Removing the golden handcuffs

Discussion in 'App Development' started by fan27, Jul 2, 2018.

  1. qlai

    qlai

    There is a difference between sharing strategy code vs sharing generic tool like back testing or charting indicator, right?
     
    #61     Jul 23, 2018
    MaxPastukhov and fan27 like this.
  2. this tool spits out 'profitable' strategies

    why would you want to sell that? t

    oh wait, because they aren't profitable

    got it
     
    #62     Jul 24, 2018
  3. While I don't believe in the future of automated trading, I still feel that there is nothing wrong with selling tools to those who believe.

    Here is the reason: there are numerous software development tools around. Prospective quants can either use them to develop something from scratch or find something more specific, polished for this exact task. It's like selling libraries/tools to software developers, or marketing tools to marketing experts.

    If there is a demand for something, there must be a supply of some kind. It's a free market.

    On the other side, I would not personally go into this niche because I don't believe. But, if I ever become a believer again, I will join the TS :)
     
    Last edited: Jul 25, 2018
    #63     Jul 25, 2018
  4. fan27

    fan27

    Incorrect. As outlined here, the tool adds automation to the strategy discovery process. It is not going to magically "spit out" profitable strategies. The user has to have an idea of what features might make up a profitable strategy.

    new approach for testing algorithmic trading strategies
     
    #64     Jul 25, 2018
    MaxPastukhov likes this.
  5. I stopped reading right there. You don't believe in the future of automated trading.

    Do you have any idea how much volume algo trading makes up today?
     
    #65     Jul 27, 2018
    qlai likes this.
  6. so it's just a backtesting tool where there is parameter optimisation?

    what is new about that? every tool i use does that
     
    #66     Jul 27, 2018
  7. Do you have any idea how much search traffic and advertising income does Google get today?
    Why don't we launch yet another Google?

    Newbie retail trader going into the quantitative niche has almost the same chance to succeed.
    As for "big guys" - yes, let's daydream about becoming the next Google. It's fun when you are in your 20th, but it doesn't seem like a great idea after you were inside all these bubbles, starting from the very first .com one.

    I'm not arguing about the fact that there are some small quants who are making money here and there. But, if you ask them if their life becomes easier over time, you will get the honest answer. For me betting on this niche is like trying to buy Bitcoin right now, following 2017 news and analytics.

    As I told previously, I'm not against quantitative trading per se, I'm against blindly following this approach. I believe that a human PLUS AI will always win against a pure AI, no matter how "smart" it may be. From my previous niche I know that everything automatable eventually loses its profitablility, becoming commoditized, up to the cost of the hardware it uses.
     
    Last edited: Jul 28, 2018
    #67     Jul 28, 2018
  8. traider

    traider

    Software development is also commoditized nowadays. All companies are investing in AI so they don't need programmers to write code for them anymore.
     
    #68     Jul 28, 2018
  9. The same here: software developer + AI will always beat any software development AI :)
    The main problem for software market isn't in AI, it's in the growing number of ready-to-use solutions. 10-15 years ago you had a chance to come into any niche, develop something from scratch in few weeks and get enough money to make a living out of it. Today software developers rarely develop anything brand new, they mostly support, integrate and extend existing systems.

    As for companies investing into AI, if you had some deeper knowledge in the niche you will know that it's not about software development at all. It's about expert knowledge and data, about inputs for algorithms that are relatively simple and can easily be standartised. That's why we see progress only in those areas that have big and reliable data sets. No number of software developers can replace that.
     
    #69     Jul 28, 2018
  10. traider

    traider

    AI doesn't need many software developers. Core team will be a few researchers and maybe 1 or 2 software engineers to code it. I would argue most software developers don't have what it takes to implement a proper AI system.
     
    #70     Jul 28, 2018