There are a few such programs but I would highly recommend that you learn to code some easy vectorised back testers.

If you are against learning to code then I can recommend metastock and I think Ninja trader can do it as well.

However you really sacrifice a lot on flexibility.

Manual coding of trading systems, which I do in TradeStation, etc., was made somewhat obsolete about 10 years ago with the creation of Code Writing Machine Learning Engines applied to Trading Algorithms such as TSL. Whatever your choice of off the shelf indicators are, it is probably far less than optimal. These ML algorithms, will choose the inputs, write the mathematics, test on blind data, reduce parameter space, mutate the math, produce a variety of reports and finally export trade ready code in a variety of languages and this code is used by large and small funds and traders worldwide who are non programmers for the most part, and, our ML automatic code writing engine is highly rated by the independent CTA, system tracking company, Futures Truth since 2008 on Sequestered Data if I may toot our own horn a bit. You should take a look at the high and low end products out there that have changed things in the algo space. 1n 1987 I was manually writing algos, but now I use a Machine.

I've build my own regression testing system I use it to test my automated trading strategies.

The TT Analytics block in X_TRADER from Trading Technologies allows you to easily add technical indicators to algos so you may build your automated trading system without writing any code. You may read more about it here: https://www.tradingtechnologies.com/help/adl/miscellaneous-blocks/tt-analytics-block/

back testing Forex ? what will be the uses for ? loosing money on the next day ?

What you need is not back testing but P&L explain based on back historical...

Now I suggest you use R for all prototyping .. and avoid hiring coder for that matter since you are a rookie. The main reason will be the Copy/Paste + minor change...

You can also use Python for the same but with less support in public ...but far more help from pro quant

Thank you all for your replies. I have recently started looking at Python however it will obviously take time to learn so in the meantime I'll have to settle for less optimal solutions. You've got to start somewhere, right?

I had Updata for years but didn't really make the most of it and have had a Metastock trial in the past but my goals and trading strategies were different then.

Thanks again.

Please, check out QuantConnect (https://www.quantconnect.com)

It is written in C#, but the API is very intuitive. We have a forum there that can help you write your own algorithms.

Gotta be careful. The market is not designed for independent genius. Only the large banks and wall street investment firms are allowed to manipulate stocks via algorithm. https://www.fbi.gov/contact-us/field-offices/seattle/news/press-releases/two-sentenced-to-prison-in-stock-manipulation-scheme

Definitely learn some code. The investment in time will be well worth the gains in flexibility and efficiency later on. Check out this tool, which is in my opinion the simplest way to get started with serious code-based algo trading - http://www.zorro-trader.com/. Python is also a great tool, but will take some time to reach a level of proficiency at which you can write your own backtests.

Check out calgo from www.spotware.com

Code does not seem like it will help you - unless one has a fairly comprehensive background across a few languages and has coded before, it can be rought to pickup on someone else's "flow" - especially if they are NOT a pro programmer (or know it's sensitive, and obfuscate a bit on coding content). All of the above of course presumes you can actually gain access to the code you're trying to evaluate to see where the Open trigger is or where it is not (should be?). Best bet is a tool set that would let you work @ high level to tune and test algo, then compile to native/lower level codeset/app as others have mentioned above. From the sounds of things, if you are back testing, you may want to simply debug/dump each input set with corresponding output especially if you are doing forensics of some kind. There ae a lot of tools that can accomplish this across a lot of disciplines (any 6S heads out there? Industrial engineers, Control engineers?) - find one that works for you intuitively.

www.quantopian.com

http://quantstore.com/

there are plenty of options. If you want Wizard-like application where you can build your strategy without programming, you can use for example EA Wizard: http://www.strategyquant.com/eawizard/

It will generate full MQL code that can be tested in MetaTrader.

Stephen, it seems you have a strategy formalized with a couple of indicators and some logic. My assumption is that this logic is something in the form of "when price crosses EMA and MACD crosses signal then buy". Perhaps the PARSAR is used to define a stop. It can be that the logic is much more complex ofcourse, but what I find a very good tool for coding and backtesting strategies is MultiCharts, even if you have no programming experience. Strategies are coded in EasyLanguage (http://www.multicharts.com/easylanguage/) and I believe this has the shortest learning curve and gives you the possibility to easily backtest your idea. You can first define your manual setup by applying your indicators on the chart and then formalize it in EasyLanguage using the same indicators. The logic based on a couple of indicators will be very simple and you can then perfectly test and optimize your strategy.

