Has machine learning, AI killed killed short-term trading opportunities?

Discussion in 'Commodity Futures' started by helpme_please, Jan 26, 2018.

  1. Dear @d08 , I agree with @trader99
    Most people think that if you have 2 or 3 strategies in trading, you're set. After 2 years of trading, nothing is so far from the truth. You need 2 to 3 strategies every few months, just to survive in trading. Good trading is no different from being a good entrepreneur. You identify a niche, you invest in it, perfect it and you milk it till it stops working. Then when you sense that returns are starting to diminish you cross-diversify into an adjacent business segments and invest into the that and repeat. Trading is no different. However, what happens in years in running a business, just happens in months in trading. I think some of the trading gods like Jamison and Hall just didn't adapt into AI when they saw their techniques diminish. Who do you think are running machine learning AI's? Its other traders who started realizing that their antiquated techniques from the 80's and 90's weren't working.

    Regarding Drawdowns and optimism. I have personally suffered drawdowns of 15% and 20% on my portfolio. (15% was the same time when Hall closed his CL futures fund. 3% of my portfolio was on CL futures). Being a full-time trader/investor, this was depressing! This is usually the time when one is questioning their life choices. However, you just got to adapt. I scoured through the forums and searched for every posting on the "the perfect hedge" and experimented with this across 200 trades and lost over 2k. But after 3 months, was able to construct a series of hedges that "decays into profit" and still gives me enough insurance for downturn.

    So, in short! Stay optimistic but with a slight touch of fear!
     
    #11     Jan 28, 2018
  2. nickynoes

    nickynoes

    The fact that you need to change strategy every other month is a function of your own psychology. I have used and know people who have used strategies far longer than that.

    You lost 15-20% and got depressed, when you should already have known the chances of such a drawdown occuring. A successful strategy will have drawdown periods, this is not the time to swap strategies but to keep going and make adjustments if you need, but most often you wont.

    Changing strategy too often will probably hurt you in the long time, since it removes the ability to forward test and adjust the strategy.

    All in all, yes the markets change sometimes, mostly in regard to volatility and direction. Not all strategies work at all times, so you either need to develop numerous strategies to fit the different market type or have a strategy which has variables you can easily adjust according to the current market type.
     
    #12     Jan 28, 2018
  3. d08

    d08

    You make money in spurts when things are working well for a longer period of time.

    PS. HFT earns money from slower traders, retail and institutional alike, they don't cannibalize each other. When you enter, you're a target, when you exit, you're a target - no matter the timeframe, an order is an order for them. Yes, if they take you for 0.05% on a long-term 20% trade, it won't really matter but saying they don't affect longer term traders is wrong.

    This scene, although from a good movie about poker, applies perfectly. We are the suckers in this one.

     
    #13     Jan 28, 2018
  4. Hi @nickynoes,
    The point of that posting was more for inspiration than anything else and for the need to constantly stay hungry! And not be complacent on the raging bull market that we have been blessed with. FYI, I specifically call on my 15% and 20% drawdowns, because I it forced myself as a trader to go seek other strategies that has enabled my portfolio to fetch me 85 to 90% yearly returns and greatly minimize drawdowns. I removed the attachment as I think this is an invitation for bad karma! If you are interested, PM me, since it appears that you are on the path to grow your trading career!

    EDIT: Removed attachment ... :/

    You're right! 2 of my strategies (Medium term, 2-3 weeks) have been working for the last 2 years and probably will for the next 3+ years till we get hit with a recession. However, its foolish to think a good thing like this is gonna last forever. As you can see, DD after the Trump win (Nov'16) and the AMZN/Whole Foods buyout (Jun'17). The Drawdown depression (DDD, just coined a term,:) ) inspired to hunt for new strategy and backtest and forward test it every month. The hedge (AMZN positive delta to neutralize my consumer staples position) went off to become a core growth for my portfolio.

    Btw, depression is something that works into trader psychology. Doesn't matter you have a 100k or 1B of trading capital. It forces us to fix ourselves.
    Cheers!
     
    Last edited: Jan 28, 2018
    #14     Jan 28, 2018
    nickynoes likes this.
  5. nickynoes

    nickynoes

    Nice curve, are you risking more of your account lately or what accounts for the parabolic last month growth?

    The ability to recover after a DDD is an incredibly important characteristic of a trader. Making money is very easy, keeping it is hard, and coming back after you lost it with renewed focus is even harder.
     
    #15     Jan 28, 2018
  6. I've removed the equity curve. But the "parabolic" returns came twice prior back in May and another spurt in November. On a log scale, this would just look more linear.

    Nope! risk has always been 30% to 40% margin cushion. Around 2% of PUT insurance. 70% equity and 30% option trading. If you are familiar with greeks, Risk management-wise, portfolio has been net positive vega and theta between 0.5 to -0.5% of porfolio size. Gamma is admittedly negative 2% of portfolio (heavily weighted on 3 stocks waiting for a vol crush), but both upside and downside protection. There ya go! I think you can reverse engineer 3 of my strategies from those comments.
     
    #16     Jan 28, 2018
  7. d08

    d08

    I've been trading for 12 years now. My main strategy has changed but only in minor ways since 2007.

    If you need to reinvent yourself every 2 months then you don't really have an edge, things don't change that rapidly. Also, to expect to come up with 2-3 new strategies every few months is unsustainable. There simply aren't enough different approaches to have 10 or more completely new and unique strategies per year, every year.

    I suspect you do superficial testing for strategies, because to properly test you need to look at every angle, explain why it works, what are the pitfalls, test on multiple data sets, develop a sometimes-complex position sizing strategy, test fill quality, figure out which types of orders work best or can be used at all. All this takes time, after that it's observing how the strategy would be worked if implemented. Only then it's maybe, if everything checks out, live trading.
     
    #17     Jan 28, 2018
    Gambit likes this.
  8. In reply to has AI killed short term opportunities.
    Attached is the results of a machine learning bot capitalizing on short term opportunities with targets of 1% per day.

    The problem is those hedge fund mangers they are using old tech.
     
    #18     Jan 29, 2018
    indicator777 likes this.
  9. It's not optimism, its bs from paper traders.

    Real men trading real money claiming the EDGE's are not there, and some punks on a chat board saying how that's all wrong and THEY bank daily.

    lol
     
    #19     Jan 30, 2018
  10. No, some commodity funds have too much capital for the available liquidity. For traders with no size problem these markets have been good. Some examples from Price Action Lab Blog are in this link. Note all strategies are in equities. Commodities is a rough market. I think AI opens door to new opportunities.
     
    #20     Jan 30, 2018