slug, like they said you can use NinjaTrader but you will have to make assumptions about what's wrong with the backtests. That in itself is a huge study, but if you work with daily data it might be less severe of an effect backtest vs real world. If you're trying to test tick data, you will get bogus results because you cannot assume market structure (ie that your order would really get filled in reality) 100% of the time, not even 50% of the time. Just be very cautious with your results and plan on a decrease in AvgTrade anywhere from 20% to 100% in real life. It is if you only care about minimal features and don't care about common pitfalls. Even backtesting has a lot of gotchas that really requires a lot of time to figure out, and knowledge about how to determine fills, slippage etc. If you don't care about liquidity, slippage and commission, then sure, a trainee can do it, but as soon as your model hits the real world it will crash and burn - not because the system is bad, but because you tested it in fairy-land. I speak from experience working on backtesters and execution for the past 2 years, programming. If you use amibroker or any of the backtesting platforms you need to check if they're not lying to you - but how? Etc etc. You get what you pay for, always.
hiring a programmer you could make a career out of.. they hold you up at Gun point if your not careful.. not following any stardardization, using proprietary languages/frameworks that are very unsaturated... such that when you wanna leave them its worse then getting a divorce.. and they all try to do it... its the perfect example of the less you know about programming the more you pay.. and its very very convex... if you know nothing you could get a white labeled open source app with no support for crazy money... the best is to take some time and get a working knowledge of it if your going to be in finance... Ninja trader is very user friendly.. i haven't used other platforms though.. but the transition from wisywig to actual programming is very intuitive.. you can see clearly what the wisywig generated and get a feeling of how to tweak/change things in the code yourself.. their forums generally have good response time with help on programming problems..
same here, exactly. I thought about hiring a prgrammer to automate my strategy. Best advice I got was, "it's a lifelong project, there will always be bugs that need to be fixed, make sure your relationship with the programmer is life long" and then he pointed me to some websites to learn programming. I checked them out, but even the beginning sites and books were way over my head, and I couldn't find a place to start. best idea I can come up with is find a programmer who specializes in trading, and pay him to teach you how to do it. Start out with something simple like, "How do I tell the computer to buy?"
I have very high language skills, but the only language I ever needed was JCL, and I was very good at it. It came quite natural. But then, I had a teacher who got me started. It's just a matter of learning the rules. We already know how to trade, we just don't know how to tell the computer to do it like we do it manually. we no longer manually type in "Buy 100 at 32.36" we already have hot keys and short cuts in place. If we can get the computer to do that for us, it shouldn't be that hard to teach it to do more complex operations. Most likely what you call "talent" we just call common sense.
how about something like community college evening courses? i've done that in the past. worth every penny
I have used many developers, this big $ to get someone good is incorrect. PM me if you need somebody.
The biggest problem most traders face when trying to automate their strategy is that they really don't understand that their strategy cannot be automated. Very few discretionary traders have a system with rock-solid rules. There is almost always a lot of intuition or experience involved that affects the trade decision. When they get involved with a programmer they soon learn that it can cost a fortune to attempt to accurately capture their subjective rules. A substantial part of most trading strategies can be Gray-boxed where most of the grunt work is done but the final trade decision process is often difficult or impossible to automate.
First off automated trading system development, evaluation, and implementation takes extreme amounts of time, patience, and persistence...years... that said... If you don't have the money to spend on a programmer for the next couple years then the next best thing is to use NinjaTraders Wizard and study and edit the source code it creates as a starting point to learn how to program your strategies. This is very easy to do, but if you can't do this with persistence then you have no chance. It is very unlikely you will build a profitable strategy just using the Wizard alone. And even if you do build a strategy that reports profitable, you better run it through Market Replay or Walk Forward test your strategy using a live data feed or you will most likely lose money. You will only learn the potential downfalls of your platforms backtesting by experience.
