Procedural languages: (Wikipedia) most Scripting languages e.g. EasyLanguage Object Oriented languages: (Wikipedia) C, C++, C#, Java, Python and Ruby <object width="425" height="344"><param name="movie" value="http://www.youtube.com/v/4_Oj4xcD52A&hl=en&fs=1"></param><param name="allowFullScreen" value="true"></param><embed src="http://www.youtube.com/v/4_Oj4xcD52A&hl=en&fs=1" type="application/x-shockwave-flash" allowfullscreen="true" width="425" height="344"></embed></object> If you know the difference between the two, you will have no problem choosing which one to use. If you don't know the difference... stay with a scripting language.
Iâm sorry you and others failed to grasp my analogy in the comparison between Ninja Trader and Tradestation. I was not singling out Ninja Trader or Tradestation for criticism or praise. Nor was I saying the Ninja Trader cannot do a Set Trail Stop. I was not trying to draw a defensive posture from Ninja Trader advocates. I have Ninja Trader on my PC and I am starting to code applications. What I am saying is Tradestation is a much more mature application than Ninja Trader. Many of the Tradestation 1000s applications, functions and strategies have had the âbugsâ worked out of them. The Set Trail Stop has been in place 10 for more years from 2000I days. That is what I was attempting to address. Today I can find hundreds of âcannedâ Tradestation trading applications that make money. Ninja Trader is just beginning to show some promise in fulfilling this expectation. Tradestation has had years to try and get its platform optimized for efficient transaction processing (still has a way to go). Ninja Trader will probably go through many releases before it gains in efficiency (this from a Ninja Trader rater âReliability: F (crashes all the damn time)â. Ninja Trader will have to gain a wider acceptance from the trading community by producing trading applications that make profits over the long term. When I can pick up the TASC magazine and see Ninja Trader programming applications in articles with detail flow charts, then it will have arrived. Tums made a excellent point. Many traders will have to learn Object Oriented programming. Since I first learned it in the 1990s it has come a long way. The bottom line is Ninja Trader has many excellent features and will be an excellent platform in the long run. When I can hook up my DB2 or Oracle databases to Ninja Trader Data warehouse and perform data mining I will be thrilled.
I'd like to chime in and say that I agree completely. Each tool has a purpose and audience who will benefit most from it. Valid points, of course. At the same time, this is the "automated trading forum" so the assumption is that the O.P. was asking about writing code for an automated trading system. Neither TradeStation nor NinjaTrader are "mature" in that sense, in my opinion for algorithmic trading. They're both okay for historical testing as long as you don't need to use tick data but most automated traders need tick data testing. Unfortunately, there doesn't at present exist any system which combines all the qualities you describe and smoothly supports deploying strategies for fully automated trading. That forces most truly automated traders to build their own custom platform. I did the same. Very painful. I got mine running after a gargantuan effort of programming. That's interesting. Do you know anyone that makes a living automated trading "canned" applications? Never heard of it myself. It seems the code in magazines is great for learning but always has holes that need filling to truly be used for algorithmic trading. Again, are you talking about automated trading? NinjaTrader seems to excel at discretionary trading. I love NT's ability to draw a line on screen and a trade automatically gets entered if price crosses the line. But after trying to use it for automated trading, I found that it has some serious shortcomings. That's valuable. But I feel a system can be built which a sufficient framework so that it doesn't require so much OO programming to build even complex models. I recently released the one I built (TickZOOM) which solves many of these problems. But it's very new, of course. Unfortunately, as a software architect, it became clear that NinjaTrader would have to start over from scratch to build sufficient performance and automated deployment into the system. Same goes for TS. TradeStation appears to be originally engineered for historical testing of bar data. Everything after that like real time trading or automated trading was an afterthought and works less than optimally. For that reason, those performance features were built into TZ first instead of fancy GUI features. A GUI can always be built around a screaming fast and reliable engine. Also, a lot of other stuff, like unlimited multi time frames, portfolio trading and stats, etc are all built into TZ that are missing in the others. However, TickZOOM does nothing for discretionary traders except maybe for historical testing. So you see. The point of my message is that every tool has a "niche" where it excels. And unfortunately, software must be built for that specific niche from the beginning. If not, it's near impossible once you have many users or customers to change the design. For beginners though, I suggest learning EasyLanguage first for 2 reason. 1. There so many EL examples of trading ideas in magazines and websites, it's valuable to be able to "read" those to learn the important trading concepts. 2. When implementing an algorithmic trading system, it will be necessary to graduate to object oriented either Java or C# to build a robust system that can handle all the exception situations. Sincerely, Wayne
Wayne, Nice Post. However I will answer one question. That's interesting. Do you know anyone that makes a living automated trading "canned" applications? Never heard of it myself. I bought eASCTREND for tradestation when it came out about 10 years (?) ago. I don't trade it for a living, but It work nicely through-out the 2003 - 2008 years in 2000I and TradeStation. It help build my retirement nest egg. RabbitOne
Wow. That's cool. Now that I recollect, I was always looking for something that could be used to make a living trading. A trend system requires more patience and you obviously had that and used it well. Congratulations. Wayne
Wayne, Thank you. But rapped up in this is an important lesson Ilearned. One of the things most traders tend over look in automated trading using weekly and monthly charts. It may be boring but it gets the job done. In bull markets volatility is lower and these trades can produce nice profits. Rabbitone
You have been around the block. So true. I learned that here on ET way back. For anyone else, if you dont' follow. Most historical tests focus on the win/loss ratio, profitability per trade. But it's much more interesting to look at the daily, weekly, and monthly statistics. By the way, TZ always calculates all those by default. That's because I use mostly daily and monthly to measure how well my personal strategies are working. Plus most good traders use exits for when they make a certain profit or loss during a day or week. that can dramatically improve your results on a monthly or yearly basis. Wayne
Think Rabbitone was actually talking about basing strategies on charts with weekly / monthly price bars rather than (as you suggest) looking at weekly or monthly trade performance statistics. Although I do agree that daily summary stats for intraday systems are interesting to look at.
Your both correct. I trade automated systems off weekly and monthly charts as well as do reviews of statistics weekly and monthly. Many of the stocks I trade with intermediate term systems such as weekly allow me gage my progress against the markets much easier. I also find these longer term trades smooth out my trading account draw down and lower my trading cost by performing fewer trades. Long term trading is its own subject with automation.