"Also you could add additional strategies that trade concurrently, something that a human may not be able to do easily if at all." That's the main purpose of automated trading. No human is able to enter like 200-300 orders per day and have such consistency as a dozen of uncorrelated systems.
Not to change the subject, but I've been hearing that from several traders right now about TWS just freezing.....and requiring a restart. Does anyone know the root cause of this problem ?....i.e. datafeed, workstation software, internet communication, etc ? Kind of scary.
syswizard, After the first time I had the freezing issue it was just days after upgrading TWS.... I then downgraded back down and never had the problem since.... PS: Bot is up $920 in my ib paper account
I suppose you could spend countless hours and money testing computer programs out, or you can learn to trade and just do it yourself. I personally like to make money now, not months or years down the road when a computer program can finally do what I can do easily. I am sure the time will come when most everything is sent through a computer program, but that time is not here, nor anytime soon as the other poster mentioned. The technology is just not there yet. Of course everyone dreams of creating a perfect program that cranks out like an ATM, but I think very few actually ever get there. Even if you did program something that works now, there's no doubt that eventually it will fail. Then you get to spend hours and money again reprogramming the program. I guess my view is simple - trading is a very difficult business. Trying to program a computer to do it for you increases the likelihood of failure even more, so why bother. Not to mention - so many little things could go wrong while you are at the beach drinking that pina colada... power outage, software freeze, computer freeze/issues, etc... Now if you stand over it and watch it daily, why not just do it yourself and have some fun!
Actually he's got a valid point. Frost started this thread to prove that trading bots can be successful, but actually it has proven what a difficult process developing a reliable trading application actually is. You have to be an expert programmer and an expert trader. It might be a good for Frost to improve in the trading area (an over 1300 loss on a day when the markets where just in a downtrend?) - there's a huge amount of basic trading knowledge that needs to be gained before he can begin to put together an automated trading system. Have a nice weekend all. Jimmy J
Okay, I'll give it a shot: Both take time and money, I'm not sure why you're so quick to conclude that one is more effortful or money-consuming than the other. That's definitely not a given. Once you have a succesful strategy that you can trade manually, automating does NOT have to take a long amount of time, but it can, depending on how you go about it. I beg to differ. How do you suppose the vast majority of traders execute their orders? Do you really believe there are no, or only a few, successful ATSs out there? This is just plain wrong. Sure, just like every discretionary trader dreams of making the right decision all the time... (not that I believe this is true, but just to show you the sillyness of such a statement) Once again, I completely beg to differ. There is no way you can prove this statement. It seems that you have a problem with strategy trading, not automated trading, because automation of a strategy, if done properly, provides nothing but advantages over trading a strategy manually. Do you believe traders that stick to a disciplined strategy (not inflexibly, but while it remains profitable) are on the wrong track and that they will all eventually fail? Being a strategy trader does not mean you have to trade a single strategy until you blow your account... Again, I don't see the logic in this statement. There may be bugs in getting to the working solution that you wouldn't encounter regularly, but once it's working, it's insulating you from things that are more likely to cause failure, namely human emotion, human fatigue, human error, which are definitely bound to happen. Most definitely. However, if you don't anticipate and make your software/hardware prepared for such situations, you have no one but yourself to blame for any bad results coming from such events. Now I'll get to the counterpoint. If you are a successful discretionary trader, why not trade 10x the number of instruments you are trading now? You could increase your profitability and lower your drawdown significantly, but unfortunately the increased workload would likely be overwhelming. Even if you could manage it, the increased work and therefore fatigue would undoubtedly lead to more errors in your trading, and decreased efficiency. Obviously, scaling outwards is much easier with an ATS. The ATS doesn't get tired, it doesn't make mistakes (in terms of following the programmed strategy), and it does not get emotional. It's also wonderfully easy to test how an ATS would perform on past data and explore what other stocks or contracts the ATS would be successful at trading. I could go on, but I don't think it's nearly as one-sided as you try to frame it. There are disadvantages, like the additional-needed skillsets to build the ATS to a proper working form and some extra expense, but in my humble opinion, the advantages of reduced workload, reduced mistakes, and the easy ability to expand outward to trade other stocks or what have you, far outweigh the disadvantages. I don't think discretionary trading is a waste of time either, but ATSs are not some futuristic fantasy...
Only true if you're going it alone, which is already setting yourself up for a very difficult task to begin with.
Ditto, you nailed it Kohanz. The only point that I could think of that you didn't mention is that personally I find that developing an ATS is more intellectually challenging/stimulating then just grinding out manual trades day after day. Obviously that is a personal choice, not everyone enjoys programming.