Acceptable risk reward ratio

Discussion in 'Trading' started by onelot, Nov 24, 2002.

  1. onelot


    i have been developing a strategy that has a win los ratio of roughly 48%... just wondering where my risk reward ratio should be at... currently it's around 1.8:1... is that average?

  2. This is governed by the formula for expectancy of a trading system. Do a search for 'expectancy' on this board and you will find more than enough stuff to guide you further. But from what you wrote, if by your risk (to) reward ratio you really mean that you risk $1.8 to gain $1, then you win-loss ratio is too low for your system to make you money in the long run.
  3. onelot


    actually it owuld be the opposite

    risk a dollar to gain 1.8 (that's usually the format i see people write it)... i have done the search and wasn't able to get enough info for determining if the above win percentage was an average to favorable system...

    thanks for your reply

  4. Pre-commission Daily Trade expectancy =
    [(win% x win size) - (lose % x lose size)] x average number of trades per day

    So, in your case:
    Expectancy = [0.48(1.8R) - 0.52(1R)] x average number of trades per day, where R is your unit risk

    Expectancy per day = 0.344R x average number of trades per day

    Let's play with some numbers:

    If you risk $100 a trade and have on average 10 trades, then your pre-commission expectancy will be 0.344(100) x 10 = $344

    If you risk $200 a trade and have on average 5 trades a day, then your pre-commission expectancy will be 0.344(200) x 5 = $344

    If you risk $1000 a trade and have on average 1 trade, then your pre-commission expectancy will be 0.344(1000) x 1 = $344

    The point is that there are many ways to get the same pre-commission daily expectancy...
  5. onelot


    ok so if the average is 4 trades a day

    then... .334R x 4 is the formula

    so if R is $75 or 1.5 ES points

    then .334(75) x 4 = $100.2 per day

    i'm actually doing more research on expectancy as we speak, but is this alright for a rookie system and a rookie trader behind the wheel :)

    assuming i can follow it of course
  6. Yes that's correct... now let's be even more precise... if you want your net, then you subtract your commissions and account for long-run average slippage... from the above (and indeed from your name :D) , I assume you are trading 1 Emini S&P contract, whose point value is $50... let's assume commissions of $5 round trip and lets also assume long-run average slippage per trade of $5... so your long-run cost per trade is $10...

    So, with your risk of $75 on 1 contract and an average daily trade frequency of 4:
    Net daily expectancy = [0.344($75)-10] x 4 = $63.20... so over the course of 250 trading days in a year (assuming you take a handful of holidays) you will expect to generate $63.20 x 250 = $15,800 a year... if your broker requires margin of $3000 and you put in an additional $3000 of "drawdown" coverage to make a $6000 'investment', then at the end of the year you will have $6000 + $15,800 = $21,800 in your account... and this $21,800 comes from a measly $6000 grub stake...

    So, to conclude, the resulting combination of your winning % and risk:reward ratio is very acceptable, if you are able to consistently maintain this over the course of the year... the dynamic nature of the market may necessitate strategy changes over the course of the year... moreover, some periods of the year are not as tradeable (Summer) so perhaps, on a conservative basis, you should reduce that $21,800 year end account by around 33% to come up with a more realistic figure of $15000...

    The main thing is that you have come up with a strategy which is positive expectancy (at least over recent times)... the name of the game in this trading business is longevity... it will take you about 2 years to gain exposure to most market conditions and to settle with a bunch of strategies to recognise and trade these conditions... until you have reached that stage of experience, statistical expectancies should not be taken as the be all and end all... what I am trying to say is that statistical expectancies only start to be meaningful after you have got long enough experience and are market-hardened... without this experience, psychological and other issues will get in the way of the actual $ manifestation of the theoretical expectancy...
  7. mark1

    mark1 Guest

    your system has a positive expectancy of 0.344
    that means for each dollar invested has a return of 34.4 cents.

    Although a pos. expect. is the base for a winning system ,
    I suggest to not optimize for the highest mathematical expectancy score achievable.

    There are other factors to be considered in real life :

    1) your personal MaxDD tolerable.

    2) numbers of trades* time :having two sys with the same positive mat.expect, the one with the highest number of trades * time, usually produces a smoother equity curve and a higher Recovery Factor

    3) High Recovery Factor( From Wealthlab web site:Recovery Factor is equal to the absolute value of Net Profit divided by Max Drawdown. Recovery Factor should typically be larger than 1. A healthy Recovery Factor is an indication that a system can overcome a drawdown.) my best sys has a RF of 20 !!!

    3) exposure: the less the best.

    Your system figures seem to be very good, but be careful not to risk high% on each trade cause given the nature of your system , statistically it can face a streak of loosing trades and you got to be ready.:cool:
  8. onelot


    i had been calculating the win loss using just my wins and losses... so where would breakeven fit in to this equation:

    do i just add that to the number of average trades a day... or do i include that in the expectancy formula:

    for instance [(win%(win size)) + (breakeven%(win size?)) - (loss % x loss size)] x average number of trades per day

    i guess those breakevens are gonna skew the results a little

    basically it breaks down to breakevens 16% / wins 39% / losses 42% Average number of trades now = 5

    I would think that i could just add it to the average number of trades...?

  9. You have lost 3% somewhere, since your probability weightings should add up to 1... I will add that 3% to your breakeven trades...

    Daily pre-commission expectancy =

    5 x [0.39(1.8 x $75) + 0.19($0) - 0.42(1 x $75)] = $105.75

    With average commissions and slippage of $10 per round trip, your daily cost is $50, so your Daily Net is $105.75 - $50 = $55.75

    The general post-commission, post-slippage formula for your numbers is [(0.282 x R)-10] x 5, where R is your $75 risk per trade...

  10. onelot


    ok, great thanks a lot you guys are very very helpful... very cool...


    my actual max drawdown is $1000... so i guess per Mark1 via WL that would make my recovery factor around 15 or so probably less taking the breakevens into account...

    i guess i should be excited, but it was good to read candletraders post about "market-hardened expectancy". that makes a lot of sense.

    what i take from this post is to not rely on just one system... good traders have a basket of strategies they apply to any given market condition... so i suppose it would be smart to have a variety of positive expectancy systems... but most likely your account size would have to be large enough to accomodate drawdowns for each of your systems... damn, this is such a challenge to get right!! what a great game.
    #10     Nov 24, 2002