Looking for 5-10 min mean reversion trades

Discussion in 'Trading' started by Chris Paciello, Aug 30, 2009.

  1. I'm trying to learn and figure out how to spot and look at 5-10 min mean reversion trades on the SPY's and QQQQ's ... Any help would be greatly appreciated.

    Just trying to learn!

    Thank you
     
  2. the1

    the1

    Are you willing to try and pick exact tops and bottoms (very short-term) of the market and scale a little if you have to? That's pretty much the only way a mean reversion system works. This is an effective strategy that takes a lot of practice and not for the faint of heart. If you want to learn it just fire up a simulator and start throwing darts at upward and downward movement. Take notes on what works and what doesn't work. Look for <b>diagnonal</b> levels of support and resistance along with exhaustion volume patterns. From there it's mostly experience and judgment. As good a strategy as any but not easy to learn. But then, what about trading is?
     
  3. Not sure what you mean by "5-10 minute", but in my view it's hard enought to make money without putting a time constraint on things, so you may want to reconsider that.

    Otherwise, I'd say pick a moving average, then look at every chart for the past year or two, and for each day write down the max distance price got from the MA before reverting back to it. Look at your data and see if you see any patterns. Then moving forward, just play the probabilities. As mentioned above, some scaling may be required.

    Best of luck.
     
  4. Have a look at the ratio of double leveraged ETF's to their corresponding inverse. Do some math, and normalize for volatility. You'll find the "mean reversion" takes place over a period of days, not 5-10 minutes. I've tried looking at my pairs systems at smaller time intervals and the main problem isn't that you wouldn't have a profitable system but that you keep oscillating between the magic level and the fair value level. You end up churning your account waiting for the mean reversion to take place, because at such small intervals less than a daily timescale you keep going back and forth just as you would a regular sma crossover, so you need to look at either open or closing prices to get your mean reverting level. Then, you can't just exit because the value reverted. Imagine you'd calculated that the mean reverting level on SPY was 100, and it kept going between 99.99 and 100.01. At that timescale, you keep entering and exiting and eventually you'll be wrong, so if you come up with an indicator from that ratio I described to you, I think you'll notice that it needs to be EOD ideally. If you want more information, I'd suggest reading my journal bwolinsky trading that you can find by searching my name.
     
  5. slacker

    slacker

  6. Can someone explain what and how to spot them?
     
  7. The big problem is how to separate the lousy signals from the good ones.
    Some over-bot situations just lead to higher highs, and vice-versa.
    This kills a mean-reversion strategy that just uses an indicator.
    BTW: I don't agree with using Bollinger Bands as a indicator for mean-reversion signals.
     
  8. Thank you so much everyone! Really appreciate the replies!

    Anybody have any charts to show some examples?
     
  9. slacker

    slacker

    A suggestion, go to

    https://fx2.oanda.com/mod_perl/register/register.pl

    get a free game account and plot a Bollinger band on 5 minute chart for EURUSD and USDJPY.

    Watch for youself. Price action on r2m works the same in any market... Tonight may not be a good r2m evening however as there is a really nice breakout underway.

    Good luck!
     
  10. Well most simply put, a "mean-reverting" series has a negative autocorrelation!
     
    #10     Sep 10, 2009