The OP needs to think a little and then use an exisiting fast track solution. As a warmup, consider how writng for publication is done these days. 1. You turn on your computer. 2. You click the Dragon Icon. 3. You put on your head phones and mike. 4. you bring up a drop down sheet that is blank 5. You click on the sheet. 6. You talk. Words appear. 7. You "scratch" stuff and re-speak what is better. You save the work and print it so your editor can work it over. This process is duplicated for p[rogramming with your platform support personel. the language I use is called "Trade Sense". I follow the diections to do what I need. If I have a question, I make a phone call. I am tranferrred to a Tecky. He gives me three numbers to type. His key stokes fix or add to the directions I am following as he punches up a solution on my coding in my computer. If I am creating a whole new part of a solution, I call and get a "coach". He is on gotomeeting shortly and so am I after I type in the number to join. He video tapes our session and sends it to me so I can refer to our coaching session in the future. I use criteria, filters, functions, rules, strategies. I have my own baskets and libraries and I can export them to others and I can import libraries from those who team with me. I pay a monthly fee for the platform. I went to a three day training program. I have ALL of the past stuff in all libraries of the platform providers. A buddy knocked out about 150 functions so we could make trade sense function as an SQL RDBMS. It took two days and cut our remaining work by a factor of five. We ae doing three applications. Three are on paper. We have some pieces to do for one of them. We have a strategy to punch in for another. The other one was our trial horse and it took a couple of weeks to finisih it doubles capital 7 times a year trading stocks as an ATS) So now I have to put a group of pieces together after finishing remaining penciled pages (of yet undone pieces) and then punch out the strategy (now on paper., too) All things are coded as pieces. pieces have names. I wrap pieces together with containers and interconnecting operators. A "verify" button highlites the automatically revised coding to make it best and if something is missing logically I am told what to do and where to go do it. What I am suggesting is that there is a "dragon" for doing systemmic code. It also makes reports for you.. You fill in test criteria(for doing the program you built in a family of contexts) and forard test from a past date. A complete traditional report is put in a file. You can print it and tweak your program. There are groups of culumns for each set of curves in the family you set up to test. you can see where tweaks are needed and you know your whole pivate library you used to create your method of making money. All lot of famous people use this system. So you can have their libraries as a starting point. They do presentations for the training conferences. Your stuff will be comparable to others. Hiring programmers is a thing corporations do in the financial industry. They try to beat the indexes and most do not. They are too big so programming doesn't work apparently. Also there is the garbage in and getting garbage out. Programmers usually do not know about trading. Programmers often try to sell traders their "inventions" of what traders really need. In this information era that emerged from the indutrial era, a person has to grow his mind to be part of what it what. I have worked through the creation of ATS's since late October, mechanically speaking. I use a C++ foundation, converted to the English language for writing on paper. Then you put the paper writngs into containers which have filling in instructions. All this is supported by coaches (In GTM's) and Teckies (9 years experience type guys) who can fill in the containers by managing yourr PC, all free. You have to be able to read. You have to be able to follow instructions. Everything is TA based and Conventional Wisdom oriented. You you can just step right out of that box and do it your way, too. I have had both coaches and teckies catch on to how money is made for the first time. They go batshit wanting your next call.
Thanks to everyone for the input. I did hire a programmer to do some work for me back in 2004 and lets just say it didn't end well. I also recommended a programmer to another group several years ago and that too ended badly. That said, I have no interest in hiring a programmer and reliving past mistakes. I always felt confident that many of my trading strategies could be implemented through a grey or black box. The reason for my initial post was to see if there was anything out there that could possibly provide me with a sort of "point and click" access to building my own box. I wasn't sure if the technology was there yet since I haven't really done anything with program trading in almost 10 years. I was recently offered the opportunity to work with a software that is supposed to offer traders with little to no programming experience the opportunity to build their own box with a week or two of extensive training. Obviously the content needs to be solid or else any box that might come out of this will be garbage. I am skeptical which is why i asked my initial question. I have seen the software and it worked very well in putting together a super simple sample strategy that I offered. Is it reasonable to assume that something like this could allow me to program my own boxes? Understand that I am extremely skeptical as any trader with years of experience would be.