Index futures automation

Can a fully automated trading strategy work in the long run?

  • YES!

    Votes: 56 67.5%

  • Hell naw.

    Votes: 14 16.9%

  • I don't know, I got my own trading to worry about.

    Votes: 13 15.7%
  • Total voters
    83
hilmy83

hilmy83

Starting this journal to answer the question: Can a 100% automated mechanical strategy work in the long term?

I have finally converted my strategy to a fully automated trading bot. I'm using metatrader 5 as the platform for now and running my expert advisor on a vps. The development process was a PAIN IN THE ASS. Took 2 months to finally get a finished program that works per the specs.

I'm running the program on mes, mnq, mym and m2k. Started at around 12k, but once account grows, I will make the transition to the emini contracts.

Started the program on 6/30. Will update every week.
week1.jpg


week1b.jpg


week1c.jpg
 
Great journal.

Can you provide a summary of back test results so we can get a feel for your system:

Number of trades per month or year
% of trades that were winners
Average win size vs Average loss size

Percentage of months that were profitable in the back test
 
hilmy83 said:
Can a 100% automated mechanical strategy work in the long term?
More...

Yes. But keep in mind that you are looking for opportunity at the margins. If you are trading deeply liquid securities and show a strong statistical advantage your trades will be copied and the opportunity will slip away. I'm sure you are using a broker and not connecting to an exchange directly, so you are agreeing to show your broker all your trades and all your results.
For this reason I do not believe you can trade highly liquid securities long term using the same strategy.
 
Millionaire said:
Great journal.

Can you provide a summary of back test results so we can get a feel for your system:

Number of trades per month or year
% of trades that were winners
Average win size vs Average loss size

Percentage of months that were profitable in the back test
More...

I didn't run the program through the typical extensive backtests. I basically just took all the ideas that didn't work over the last 2 years and tried to find common denominators. Once i got the idea down, started programming it and ran few weeks of historical testing and forward testing , just to make sure it works per the specs.

So the only assurance I have that this can be "profitable" are segmented historical manual trading results. But basically, the results were:
  1. daily trading (1-2 trades per session)
  2. around 80% success
  3. profit factor of about 1.5-2
  4. holding period of 1 hour or less
Main thing about my past failing strategies is the position size. I would start small and just seem to want to increase faster once i see few weeks of success. But I always end up burning the account.
 
Need three things:

1) Total trades
2) Average gain per trade
3) Standard deviation

Then compute the population standard deviation. You can't fight the math. If you aren't doing this then you are just guessing.
 
hilmy83 said:
Starting this journal to answer the question: Can a 100% automated mechanical strategy work in the long term?
More...

Well, let's hope it works for you and you truly print money.

But I come from the opposite end of the spectrum...thinking if something like this exists, or was possible, it would have been discovered by now by people much more smarter than you, and/or fully exploited already where it doesn't work.
 
MacBookProHo said:
thinking if something like this exists, or was possible, it would have been discovered by now by people much more smarter than you, and/or fully exploited already where it doesn't work.
More...

The hope is that there are some good intra day edges out there that scale to making just a few million a year at most. That the multi billion dollar hedge funds who employ those super smart people wont even bother looking for.
 
Last edited:
You must log in or register to reply here.
Back
Top