When it comes as of writing November 2016 – in terms of NinjaTrader 8, MicroTrends is very behind where i wanted us to be in migrating our technologies to NinjaTrader 8 from NinjaTrader 7. You will happy to know we are at 1000% effort and time on this now and very happy with the technology NT has done a superb job and the latest C# .net technology is a light year ahead now. WE will be releasing NinjaTrader 8 new and migrated components November onwards. The reason we are behind was due to a long project i will summarize if i can:
In 2015 to 2016 – my focus was elsewhere instead of MicroTrends public products and projects – I developed and traded automated trading systems supplied by MicroTrends algorithmic trading consultancy for a commercial hedge fund Clearpoint Capital (CPC) and was given a 5Mil cap to trade with and a risk mandate of 1%, to trade alongside their discretionary fundamental trading model – I was treated with amazing hospitality and every courtesy, expense and wish furnished… my job apart from enjoying Australia was to implement mechanical systems using the MTDS7 as the foundation with some strategies based on the MTDS USAR and the MTNTF7 Fib Grid systems the later. I could write a small book on the pains and successes.
I was based on site at the invitation of CPC and over time implemented some swing/ day trading algos which are proprietary to MicroTrends, we aimed to diversify across 20 futures instruments or so with different volumes, ranges, behaviour and seasonals… which resulted in a massive testing cycle of 3 months testing, visualisation sense checking, out of sample, walkforwards and realtime sim trading pre production – until going live, then ongoing stats and assessment against baseline -and variants of settings based on live trading feedback… trading 3 different timezones and requiring fault tolerance and redundancy, therefore in In addition and lots of infrastructure and data feed costs.
We ended up with viable automated systems on 6 markets live trading with 3 automated systems 2- systems but with parameter-set and time-series diversification -a total of 18 in parallel and that took 6months to go live, for compliance we had to follow a rigid set of plans before trading…and some systems were taken offline and replaced by evolved versions before we were stable, i enjoyed immensely using MicroTrends MTDS7 and NinjaTrader technology and enhancing it on the job as required. It was a fantastic journey with a lot of success algorithmic trading the FDAX which is beyond all doubt my favourite market….trading up to 8lots per trade entry certainly had its moments 🙂
We also placed discretionary trades based on the MicroTrends NinjaTrader Framework – Multi-time Frame Analysis – Fib Grid systems and had great success with those on markets screened for volatility and market structure- such as Silver, Gold, Coffee etc Our challenge with that as merely to trail the stop/brackets at the correct intervals… so automatic trade management was a key requirement missing for this components- if volatile moves occur you need to scale out/trail in to lock in profits to prevent giving it back on a reversal profit taking sell off or short covering rally -all those points are easily determined in advance -you just need to be there to move the stoploss brackets in.
Due to the time zone and style of trading implemented and the need to constantly diversify, i ending up working until 6am, 5 days per week and then on weekends preparing multiple systems, stats, backtests, real-time tests, live results and manage 3 live trading servers, 3 pre-production, 3 back test servers and so it was in the end a 7 day per week task – instead of 1 server with an easy to manage end of day trading system – this was not a tenable approach for me or anyone – so in August we switched off to change route and focus free from the overhead of trading and monitoring, we had great success with the algos… but our operational foot print was too big and our efficiency was too low…perhaps 66% instead of 85% as a goal…
Imagine the cost and maintaining 9 servers, backtest, pre production, and production servers as they used servers with different time zones had costs of data feeds + my office costs- + my retainer 😉 + my annual bonus… then 24/7 monitoring – which in the end made this approach not the best route. 1. we needed to scale to a wider and more manageable trading portfolio… 2. the basis of mistakes and inefficiencies in the limitations of the technology had cost around 50% of profits on some markets, according to a real-time baseline versus the real-time live results, crashes, stop/restarts/ downloading data crashes, or connections failing at the session open were the culprit…
This is the real difference between the hypothetical realtime sim world to the realtime live trading world – and it is not just the difference between getting in the cage and fighting a UFC champion or sitting at home and being a keyboard warrior – but more importantly the difference between the realized real world view and the hypothetical universe which is required as a guide and benchmark – the efficiency of live trading is the key to trading success.
So, on FDAX, FGBL, SPI, Nikkei where i was present and monitoring we hit it out the ball park or made a profit most weeks… on the other systems where i was not in direct attendance we routinely suffered opportunity costs -missed profits – technological issues prevented the mechanical fully automated trading systems working in a viable way due to inefficiency, the problem we were using a system created for day trading not swing trading. NinjaTrader 7 had some bad habits such as crashing when loading data so that other strategies were stopped and lost their context, working orders were cancelled and positions had to be flattened before you could restart a mechanical trading system, s0 we had to have separate servers to reduce the effects of errors, professional data feeds for each one. For me although trading was successful, I was simply spinning plates and juggling and it burned me out somewhat……we could not scale with this approach… fine for someone wanting to sit in front of a machine for 5 days per week – 12 to 16 hours per day – but even then limited to the amount of systems you can easily manage.
