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

    You're right, I should just focus on C# for now since it has real practical use for quantower.
     
    #521     Aug 24, 2023
    fullautotrading likes this.
  2. hilmy83

    hilmy83

    upload_2023-8-24_16-59-31.png

    Posting cause i've waited a while for a day like this (solid bearish day). At this point, the strategy has gone through a spectrum of market conditions and still function (since 2/20).

    upload_2023-8-24_17-5-24.png

    There is still a scenario in my head where I think this algo could fail. At this moment, it's at least a 1 of 126 probability? (trading days since I started this).

    I mean there is no such thing as a perfect strategy right. We have hedge funds managing billions still collapsing when that black swan hits. So I'm wondering when my "black swan" will hit..
     
    #522     Aug 24, 2023
  3. hilmy83

    hilmy83

  4. hilmy83

    hilmy83

  5. I suppose that depends on what's going on 'behind the scenes' intraday. If you're averaging down or doubling down on trades to make back initial losses, you may one day meet your black swan.

    Your equity curve on a daily basis is extremely smooth. Is it just as smooth if you were to post the equity curve of all your trades?

    Regardless, so far, very good.
     
    #525     Aug 26, 2023
    NoahA likes this.
  6. hilmy83

    hilmy83

    Yes, there are some position adjustment rules I built into the algo. My assumption is that when it gets into a range after sometime, it would breaktout eventually. So I have a sequence count as to when it increases that position size to take advantage of that breakout. It happened couple of times since I started.

    So the scenario that I have in my head is if it never actually breaks out enough to profit :vomit:. So you can see why this would fail in ES lol.

    This is the summary I pulled from MT5.

    upload_2023-8-26_10-5-42.png
     
    #526     Aug 26, 2023
    NoahA and Laissez Faire like this.
  7. hilmy83

    hilmy83

    Also keep in mind that during my MT5 implementation, I encountered so many issues that forced me to trade manually at certain times. I switched over to QT sometime in May. So the 'jaggedness' is also due to algo errors that I had to deal with.
     
    #527     Aug 26, 2023
  8. Quick comment as I have to run, but just by eyeballing that equity curve it seems like you've had several days where your intraday drawdown exceeds weeks of prior gains, yet you were able to recover and end the day green.

    So, I imagine your algorithm in some way caps upside (seems like your average daily gain is around $50-100) while you allow it to go deeply red and average down with increasing risk to end the day green.

    This can also be understood by your largest profit/win (1411 / 867) as your daily P&L never was in that ballpark.

    Would that be correct or not?
     
    #528     Aug 26, 2023
    shuraver and NoahA like this.
  9. hilmy83

    hilmy83

    Well that's a balance chart, not equity. So I'm not averaging down trades, but i do widen my reverse (stop) and target and also adjust the position to recover (not quite martingale, it follows a sequence after few reverses fail).

    But I do emphasize consistency over the large inconsistent profit targets, so the wins are capped as you say.
     
    #529     Aug 26, 2023
    shuraver likes this.
  10. NoahA

    NoahA

    I like the issues that LF is bringing up. I have always wanted to ask you about the individual trades as well, and if you wouldn't mind sharing, something like this.

    2023-08-26 0953.04.png

    It was more so for my benefit because what I am amazed by is how you can always end green. I wanted to see what your trades look like when you hit a rough patch, and if you pull off bigger winners, or average down, etc. (my snapshot doesn't show quantity, since its always just one contract, but if you could have shared yours, that would be awesome, and provide great insight, not into your strategy per say in order to not create liquidity problems for you, but where the consistency comes from and the thought process behind the algo's consistency).

    I am nobody to give advice, but from what I read, some do say that you can either trade to make money or to feel good. Of course its amazing to know you have a system that is always green, but having a system that makes more money is even better! If trading manually, its even more important to be green because if in a drawdown, its probably difficult to put on the correct trades and hold. But an algo would have no trouble doing this. So it seems to me like an algo would be perfect for this.

    What I imagine is that on some days, you could end up with $100 vs. +$30, but this would mean that some days you finish at $0 or -$20. But if there is a positive expectancy, then for sure it makes sense to do it like this.

    But I also see the merits of taking a profit and not wondering about what could have been. Its just that an algo, without the emotional bias, is much better suited for doing the thing to maximize profits vs. what feels good, which is what a manual trader needs.
     
    #530     Aug 26, 2023