Index futures automation

Discussion in 'Journals' started by hilmy83, Jul 3, 2022.

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

  1. YES!

    56 vote(s)
    66.7%
  2. Hell naw.

    15 vote(s)
    17.9%
  3. I don't know, I got my own trading to worry about.

    13 vote(s)
    15.5%
  1. hilmy83

    hilmy83

    Saw Pax's videos and learned a bit how he trades breakouts.

    Pax method is pretty much trading the ORB (30 secs) using at min 4 cars. He would set the stop just under b/e at a few ticks out (he calls it paying for the trade) and re-enter. After 2 stop outs, he would scale down by half. After 4 stops he would be out for the day.

    Once he's in a good trade, he'll scale out and leave 1-2 cars for possible runner

     
    #161     Dec 28, 2022
  2. hilmy83

    hilmy83

    Well, coming to the final days of the year. I'm still trying to test the updated algo against partial fill. It takes time to do because you can only get partial fill in live trading. And most days my TP are complete fills.

    But going forward, I'm basically going to trade just the MES for the full algo release (back to my original target market lol). Based on how I'm going "all out", MES would give me a lot more margin of safety. I can't trade multiple markets this way and MNQ is just so volatile and that 2x margin requirement add up quick.


    So based on my calculation i'm trading 1-2 contracts of MES for every $10,000. Tiny start lots. Typical day woudl net around $50-70, until the next account increment. it would net around $1000-1200/month to start.

    upload_2022-12-28_9-46-37.png

    But it's fully automated, so it's like having a passive income stream (that can potentially blow up lol). But the risk profile I chose can withstand 50 points on the open for S&P .
     
    Last edited: Dec 28, 2022
    #162     Dec 28, 2022
  3. hilmy83

    hilmy83

    The EA is finally complete now. It's able to maintain limit target partial fills

    upload_2022-12-29_10-2-36.png

    I even tested disconnected VPS scenario and EA resumed current setup as soon as connection is on.

    Now the only thing that will screw up is the strategy itself.

    upload_2022-12-29_10-5-52.png
     
    Last edited: Dec 29, 2022
    #163     Dec 29, 2022
  4. hilmy83

    hilmy83

    Plans for 2023:

    Plan 1
    1. Run the algo on just MES with 10k
    2. Top off account to 20k after 3 months of consistency
    3. Aim for 50k account end of the year (projecting around 200% total return minimum)
    4. Try to win the financial trading competition in the enhanced growth futures division (projecting about 150% based on 20k min capital) https://financial-competitions.com/
    Plan 2
    1. Blow up account
    2. Find the nth trading strategy that will work this time
     
    #164     Dec 31, 2022
  5. hilmy83

    hilmy83

    Plan 3

    Run on index futures (mes,mnq,m2k,mym) with 1 contract per 10k. And also adding b/e stop to get to flat quicker and minimize margin use. It's really a give and take situation that i'm still thinking over.

    upload_2023-1-1_13-50-27.png
     
    #165     Jan 1, 2023
  6. hilmy83

    hilmy83

    First trades of the year



    upload_2023-1-3_10-0-53.png

    So I decided that since I'm implementing a b/e stop in the EA, I might as well run it on multiple markets just so I can profit off larger sample size. Based on my backtests, my b/e stop would only give me about 40% success rate overall. This would give me about 2 out of 5 markets being profitable, while the rest being "breakeven"

    I think this combined with very modest starting lots should ward off sticky margin situation.
     
    Last edited: Jan 3, 2023
    #166     Jan 3, 2023
  7. hilmy83

    hilmy83

    I decided to edit the way i trade slightly based the premise that price movement is mostly a random event.

    https://www.elitetrader.com/et/threads/flip-of-a-coin-but.372081/

    I want to focus more on the tendency of price to retrace.

    Why do I think retracement trading works better than trying to predict entries or market bias:
    1. retracement is more objective. It's hard to say a market will have a trend bias or range. But I know market will move and i know it will retrace, that's it.
    2. Retracement is natural occurrence due to ebb and flow of orders (taking profits, new orders, etc.).
    3. In an algo, high volume driven market like the MES/ES, the periodic corrections is almost predictable during normal trading times

    The downside of this is when price goes into consecutive running phase with almost no retracement. I have few ways to manage that:
    1. Use manageable trading size to allow for the market to move until retracment happens
    2. Set early breakeven stops to exit as soon as the trade is caught in a run away price
    3. Set a max size as to not continuing building larger position
    4. Max time stop where EA closes all position at specific time

    I think combination of size, b/e stop, and maximum position can help navigate the infrequent times price do run. But I think as long as I trade in high volume periods, I can avoid situtaions like that.

    I think the only times where price is "less" random is during news releases, but most times (even in trend days), it's more or less just random.

    Anyway started using the updated algo only on MES. It's directionally neutral and I trade this during the high volume periods of the day.

    upload_2023-1-11_9-20-10.png
     
    Last edited: Jan 11, 2023
    #167     Jan 11, 2023
    NoahA likes this.
  8. hilmy83

    hilmy83

    #168     Jan 11, 2023
  9. Overnight

    Overnight

    It is because too many users of AMP have been looking the gift horse in the mouth for too long, and have not respected what they are doing for them.

    They are a futures broker with very little segregated monies, trying to support assholes who take them down to the bone. So they have to do this, increase margins beyond what is expected. No big whoop for responsible traders.
     
    #169     Jan 11, 2023
  10. hilmy83

    hilmy83

    Backtesting in MT5 is complicated when it comes to futures. SInce you have different contract months, you can't test continously (unless you use continuous contract). Also DST doesn't adjust automatically in test. And also I don't have a "holiday" filter added (i don't trade half days).

    So you have to do it in segments. But I finally able to do that since recently got a "reset" code added to the algo.

    9/12/22-11/2/22.

    upload_2023-1-13_18-7-55.png

    I
     
    #170     Jan 13, 2023