The problem stemmed from using NT7 strategies – these are really only suitable for attended day trading/swing overnight Sunday to Friday at maximum quite reliably but you need to make sure you are there to prepare data download,refresh and set it off for the open on Sunday evening – then you can trade all the way through with a NinjaTrader Continuum connection all the way to Friday Close, you need to close out Friday or do some special arranging using mutliple accounts with a fresh batch to start Sunday… Fantastic, but really i needed something to trade all week and start again the next week and not lose context of its positions and orders and to be able to reconnect and reload data… Sometimes a disconnect and reconnect caused havoc, cancelling working orders and leaving on unguarded positions, routine stuff… or a connection not waking up at the session open Sunday would miss a gold move – and that was the one needed to make the system viable……
So i needed to replace this solution to provide us with a swing trading/position trading capability – restart, stop/start, disconnect/reconnect – over a period of weeks/months on the same trade and not lose context…and to allow external adjustments of trades and orders and not go out of sync… CPC had no alternative but to as order me to investigate a multitude of alternative trading platforms: I investigated CQG API, TT XTrader API, Smart Quant, Bloomberg, Rithmic, Orc, Aparma, Modulus, AgenaTader, even KDB+ and a Fix Engine after exhaustive and non biased examination of all technologies and many dead ends… or time and budget restrictions, actually we ended up with NinjaTrader 8, nothing else ticked as many boxes as the only cost effective and technologically capable solution – which is good for me… I can migrate from NinjaTrader 7 to NinjaTrader 8… still using charting technical price and signals indicators and not have issues trying to port them to other lesser featured platforms.
Using NinjaTrader 8 we can remove the problems with NT7 crashing and not loading data properly for example – also connectivity in NT8 works better and i can make sure it starts on a Sunday open to trade gold and not miss the open for example…… but we have to write an extension a set of proprietary MicroTrends technologies to enable true swing/position trading capability in NinjaTrader.
So, what about the winning algos in NT7? Well they are great for a few markets but our mandate is the systems need to scale across 30 futures instruments not 6… which were medium high frequency trading systems- we wanted easy to manage low frequency systems – so then the only alternative I had to write an EOD system which allows us to scale to a large number of markets, so I can combine the winning algo pricing and rules, in as additional setup criteria when conditions are correct to use it in the system.
Why NinjaTrader 8 is the proven choice.
What we need is at a glance to see positions and entries across 30 markets+ as defined by the Fund Asset, stay in control with discretionary/automatic entry and automatic trade management with the ability to trade it all from a market scanner/chart view, stop/restart/ and not lose context of positions or working orders and externally adjust manually and the automatic system has no issues adjusting. We can do this with the market analyzer with modified columns and function, allowing us full control and even the ability to monitor connections, reconnect and even download data on a schedule, and the system all restarts and continues without a hitch… Our Efficiency is then where we need it -say 85% +… Operationally cost wise, we can use the minimum for data feeds and equipment, this technique is something NT8 cannot do out the box, it requires advanced specialist development skills MicroTrends has in house, we can leverage the NT8 open structure we can access the lower API levels directly and access and build a solid foundation for swing and position trading… incredible! Thank you NinjaTrader!
NinjaTrader EOD system swing trading/position trading automatic trade management, stop, restart continue…discretionary + automatic trade management or fully auto + automatic trade management.
So, to conclude now i completed this EOD system in November and we are in a test phase, CPC will forward test and roll it out in 2017 and our high frequency high maintenance trading is finished allowing me a normal life once more- i will migrate the EOD system to NinjaTrader 8 eventually and we plan to build similar systems for traders who are interested in End of day Trading systems for NinjaTrader.
As a result of this work we are happy to anounce MicroTrends has also planned a release of a similar product. If you interested in a swing/position trading which overcomes NinjaTrader strategy behaviours and allows control of a portfolio with automated or discretionary trade entry and automatic trade management please contact John at the Help Desk [email protected] who will provide details and make sure you are on the list for this limited sale product.
But for now, this means i am now back working full time on MicroTrends business – migration from NinjaTrader 7 to NinjaTrader 8!!! And will also be consulting and continuing with the above at a more sedate existence!!! So now you know why we are late with NinjaTrader 8!
NinjaTrader 7 & 8 Product Releases Coming Soon – Please join our mailing list for news announcements.