Math/Stats theories sharing

Discussion in 'Strategy Development' started by j2ee, Jun 14, 2013.

  1. j2ee


    Other than Monte Carlo, walk forward, Sharpe ratio, what else is out there to build a better trading program?
  2. Sergio77


    Imo none of the above build better trading systems. Fitted systems give excellent MC results, Sharpe ratio punishes good performance and walk forward is just optimization. The best test is that your strategy must work across many instruments without modifications
  3. j2ee


    I think there are two reasons that changing the parameters in a trading system make sense:

    1. Fundamentally different markets like stock comparing to commodity would have very different. Then same parameters would have very different result even the trading system has some level of predictive power. One of the main different would be many commodities really affect a lot by the weather while stock market is not, then this reason would affect how a trading system should be made for a market.

    2. A future with like 20000 point comparing to a stock like $5 dollars would definitely need different parameters to apply the same ratio of math to calculate, as long as you are using the price to calculate in your trading system. 1/20000 is very different from 1/5, then no way applying exactly the same formula. As least need to change the parameter base on ratio.
  4. Eyez


    Yea, lets share working quant strategies with complete strangers online, excellent idea. You first
  5. j2ee


    your comment is off topic.
  6. Sergio77


  7. Humpy


    sorry to pour cold water on your parade but past events have little or no predictive powers to forecast future events.
  8. IMO, a big part of trading is all about coping with the "little" you refer to above, i.e. very low signal-to-noise ... so overleverage is often the villain ... one needs to be sure one can still be around for when the good times start to roll again ...
  9. j2ee


    welcome to prove it.
    #10     Jun 18, 2013