Comparison of software in several programming languages

We compared tradestation, metastock, ninjatrader, tradersstudio, multicharts, wealth-Lab, rightedge, openquant, and other platforms with the largest number of elitetrader.com platforms.

Both tradestation and metastock have a large number of out-of-the-box code users (many of them are well-qualified or professional trader). Their programming languages are relatively simple and Their strength lies in the convenience of developing various indicators, however, the backtesting function is weaker than other functions.

Other platforms have relatively strong backtesting functions and have their own strengths.
• Openquant, wealth-Lab 5, ninjatdge, and rightedge are all Based on. NET and use the C # language.
• Wealth-Lab 4 Adopts Pascal-like language
• Multicharts uses the power language compatible with the EZ language of traderstation
• Tradersstudio uses basic-like languages
• Amibroker and metastock are similar. They use the formula language based on the sequence. The language of amibroker is between C and basic, similar to mt4.

Compared with these platforms, amibroker has the following advantages that I prefer:
•Fast Running. Some users I have seen many times say that AB is the fastest software they use, especially the backtesting performance, which is the fastest among all software. I installed ninjatrader and AB in the VM. The loading speed of NT is much slower, and there have been no response in the middle several times. AB load is very fast.
•Extremely flexible data sources. I like it very much. At present, I have used FXCM, quotetracker, and IB as the data source. The results are good. It is also very convenient to use amiquote to download EOD. I was hesitant to use ninjatrader, but it was too inflexible to see nt data sources. At least there is no convenient data like amiquote. You cannot use the DDE data source. Therefore, FXCM or other data sources are unlikely to be used.
•As a quick development and testing environment. I have seen some veterans say they use AB to quickly experiment with many strategies. Because AFL is based on series, it is much easier to operate than those Based on. NET. I also see some code comparisons. ninjatker is much more complicated than amibroker. An old trader complained that nanjatrader's C #-based language is too difficult for non-programmers. Note: amibroker seems to be relatively strong in the EOD test and is not clear about the use of intraday data for testing. Update: v5.2 can even perform backtesting and scanning on tick.
•Easy to integrate Interfaces. There are many ways to generate a transaction ticket using AB in the future. Do not pay attention to whether emails can be sent.

 

Considerations for the analysis and testing platform
I checked the evaluation of some other tools on the Internet:
•Ninjatrader (NT)From the perspective of its operation model, it is still closely related to dealers, and the data source is not open. Some people have commented that the direction of NT is to build a transaction platform. In terms of development and testing, the. NET-based nt5 is too resource-consuming. This is how I use nt5, And it is slow to load each time. Ninjatrader.
•Wealth-lab and rightedgeAll are based on. net and C #, but wealth-lab is mainly used for testing and experiment. It is not a complete trading platform, data sources, brokerage, and automatic trading interfaces are not built-in. In addition, wealth-lab was recently acquired by fidelity in some U.S. markets. The difference between wl4 and wl5 is also great. From this perspective, wealth-lab does not need to be considered.
•RightedgeAccording to the comments, it is not as comprehensive as openquant, so it is not considered for the moment.
•OpenquantIt is a retail edition of quanthouse (for institutions) quant developer (originally smartquant technology was acquired by quant house ). It is also based on. NET and C #. I read the document and found that the structure is well organized. Openquant also provides positions, capital control, and other functions. It also provides brokage interfaces for automatic transactions. I saw a trader using amibroker saying that he used amibroker for rapid development and testing, and then made more detailed analysis, deployment, and transactions on openquant. Seeing some code, I personally feel that the Code workload is still very heavy. Additional personal comments (pasted from I feel bad. Creating strategy from the metaproject project seems too cumbersome. A set of strategy is divided into the market, entry, exit, money, risk, and other parts, a bit like the original introduction of turtles
"Market: What to buy? Entry: when to intervene ?" Standard format. Each part is written as a component similar to. NET and then merged into a set of strategy. This is a bit like ts5.0, but it seems that there are a lot of complicated, cumbersome, and unrelated code to strategy written with pure oop.
This software is intended for programmers or programming enthusiasts, especially. Net programmers. We are not sure whether to evaluate and consider openquant in the future.

 Most people's opinions after reading are as follows:

I think amibroker has higher programming requirements than tradestation and metastock. After all, amibroker has better functions. However, it is much simpler than those Based on. NET and C # platforms.

It is much simpler than mt4. I used to develop a framework with mt4, but it was not quick enough to experiment with different strategies.

Amibroker, the software processes data very quickly, has complete data interfaces, and many people are using it. I personally think the only drawback is that it is in the fully automated transaction part. If you use IBC to interconnect with IB to control the order, the amount of code is relatively large. It is also difficult, so you have to work hard.
QD: for gamers of the ashes. There are two usage methods: one is to write the transaction system directly under the QD interface, and the other is to develop your own transaction software using QD APIs. You can install QD even if you don't need QD. Reading the help documentation of QD is helpful for developing transaction systems. The disadvantage is that QD does not provide subsequent services (if you use the dversion, you generally cannot use the genuine version .), When the broker API changes, it is troublesome to modify the relevant program. QD can support Ib's advisor account, but there are still some problems.
OQ: It is no good for IB to run the formed transaction system with a separate account. Benefits from the use of event processing mechanisms. Compared with qd, OQ is not flexible and QD functions are more powerful.