I am sorry to write so many replies, but I want to state that this is exactly what I am trying to do, and I agree 100% with the logic behind what you wrote. Thanks for your clarity and conciseness. Very good point also in stressing out "don't pull the plug on the system during the inevitable drawdowns" - because only when you do that, you know that you are fully automated. I know because I pulled that plug a few times before (because the system wasn't good enough and didn't have any money management).
How long before you were fully automated?
- Thread starter travis
- Start date
I am sorry to write so many replies, but I want to state that this is exactly what I am trying to do, and I agree 100% with the logic behind what you wrote. Thanks for your clarity and conciseness. Very good point also in stressing out "don't pull the plug on the system during the inevitable drawdowns" - because only when you do that, you know that you are fully automated. I know because I pulled that plug a few times before (because the system wasn't good enough and didn't have any money management).
Thanks. I understand now. Probably your system was good but transaction costs exhausted your edge and you end up working for the broker for free. I have gone through this in the past. I made 20K profit and my broker received 40K in commissions. I promised to myself it will never happen.
You post is very interesting. I am in the process of fully automating a forex trading system and it is very useful to me to hear what others have done along these lines and their experience.
Quote from RainMann:
Travis
I started about 4 years ago with forex and some option trading and finally made the move to the E-minis about a year ago. Being a programmer at heart (many many years doing this) I immediately began scripting eSignal and TradeStation but not with the notion of Auto-trading. I simply wanted to know how a certain trade would perform if I took the trade every day. It was actually a statistical method for me and I could test stop values and target values. Anyway, it grew from there.
And it was just last week that I finally toggled the switch from SIM to LIVE on one of the strategies. Just like you, I am trying to make decisions for it, completely forgetting the fact that I had already made the decisions when I programmed it. Each time I turned it off, the trades came in fine.
Anyway, it's going to be an adventure I'm sure. So best of luck to you and your systems and may your connection never go down.
RM
RainMann, I read with interest and understood everything you said. Thanks for your contribution and congratulations for the speedy automation. On this poll, I voted "> 10 years".
Yes, it is extremely interesting for me to read you as well. That is why I started the thread. On the one hand I feel reluctant to speak about my trading and my systems (and I suppose others do, too), because all other traders are in a sense my opponents (zero sum game and similar considerations). On the other hand, I feel like the other traders are my friends, also because they helped me in the past. I think it will still work out, as long as I don't have too many friends, otherwise I don't know who will keep losing their money to the market and ultimately to us.
There's one lucky guy at the top of the poll, who voted "0 to <= 1 year". It took him less than a year to do everything, ever since he has set his eyes on trading. Very interesting.
He was probably a programmer to begin with. He then either came across the right person, someone who advised him on what to do, the right magazine, or the right forum (maybe this one). If your brother has been doing exactly the same thing, and you're a programmer yourself, it's likely to take you less than a year to do everything. Or then maybe he was just someone already convinced from the start of the best idea - "the way to go is automate it". Lucky him.
I wasn't even a programmer, and had to learn everything from scratch. But I am glad I knew English, otherwise it would not have been possible for sure. I would still be evaluating whether to download Tradestation since it's in English, or whether to open an account with IB, and pay 2 dollars commissions or keep the Italian account, with 20 dollars commissions and no support for automation...
It's not enough to be a programmer and to know English, you need to be willing to risk your money, to spend a lot of time finding a strategy, and to believe that something like this can be done. That excludes 99.99% of people here in Italy. But then again if you are really determined you won't let English stop you, and you'll find someone who'll help you with the language, so maybe I'd say at this point there must be a few hundreds people in Italy with an automated system (that works). How many in the US? Certainly thousands, but would you say more than 10,000 people? I am just shooting rough estimates of course.
He was probably a programmer to begin with. He then either came across the right person, someone who advised him on what to do, the right magazine, or the right forum (maybe this one). If your brother has been doing exactly the same thing, and you're a programmer yourself, it's likely to take you less than a year to do everything. Or then maybe he was just someone already convinced from the start of the best idea - "the way to go is automate it". Lucky him.
