What are good results for daytrading

C

C++Programmer

I'm a starting out C++ programmer interested in desinging custom trading systems using some pretty advanced ideas. I was wondering what are good results for daytrading.

Criteria: Percent after slippage, trading costs, and other fees based on an account of 100,000 dollars

What would be:

1st base?
2nd base?
3rd base?
and a home run for results?
What are the best one year results you have ever seen?
 
I witnessed a thread recently on another board in which the discussion centered around whether 1% per week was a reasonable expectation. Views about this were quite mixed, but that seems to be the magic, easy to visualize number that a lot of traders have in the back of their minds. It's complicated a bit by the question of what the 1% is relative to (i.e., to your current trading account size, or to some more vague notion of what you consider to be your total amount of trading capital). There's also quite a bit of sentiment that % profits decrease with trading size, so that attaining a particular % profitability doesn't mean that you'll be able to maintain that rate over time.
 
I have approxiatmly 90,000 dollar in my account. Over the last month I’ve consistently made over 1.5% a day without margin using my computer day-trading system.

I’m trying to figure out if this is good or not. Should i try to start working on a new system?
 
I expect Hitman is sending you an email to recruit you for his team as we speak. Seriously though, 1 month is a very short trial period. How extensive and rigorous was the backtesting of your system?
 
This may not be the answer that you are looking for, but I think it depends on what risk/reward you are looking for. You post does not mention how much risk you are assuming. Now, 1.5% of a 90K account equates to $1,350 per day (or $351,000 per year). Personally, I think I could live on that.
 
Some questions??

What is your worst possible drawdown?

Can you tolerate this?

How does this system react in non trending markets?
How does it react in a bull market?
How does it react in a bear?

Usually one month isn't enough for backtesting. Quite a large # of systems do well in certain times and very poorly in others. Can you figure out when it will do poorly? How will you react when it does?

Robert Tharp
 
I'm an avid C++ programmer who just started making systems that perform pretty much the same in any market (I hope). I'm just trying to figure out what should be my baseline for results. I'm trying to use some ideas and concepts that no-one has implemented in a trading program. The worst result so far in the last month was when i lost 1.9% on dec 7. Besides that the system performs close to 1.5percent a day and does really well on some days.
 
some questions to think about:

1. What does it trade? futures (slippage guaged easily) or stocks (@!#$@#$% !!!)

2. How frequent does it trade? if it does 40 round turns a day, then 1 month is plenty of testing, if it does 4, then you need 1 year of data.

3. How is your entry/exit implemented in your program? I mean watch out for hidden optimization, which can produce amazing results in simulation, if you only have 1 losing day in a month, then I bet you must have some hidden optimization in there. I mean Tony Oz can not go 1 month with only 1 losing day.
 
It trades stocks long and short.

Part of my program goes through about 7000 stocks each night trying to figure out best daytrading stocks the next day factoring in: historical bid ask spread, volatility, and volume.

The second part of my program performs the magic by getting in and out of stocks around 8 times a day at the right time to maximize profits.

I also keep track of intra-day stock data each day and save it for later.
 
You must log in or register to reply here.
Back
Top