It wont code the entire trading system for you (not sure anything can) but I would highly recommend the Adaptable Blotter (www.adaptabletools.com) as the environment in which to do your trading. Its data agnostic and asset class agnostic so you can trade anything you like and its very powerful functionality will allow you to do everything you list in your question: i.e. select indicators and settings and create automated trade logic. It will give you out of the box most of what you need and let you configure it exactly how you want.

@Jonny to me this looks like a highly customizable top layer for various data feeds/platforms. While this is certainly a nice product, the website doesn't explain how to run a backtest of an automated strategy and looks more like a front-end tool. Or does it rely on a integration with a third party platform that offers the auto trading features? For developing strategies and backtest them it looks a bit complicated but perhaps I'm missing the point.

So-called self-coding or drag-and-drop coding methods are technically possible, but only on the most basic buy/sell-hold level strategies. Not if your strategy requires any degree of sophistication (multiple if/then/else statements). I cannot imagine any trading system that would NOT have multiple if/then/else statements, these would have to be adjusted to each platform you intend to trade.

Plan to hire a freelance coder. hourly projects are best. But even if you find a coder you can trust and is competent, then the challenge is the willingness to go through 100s...1000s[...10,000s] of hours taking screenshots, and explaining EXACTLY what you want in a logical format...working through the bugs until you get exactly what you desire. Ahh yes, spiral development :)

I have seen some very basic rule-based wizards & tools in various software packages, however I've found that they are lacking in one sense or another when comparing to the complex capabilities found in actual code trading environments.

Sounds like what you want is a trading software package that uses a graphical programming language (something node-based with flow arrows, much like what Unreal3D and Blender3D use for various portions of their software). See https://docs.unrealengine.com/latest/INT/Engine/Blueprints/index.html for an example of this programming paradigm.

I'm not currently aware of any trading software that has this kind of high-level visual language, however I can see the merit in using this sort of language, as it lowers the barrier to entry.

Perhaps there's a growth market here for trading software companies.

There are many suggestions here, the last thing I would recommend is "learn how to program python". You waste your time. I have created a little video how I would do it, and from there it is easy to take it to another level if you wish to do so. https://youtu.be/WgYr_nKlJ8Y

Most important is to get an idea if you are on the right track without spending hours and money. Let me know what you think.

@Hans You are right. It IS a front end tool only and yes it relies on integration with a third party platform to actually execute the trades. I understood the question to be seeking advice on a platform which will easily facilitate indicator selection, trade settings and to evaluate logic of the sort you identify: "when price crosses EMA and MACD crosses signal then buy". But the actual execution on buy is not something the Adaptable Blotter itself can do.

Your suggestion of MultiCharts is a good one and it does look very impressive.

The key point with all those visual and other tools available on the market is that you have NO IDEA about the reliability of the backtesting engine of any of them. I remember my experience with Trading Bloxx (or maybe it was another similar "visual" software of the same kind, I don't remember, I don't use them and this was a kind of a game courtesy one of my friends) — when I exported an EL code from it (yes, some of these "visual" platforms even are able to generate executable codes) and ran a backtest in MC — well, I saw just a very ordinary equity curve oscillating up and down around zero. At the same time it looked skyrocketing in the "visual" platform itself.

Even worse: most of these platforms won't allow you to use your own data, and since you're going to test an fx strategy, then the variety of data sources and their quality is similar to the variety of species in a menagerie. So, not only you won't have any idea about how the backtest was done, but even based on what data.

So, my bottom line: if you want to seriously test anything, and be confident in your results, then learn the market structure to know what you're doing and programming to be sure you tested it right. Period.

@Stephen Jacson, interesting and highly up in time question. There are many suggestions, but the answer is control, control and control - to be in control, you need to make a research.

Hi for what it's worth try this: http://www.eabuilder.com

Alphadvisor (www.alphadvisor.com) is possibly the best I have found so far. Nevertheless, nothing like the ability to code (R, Python, mql4 or whatsoever) to program your strategy in the way you want.

