Profit Factor

Discussion in 'Strategy Building' started by Vixyup, Feb 1, 2018.

  1. Vixyup


    1st post here been an advid reader... for all the guru traders what do you see more important profit factor or win loss percentage? I have noticed with tighter stops you can significantly increase profit factor which intern decreases win %. I know this is a simple inverse relationship, yet it seems critical... i know it depends on strategy etc but would luv someone to elaborate.
    .sigma and aldrums like this.
  2. ironchef


    Well, an amateur's (my) intuitive answer is as follows:

    A tight stops reduces your down side risk, what you pay for is occasionally those tight losses might turn into upside winners if you loosen up the stops. It is similar to putting a collar on your position. In general, lower risks mean lower returns (CAPM).

    I am looking forward to comments from professional traders.
    .sigma likes this.
  3. d08


    Win % is mostly useless. Profit, drawdown, profit factor are 3 metrics that matter the most. A lot of churn via tight stops will increase the amount you spend on slippage.
    Slope Trader, .sigma, Sprout and 3 others like this.
  4. Xela


    This - exactly.

    Funny how some subjects seem to catch the zeitgeist, in this forum. (Or maybe I just notice them because I'm thinking about them?). Expectancy matters. Drawdown matters hugely. I find PF very useful, myself, sometimes partly as a quick way of "identifying and excluding things not worth giving any more time to".

    Win-rate alone is a dreadful, mostly useless and very often misleading metric, in principle.

    My thoughts are what I said here, yesterday or the day before ... [​IMG]

    And welcome to EliteTrader!
    Last edited: Feb 2, 2018
    Sekiyo, Sprout, comagnum and 2 others like this.
  5. henry76


    People often use win % as a less obtuse meaning for profit factor , call it what you will , keep your wins as big as your losses and have more wins than losses ( and other way round , wins bigger than losses and equal number of wins to losses) and you won't go far wrong. You can keep your consensus of carefulness , and I'll keep my profits.
    .sigma and comagnum like this.
  6. qxr1011


    Regardless of how many tricks one plays with the calendar, the pregnancy still lasts nine months....

    Profit factor, shmofit factor.... - the robustness of the system that's what's matter ! And every trader feels with his guts how robust is his system - usually not much....
    .sigma and henry76 like this.
  7. The win/loss percentage doesn't mean anything when evaluated in isolation. Consider two trading systems A, and B, with 3 returns each:

    A: {+1%, +1%, -10%}
    B: {-1%, -1%, +10%}

    Notice that system A has a better win/loss percentage than system B, but clearly, system B beats system A by a mile, and any rational person should indeed prefer system B.

    Profit factor is a much better performance metric (compared to the win/loss percentage), because the way it's structured, it aims to evaluate the profit per unit of risk. But it's still problematic, because it completely disregards the shape of the return distribution. Again, consider two systems A and B, with 11 trades each:

    A: {+$100, $0, $0, $0, $0, $0, $0, $0, $0, $0, -$50}
    B: {+$10, +$10, +$10, +$10, +$10, +$10, +$10, +$10, +$10, +$10, -$50}

    The profit factor (PF = 2.0) is the same for both A and B. But the systems are very different, and PF fails to distinguish between them. Most people would prefer system B, because it has a better (i.e. "smoother", "flatter") distribution shape. By comparison, System A looks "suspicions", as its entire gain is due to a single return, which could be an outlier, not indicative of the overall performance.

    There are literally hundreds of other performance metrics, but there is no consensus on which one is the "best". The bottom line is, performance measurement is a nontrivial exercise. If you dig deep enough into it, you would encounter the concepts of stochastic dominance, utility functions, risk aversion, and all that good stuff.
    Last edited: Feb 2, 2018
    .sigma, Cat88, Sprout and 2 others like this.
  8. d08


    In my early years I was obsessed about win/loss percentage. It made the difference between a bad system and a good system for me. Nowadays I only started looking at it again because it's been discussed on ET. I believe it's actually harmful to look at it because if something says 55% success rate to you, you start second guessing things, despite the important risk metrics actually being excellent.
    Sekiyo, Cat88, Xela and 2 others like this.
  9. henry76


    The point is that when people mention win loss percentage they very often mean in your world profit factor , your arguing semantics about a subjective meaning , qx ( as above) calls it robustness of system , others call it Alpha , some profit factor etc , whatever you want to call this nebulous expression of some sort of positive outcome , it is the key and the only key.
  10. d08


    I'm talking about "the" profit factor which is gross profits / gross losses for all trades. It's very clear to me, there's no room for interpretation, am I missing something?
    #10     Feb 3, 2018