question about maximum trading frequently

 

Hi, this might sound like a strange question but here it goes anyway.

If I wanted create a trading robot that trades as often as 'humanly' possible while still hoping to make money in this market what is the maximum trading frequency on the order of?

Is it thousands, Hundreds or Tens per day? Suppose I used some type of very aggressive scalping strategy that uses any currency pair. Could that kind of algorithm yield more than a few hundred trades a day?

I welcome any input on the subject from anybody that knows Forex well enough to answer this.

Thanks in advance.

 
I don't think the frequency of a trade have anything to do with its profitability.... for better nor for worse.
 
ubzen:
I don't think the frequency of a trade have anything to do with its profitability.... for better nor for worse.


Sorry, i might not have made clear enough the thrust of my question (remember i prefaced my it with 'this might sound a bit strange').

I am not making some sort of implicit correlation between maximum trading frequency and profitability.

I am specifically interested in what is possible as far as maximum frequently I can open and close trades with running a still 'usable' algorithm.

The traders who use algorithms that trade the most frequently in the forex market, what number of trades do they achieve per day?

I need this for a research project where I need to do statistical analysis of the trades that have been made. For this I need a large sample of trade within a relatively short time period.

 

If profitability is taking out of the equation then the #-of-trades is dependent on the technology itself. So, somehow I think you're referring to profitable high frequency traders. When I was new here, I asked a similar question and the response I received from a senior member at the time was something like it's not possible to implement within mql4. I now agree with him to some extend.

The high frequency scalpers you're reading about are probably the same ones I've read about. Put simply most of their style is Not speculative. They either place high-frequency trade or high-volume trades when the opportunity presents it self. Example: when prices are mis-aligned (arbitrage). These guys do not use mql4 and the speed of light (it-self) is a limitation for them. They built servers next to the brokerage in order to beat the competition to the price.

Coming back to speculators and humanly possible. I've had the opportunity to observe some scalpers at work before. I think on average perhaps 4-trades per hour would seem ideal. You can definitely get algorithm's to place those numbers of trades (usually you'll use smaller time-frames). However, not everything those high-frequency traders said I believe. Like you can look for 5-pips profits with stop-losses of 2.5 pips. etc. But that's just Me.

Also #-of-trades and #-of-independent-events have to be distinguished. Two traders can trade the exact same system one place one-order lot of 1.0, while the other place 10-orders lots of 0.1. From my experience, high-frequency independent-events trades is a very difficult task to make profitable, in-the-long-run.

 
ubzen:

If profitability is taking out of the equation then the #-of-trades is dependent on the technology itself. So, somehow I think you're referring to profitable high frequency traders. When I was new here, I asked a similar question and the response I received from a senior member at the time was something like it's not possible to implement within mql4. I now agree with him to some extend.

The high frequency scalpers you're reading about are probably the same ones I've read about. Put simply most of their style is Not speculative. They either place high-frequency trade or high-volume trades when the opportunity presents it self. Example: when prices are mis-aligned (arbitrage). These guys do not use mql4 and the speed of light (it-self) is a limitation for them. They built servers next to the brokerage in order to beat the competition to the price.

Coming back to speculators and humanly possible. I've had the opportunity to observe some scalpers at work before. I think on average perhaps 4-trades per hour would seem ideal. You can definitely get algorithm's to place those numbers of trades (usually you'll use smaller time-frames). However, not everything those high-frequency traders said I believe. Like you can look for 5-pips profits with stop-losses of 2.5 pips. etc. But that's just Me.

Also #-of-trades and #-of-independent-events have to be distinguished. Two traders can trade the exact same system one place one-order lot of 1.0, while the other place 10-orders lots of 0.1. From my experience, high-frequency independent-events trades is a very difficult task to make profitable, in-the-long-run.

Thanks for that explanation. I started to suspect this when I started placing real trades.
 
marketpirate:

Hi, this might sound like a strange question but here it goes anyway.

If I wanted create a trading robot that trades as often as 'humanly' possible while still hoping to make money in this market what is the maximum trading frequency on the order of?

Is it thousands, Hundreds or Tens per day? Suppose I used some type of very aggressive scalping strategy that uses any currency pair. Could that kind of algorithm yield more than a few hundred trades a day?

In the first place, creating an EA that is profitable in the long term is hard enough to start with. Then you have the question of the spread. If you are running a retail account with a spread of say 2 pips, the price has to move by that amount before you can break even on a trade. You therefore can't make "high frequency" trades. In order to trade at high frequency you need a professional low spread or zero spread (commission based) account and low latency by getting a feed straight into a central dealing computer.

Assuming a random market, if you use a 4 pip stop and a 4 pip take profit with a 2 pip spread you will not win half the trades. More like 25% wins in fact, giving a profit factor of 0.333. But if you increase the SL=TP=40 pips then the random profit factor shoots up to 0.905. Thus trading on a longer timeframe reduces the odds against you due to the spread. You then need some sort of system to push the odds past break even and into profit - a non-trivial undertaking.

Reason: