Risk Management When Trading Spreads

Discussion in 'Risk Management' started by spread'em, Feb 4, 2018.

  1. spread'em

    spread'em

    I had a question for the more veteran spreaders out there - how do you approach risk management?

    Usually I take longer term views when trading spreads and will usually take trades where the reward outweighs the risk, for example 1:1.5 risk:reward but of course each trade is unique. Once a level is reached (target or stop) the trade is closed. Pretty mechanical.

    I know of some techniques which use stat-arb and in those cases the traders will often average in/average down positions. How does this approach work with regards to RM? Do you run stats that tell you your hit rate (say 70% winners, 10% scratch, 20% losses) and then work backwards from there tweaking certain variables?
     
  2. Expectancy: that is the magic word.
    It doesn't have to be all Kelly or Van Tharpie or even my personal fav. Al Sherbin. But simply paying attention to the probabilities of the $$numbers you're considering, and you're half-way there. But you have to work with a reward and risk number that reflect the expected value of the position.

    Sometimes, a cost-or-price is 100% known. Then that part of the story is over: it's 100% of whatever that cost (long stock? long vertical spread?) was. Maybe you *sold* a vertical spread -- your received price (in 'premium') is alllllll that you're going to get.

    But when you don't know part of the equation, DO NOT simply throw an unqualified target at the expectancy calc.

    Sorry. Closing minutes of the Super Bowl. Pretty good game so far, but trying to prep for Monday, too. And I don't think I did a good job here. :confused::(:cool:
     
  3. Neuroway

    Neuroway

    I view any trading system as a car. Leverage is the gas pedal, stop losses are the breaks, the balance between regression to the mean / trend following is the steering wheel, capital is the gas, the speedometer marks the % gain / period ... and risk management is the speed you want to keep the thing at. Of course, the faster you go, the more risky it gets.

    For me, a trading system that works must be able to ride the market whatever it's trending or not, by automatically (or semi-automatically) balancing itself between trend following and mean reversion.

    If your trading system works, it works. It may be a big fat slow gas guzzling, underpowered SUV or a hyper-nervous 2,000 HP Lamborghini, it's all the same. It'll take more gas (capital) to operate a SUV from point A to point B, but eventually, it'll take you there. A Lamborghini will take you as fast as you want from A to B, if you don't crash it or be caught by the cops first...

    So... The trick is leveraging your cage to a state where you feel perfectably comfortable at the wheel. Well... Perhaps a bit awaken if you like adrenaline and need something more than a comfy, slow ride. You can adjust leverage as often as you wish, just like the gas pedal of a vehicle.
     
  4. spread'em

    spread'em

    Thanks to a degree...I mean I know how to manage risk from various points of view (worked on a MM desk and traded spreads) so not really looking for any hints or tips per say so no need to be so allusive and mysterious in your answers...

    I'm more looking to understand how the other side manage risk, those who trade mean reversion/stat-arb spreads. I understand that averaging in is common, but would like to know how it is approached. For example, if you are trading in a 5 tick range, how do you determine your stop? Do you use VaR or something like that to measure the odds of a 2/3 StD move against you and use those odds to work back towards what your R:R should be? I'm just curious as I've never really traded from that POV.