Backtesting is useless

Discussion in 'Options' started by JCDST1979, Jun 30, 2020.

  1. JCDST1979

    JCDST1979

    In the book Fooled by Randomness, Nassim talks that the history that unfolds is just one of the multiple possible ones, this makes backtesting completely useless, don't you think?
     
  2. southall

    southall

    There are plenty of hedge funds that have made money using backtesting.

    But as markets become and more random, the effectiveness of backtesting reduces.

    But I dont think the markets have become completely random yet.
     
  3. I think you may be taking an overly simplistic view! While the past does not infer the future, parts do rhyme. One can learn from history. IMHO, an important role in backtesting is to observe how your detailed trading strategy would have performed if applied to periods of history. This is valuable. The notion of backtesting to develop a trading strategy is one that should instill doubt and fear. (Attributes of curve fitting apply and relate to your "useless" comment)
     
  4. Turveyd

    Turveyd

    Simple things and back testing maybe, better off back testing only a few days all of your different methods in the presumption that it's made money last 3 days so good odds it'll make again tomorrow, then run that till it stops, while back testing the other 10 methods until 1 is out performing, then switch to that.

    I doubt anything 10years ago is relevant to todays market then the bits that are relevant likey better off flipping a coin.
     
  5. Only if you believe the markets are random and as a result unpredictable.

    And back-test what exactly?
     
  6. lindq

    lindq

    An experienced trader knows that strategies developed with backtesting are not timeless.

    Markets change, and the clock keeps ticking.

    But that does not at all negate the value of backtesting in the short to medium term. To call it useless is to denigrate an important tool that can save a lot of time and capital.
     
    drcruz, VenusMann and Space Alchemist like this.
  7. Atikon

    Atikon

    Have you ever heard of Earnings Management? Lookup how price pinning works with regards to street earnings. Most companies hit EPS down to the Dollar +1 Cent. Makes Earnings Upmoves somewhat predictable. Idc about Technical Analysis, but I respect the fact that ppl like Simons have figured it out. Also Backtesting is only one Part of developing a solid strategy, your game plan has to work going through robustness test as well.
     
  8. Atikon

    Atikon

    Forecasting Models for Volatility, Technical Indicators, Ranges/Deltas. Whatever you like
     
  9. JCDST1979,

    I respect your opinion, but if back testing is useless how else would you know if what you are doing will make money?

    I like back/forward testing because it allows me to practice my trading skills, day to day process, thinking, testing my discipline and patience, testing my risk management per minute, per 5 minutes, per trade, per day. Back testing test ME. Back testing keeps money in my pocket and keeps me form making future mistakes. It let's take what is on paper or idea and put it to test.

    Back testing improves my odds of winning. Backtesting is free practice.
     
  10. Sekiyo

    Sekiyo

    No it doesn’t.
    It can tell you what doesn’t work.
    No matter how small is the sample size.
     
    Last edited: Jun 30, 2020
    #10     Jun 30, 2020