I use Amibroker for automation but find that simply look and see works pretty well..........you can see my daily video at ronjaenisch.com

@Stephen Jackson, Have you looked at BloodHound? http://www.sharkindicators.com/bloodhound/

@Krishtop It is interesting what you say about T-Bloxx. How can a strategy be successful in one software and not in another? Often it is related to settings or it can be the money management rule. Just to accept it isn't good enough and to disregard the system isn't either. May be either one software isn't doing the job right or so different that it matters? Comfort of backtesting often comes limited options that help to create winning systems.

Using external data is also very important. Using clean data can make the difference. You are absolutely right, software that does not allow the use of eternal data is kind of useless for backtesting!

No matter how one wants to look at it, anyone wishing to play the game “trading” using automation will need either to program it themselves, using available trading software, or have someone else do it for them.

The markets are a lot more sophisticated than a few decades ago. With millions more participants, and most of them computerized. Over half of all trading is now done by machines. And if you want to win at this game, you better be ready to face them all, or code your programs in such a way as to mostly bypass those machines.

True, EasyLanguage, and by the same token MC, are relatively easy to learn. But, still, there is a learning curve. Wealth-Lab has a strategy building wizard (drag and drop) where a novice can almost immediately design simple trading strategies. I think a novice has to go through this apprenticeship phase, to get familiar with the nature of the problem he/she wants to solve, and get familiar with needed tools.

...more

A newcomer to the game should realize that he will be up against the brightest, the best equipped and the most sophisticated traders in the history of humanity. Never before could so many do so much. The amount of brainpower put on this trading problem is tremendous to say the least.

I can definitely say to a novice that a simple trading program won't do the job. They've been tried by thousands and thousands of newbies who are no longer in the game, or have abandon even trying to program what they have in mind.

The game is so complex that we all have diverging views of what it is on some basis. We can all have automated trading strategies that work for some reason or other, and for most of us it is sufficient. We all have our preferred methods of doing things. It is probably why we are still hanging around, as part of the survivors. But, I think, most (not novice anymore) would agree that it took a lot to get there.

…more

But the novice has not reached that level of knowhow, experience and confidence to even think that the first “simple” idea that comes out of his head is a winner. Give me a break. Let the novice jump right in, but, no matter how, he will have to learn bit by bit all he/she might need to do a decent job. And that journey is accomplished one step at a time, just like in any other endeavor.

@Guy I absolutely agree that there is a lot to learn, even the basic things. I am not even talking about things like Position Priority which becomes important depending on the money management method used in any given portfolio or other important tools like Monte Carlo Lab, Walk Forward Optimization (which I call Parameter Stability Testing). Features like Worst Case Scenario should also be part of any testing environment just like limiting the position size to a certain percentage of the daily volume... This list goes on forever.

It is true that drag and drop of my strategies are sometimes hitting certain limitations but before any novice system trader without any programming knowledge gets there it will take a while. At the same time I believe it might be better to hire a programmer for the special occasions, it will dafe you a lot of time.

Nope.

Yep

If you can not code it, you'll never know how it really works. Unless it is something very simple.

@Costas, and if it is something simple, it might not be worth much either.

Ofcause you can develop strategies without coding today - to know "how it realy works", you make a backtest - and you compare when the strategy trigger with some visual indicators of your criteira (if your development platform supports this). You can also let the strategy run for some time, and compare the live backtesting with the traded results (if your development platform support live backesting). And you can finally make a backtest and compare with the live traded results.

The worth of something simple to a novice trader is the experience gained with the programming of a strategy, the platfrom, backtesting and optimization procedures etc., and then conclude for him/herself to go back to the drawing board. This going back to the drawing board is not only technicalities of various sorts but more importantly the understanding of the inefficiency your are trying to exploit. But you have to start somewhere and I believe there is value in every step in this journey. Even losing money has some value.

You can probably do this through Tradestation. This may require you to learn their programming language. But indeed, you will need to be confident in your system and the best way to get this, is in developing it by yourself

Soren, with all due respect, please stop your promotion. You don't compete against "sophisticated systems". You still compete with your own idiocy or greed and fear. According to the latest researches 90% of the most important trading decisions made by those who trade real money, not several hundreds of thousands, are made by human discretion.

And also please refrain from stating that a "visual programming" platform, limited BY DESIGN to only the choice of its creators, could be more flexible than a platform which supports full access to any basic element via code.