I wasn't even a programmer, and had to learn everything from scratch. But I am glad I knew English, otherwise it would not have been possible for sure. I would still be evaluating whether to download Tradestation since it's in English, or whether to open an account with IB, and pay 2 dollars commissions or keep the Italian account, with 20 dollars commissions and no support for automation...
It's not enough to be a programmer and to know English, you need to be willing to risk your money, to spend a lot of time finding a strategy, and to believe that something like this can be done. That excludes 99.99% of people here in Italy. But then again if you are really determined you won't let English stop you, and you'll find someone who'll help you with the language, so maybe I'd say at this point there must be a few hundreds people in Italy with an automated system (that works). How many in the US? Certainly thousands, but would you say more than 10,000 people? I am just shooting rough estimates of course.
I find this forum extremely helpful and I have learned a lot so I never hesitate to tell other people what I think works. Also keep in mind that only a fraction of traders participate in this forum. I think the majority of traders are gamblers and will lose to those who are systematic and exercise good money management. Usually that type of traders think they know everything and it is simple. When they find out that their winning probability depends on what other people do, it is too late. If you and I do something advanced that has an edge, we automatically limit the odds of gamblers. This is important in my view. Markets reward consistency, not gambling. Maybe some gamblers will make money by chance but 95% of them will lose.
In order of importance:
1. Low commissions. Do not trade if commissions are not rock bottom. In forex do not trade if spread is more than 2 pips, absolutely the maximum.
2. Only systematic and, better, automated. The media sooner or later hypnotizes discretionary traders. Maybe not all but 95% of them will be convinced to make the mistake.
3. Anti-martingale money management. As low risk as possible as percent of current bankroll. I use 1% max. Often 0.5%.
4. No optimal money management. Optimal methods do not account for tiny black swans. This is what killed many good systems and funds. Markets will generate tiny black swans to kill the best of systems that use optimal money management for geometric equity growth. Kelly formula has many problems in practice. The most important is that the parameters always change and are not known in advance. (Read this article for details, section: application of the Kelly formula)
5. Trade-off high risk infrequent setups for high frequency trading with small percent risk. Many trades with small risk as opposed to a few trades with high risk.
6. Know ALL the details of everything you use, including indicators, trading systems, execution software, markets and instruments, order types etc.
7. Now my rule from experience: most of the times, a nice day is nice from the morning. Just a few times you wake up to a storm and then things get better. Figure out a way to get out mechanically of losing trades. Hint: time limits, bar limits, day squaring of positions, etc. depending on style.
and much more...
I agree with everything intradaybill wrote. In detail, this is what I think:
1. Agree
2. Agree
3. Can't do it yet, because I have too little capital
4. Can't do it yet, because I have too little capital
5. Agree
6. Agree
7. Agree, but I don't want to risk over-optimizing by introducing such rules
Also, when you say "Kelly formula has many problems in practice", you are reminding me of something I read in a wikipedia article on kelly:
http://en.wikipedia.org/wiki/Kelly_criterion
"Since in reality we seldom know the precise probabilities and payoffs, and since overbetting is worse than underbetting, it makes sense to err on the side of caution and bet less than the Kelly amount."
As soon as I'll have enough capital to do so without hurting my chances of ever becoming rich (consider I have started with a capital of 4000 dollars), I will start "underbetting", as you and the article both suggest.
1. Agree
2. Agree
3. Can't do it yet, because I have too little capital
4. Can't do it yet, because I have too little capital
5. Agree
6. Agree
7. Agree, but I don't want to risk over-optimizing by introducing such rules
Also, when you say "Kelly formula has many problems in practice", you are reminding me of something I read in a wikipedia article on kelly:
http://en.wikipedia.org/wiki/Kelly_criterion
"Since in reality we seldom know the precise probabilities and payoffs, and since overbetting is worse than underbetting, it makes sense to err on the side of caution and bet less than the Kelly amount."
As soon as I'll have enough capital to do so without hurting my chances of ever becoming rich (consider I have started with a capital of 4000 dollars), I will start "underbetting", as you and the article both suggest.