex engineering then static data mining cast process. Certainly I can not use public data at the level that I am trying to trade at. I need to navigate things like deltix or build an GUI from CME or such…this whole historical data does not match or dont expect yesterday to not be today or tomorrow. This is issue needs to be dealt with! And I see how to go about it…but the engineering is complex.
In my opinion this lack of accuracy is not a result of market condition change or regime change or the heterogeneous nature of decentralization markets as in forex… Change is the very thing we are attempting to quantify…so to say because of change is the reason its not….something is off here….and from my years of working with folks building the self driving cars…Data mining is one of the main causes for this inaccuracy and I feel there needs to be more work in building a replay model to match a live market feed.
In the autonomous car community these guys talk about this issue with data as well. The static cast is very inaccurate. So they build three dimensional models and work form a replay….I have also heard some data scientist out of Finland working in this Research as well in how replay is more accurate than data mining. This is issue one. Two is you need to be able to study the math from the interface. Automated stochastic, slope rate stacked on the automation, separation value……on a 240 time frame, day and week. At 3 deviations from the mean….ATR set to a slope rate deviation on a scale of day over week on 3 time frames and stacking all this into a relationship derivative….when you set your bar linking to sync and you pull back to your loss and hit replay….there is a much clearer picture as to what’s going on mathematically. ITs easier to see see what the math is doing then just guessing how a math formula will perform on some numbers……So I need a deeper study than data casting?
Ill get there eventually…but if there is some people doing this on a more complex manor. It would go much faster??
Hello Philip,
Sorry to say that we don't have any C++ course recommendations for your requirement.
Thank you