@Alex, perhaps a good idea to tell about how many years you have spent tesching your self how to develop automated trading strategies by programming, and tell about the successfull strategies and all the money you have made. Would you advice beginners to do the same ? --- a question I ask, since not long ago you wrote you wouldnt do the same again - but still you recommend others to do what you did ?

I you didnt succed, with all your sophistiacted neural network implemtations etc, how would you expect a beginner to succeed ?

Soren, you obviously take me for someone else. I've never used neural networks, and other sophisticated methods. Nor did I write that I wouldn't have done the same (what that same?..) again. The brief answer to your previous comment is definitely yes.

A pitty you dont recall your sel-reflecting comment, probably your best comment ever. I was referring to Edgesense - no neural network ?

No.

If mt4 platform. Try this: http://www.forexeadvisor.com/

This is quite funny, by calling Visual programming you do realise that most of the time you are stuck into data driven design. Meaning debugging is barely not possible, since it is mostly context and combination. It is fine, when you have small combinations. Drag & Drop, Copy / Paste later and you now start to have a gaz factory. Plus you won't be able to use algorith found over the net or magazin, you will rapidly be blocked.

C++ the most advanced and versatile language has only 113 keywords

80% of them will never by any of use for a trader ( usually for, if or switch and that's it )

Same goes for many other langages...

Debugging a sequence flow is easy, it is a step by step process

Debugging a multithread data driven error is context hold procedure, you need to reparallele the whole combination, most platform won't provide this power anyway. So you will loose days to find error of design.

this may help - http://fxcodebase.com/code/viewforum.php?f=17

supports MT4 and FXCM platforms - FXCM also supports C++ development

@Stephen, have you looked at BloodHound? http://www.sharkindicators.com/bloodhound/

Hi Stephen, I developed a system in JAVA that allows fast implementation and testing of different strategies. Equity Curve analysis will be done in Matlab. I also have long-term daily and some intraday data for back testing, forward testing is also possible. Some initial test I could do for free. Contact me if interested.

There are 'apps' like StrategyQuant, ... that allow you to specify indicators, ... whick will generate/code trading systems for you. But you should not expect that these systems will be profitable. There is much more in system developement than chosing an indicator.

Actually apps like StrategyQuant can and do generate profitable strategies, but it is really not as simple as selecting a few indicators and expecting a profitable strategy. It is based on machine learning, so the process involves some understanding of statistics, robustness, etc.

It also depends what you are up to - StrategyQuant generates random strategies for you, it is basically looking for an edge on the markets.

If you want to test your own ideas without the need of programming you need an app where you can build your strategy in form of some building blocks or simple conditions - but this will always have limitations compared to programming strategies.

Full disclaimer - I'm author of StrategyQuant, so I'm inherently biased.

Our group company TickCOM is probably the provider which is most far in the field of non-programming - no limit development - and the ability to add custom features by programming. There need to be the ability to add custom XX, as no visual development platform is complete. www.tickcom.com

maybe j-forex, as it produces the code by grafic block design, and you can copy the code as the code for the blocks are clearly displayed.

Hi Stephen, take a look at https://www.quantopian.com/

If you're a rookie in programming there are great tutorials.

For a completely different, visual approach to development and testing take a look at www.qbitia.com which won Best Trading Development Software this year at The Technical Analysis Awards here in London. Programmers will hate it as there is no programming. You drag elements to create the process. It is currently connected only to CQG's execution platform but that will change. I see it going far.

Good to hear The Technical Analysis Awards is still alive - anything new, or still the old stuff they keep shortlisting ?

Cut to the bone - this is about to develop automated trading strategies which can perform in the market - to do that, you need a development platform which makes this possible. The limitation is not You, but the development platform you use.

Our ForexGridMaster MT4 EA automated trading robot in development since 2005 enables traders to create an unlimited number of strategies targeting low to high profit, no programming necessary ... http://forexgridmaster.com.

You can go with desktop apps like amibroker or ninjatrader which makes your backtesting process easier and get quicker statistical outputs. Creating a simple strategy in Amibroker/Ninjatrader and backtesting it is a child's play

With any 3rd Party Platform - check out their terms and conditions of use very carefully before you commit. Most secure some proprietary claim or other over your development and/or restrictive use other than from their own Platform (e.g. Quantopian).

Also make sure your strategy is kept on your PC, and not exposed to the platform provider or broker.