Fully automated Portfolio trading system looking for partner.

Discussion in 'Automated Trading' started by Sun Light, Sep 23, 2021.

  1. Background:

    I am not sure if this is the right forum to put it out, but is my best guess so here it is.
    In last few years, I have developed a fully automated trading system to trade Stocks.

    It starts early morning, updating database, earnings events, news etc. Then it submits several orders before market opens, valid for that day. Once trading starts, it scans entire market periodically and enters new trades.

    On average it takes about 15 new trades each day. Typical holding period is 3 to 6 weeks. Typical number of unique stocks held in portfolio are 350. Exists are calculated automatically.

    There is High level flags for leverage, long to short bias, risk, sector weight etc.

    It is close to 98% fully automated. I do not think any portfolio management system with hold period of few weeks can be 100% fully automated.

    The 2% manual intervention it needed is on Setting up high level flags (changes in few months), managing about 2% of portfolio which is in severe loss (checking news, Human Fundamental and sentiment analysis) etc., Manually reducing leverage when needed by closing positions.

    What I am looking for:

    I would like to take this product to next level.

    I want to partner with right professional who brings complementary skillsets to the table. Professional with Industry experience with Fundamental analysis, trading experience for big firms, has connections to find funds, Regulations. You need not have technical skills, but if you have it is a asset.

    The product is developed to trade portfolio level with Risk management as a core. I will be checking my private messages as well visiting this forum once daily.

    References, proofs, more details of systems will be exchanged privately as needed.
    questions and suggestions (other ideas to take this to next level) are appreciated.
    Leob likes this.
  2. You asked too much but did not mention how wonderful your system is and your own background
    VPhantom and Sun Light like this.
  3. What is your definition/expectations for wonderful system?

    Background: Software Engineer (20+ years) and trader. Only Stocks listed in NYSE/NASDAQ. No options, no forex, no futures.
  4. deaddog


    That's a lot of stocks. Does the system outperform the index ETF's?

    Can you throw some recent results or backtest numbers at us?
  5. traider


    Are you offering strategies or infrastructure?
  6. Yes, that is lots of stocks. (it's a good question asked) Only with high numbers, the signal/noise ratio improves. in spite of studied electrical Engineering, it took me good 10 years to actually realize it in trading/system development and that's a shame.

    Yes it outperforms Index ETF. I think that should be given.

    backtest numbers will be for individual components that goes in the system. It can never be identical to actual performance of entire system, in Backtesting or Realtime mode.

    Imho, it is kind of myth to look for great Sharpe ratio or similar metric. and this is with real trading experience of trading individual systems with great back testing Metrics, only to see that they do not perform that well in Real time.
    Last edited: Sep 23, 2021
  7. Not sure what you mean. I am imagining a Technical partner(me) + Business side Partner (you?) to take this to next level. If you are asking if strategies OR signals for sale then NO.
    Basically, there is no singal strategy. It is at portfolio level, combination of many strategies, Fully automated. I do not think any single strategy will work as good as the whole team.

    Let me explain that better. System = n * strategies => many entry signals.
    Let say system generates 40 signals on a day and it is able to accumulate 20 new entries at a price it is looking for. One of them is AAPL. you liked that signal and you too that entry with 20% of capital. and it tanked. You will not like that now. System is ok, as it expects some of it to tank. That is part of risk management at system level. Point is, it can not traded as individual signal. Hope that explains.
    Last edited: Sep 23, 2021
  8. Drawdown, profit factor, win rate, holding period, long only or neutral,etc?
    gkishot likes this.
  9. Drawdown: good question. less than 15%

    Win rate: at portfolio level it is less important metric. As most of trades will end in breakeven or many in small losses and few in big wins.

    profit factor: again at portfolio level it get's complicated. It beats index ETF (as a system) is good enough info here at public forum.

    Holding period: good question. 3 to 6 weeks average.

    Long/short/neutral? : All three. When to go long-short as 100-0 vs 80-20 vs 50-50 vs 40-60 is decided by system. With little intervention of Manual flags at system level. AI is not good at detecting reversals. (I am hoping the other partner with Fundamental analysis aspect and street experience will add value here).
  10. deaddog


    Yes it should be a given, but unless you ask you never know.
    The only metric I look at is equity curve compared to a buy and hold of SPY.
    #10     Sep 23, 2021
    Sun Light likes this.