Profit By Combining RSI And VIX

About the Author Jeff Swanson

Jeff is the founder of System Trader Success - a website and mission to empowering the retail trader with the proper knowledge and tools to become a profitable trader the world of quantitative/automated trading.

follow me on:
  • Would love to see a newbie crash course or a brief Trading for Idiots report!

  • Thanks much Jerry.

  • George says:

    Jeff, another great article!
    I have Connors books and have been testing and coding these in MC for the last few months.
    I have one question for you. In your latest article you state that:
    “Is this a complete trading strategy that you can trade with your own money? Probably not.”

    Besides the lack of stops, which I completely agree with…, what else does not make this or other Connors RSI system incomplete? The Profit factor, average trade size, % profitable trades, etc… seem quite good.

    I’ve tested these and performed confidence interval testing (assuming normal dist, which might not be correct), but in any case these systems seem quite good.

    What do you see that is missing?

    Thanks for the great website.

    George

    • George, the biggest thing missing is probably the stop loss. Most of the other factors look good. A position sizing method could also be added but more importantly, you could use TradeStation WFO to test the robustness of the input values further. While it’s true we did this to some degree in the article, a WFO cluster analysis would provide an even more robust test. Also, using a Monte Carlo simulator over all historical trades will give us a better idea of potential drawdowns and returns.

      Glad to hear you are enjoying the website!

      • Mark says:

        I’m a few months behind in your posts, Jeff, but I’d like to see a post discussing WFO and your take on it. I’ve tossed in a comment or two on WFO in previous articles.

        • Mark, that’s a good idea and one I’ve been throwing around as well. It’s on the do-do list.

  • George says:

    Jeff
    Thanks for your reply, I never considered using a Monte Carlo routine to stress the drawdown – that’s a great idea.

    I’m not sure what Tradestation WFO is. I’m using MultiCharts and have had no problems importing the ELD file.

    Thanks Again
    George

    • George, WFO is TradeStation’s Walk Forward Optimizer. If you have access to a Monte Carlo routine that would be a big help.

  • choubix says:

    Hi, came across this post while looking for trade ideas and strategies on volatility. What underlying do you trade when RSI hits the tresholds? Tks

  • Nice work, I’d been working on VIX signals recently, but not combined with the RSI. I think one of the risks with using the VIX for entry signals is it’s ‘recent suppression’ thanks to QE and the growing popularity of the VIX as a trading product. In any case, the backtesting points and time threshold parameters you talk about, are well made.

  • Andrea says:

    Good morning,
    in my opinion, to stress the drawdown is sufficient you test the system from 1/2006 to 1/2010
    The result will not be as exciting.
    I’m italian, I’m sorry for my bad english
    Bye
    Andrea

    • Andrea says:

      I’m sorry,
      I deepened my studies, I apologize for my previous opinion and I offer my congratulations to Jeff
      Bye
      Andrea

  • Mike says:

    Jeff – Can you do an analysis of this strategy as a bearish trade using the opposite criteria?

  • Thomas says:

    Hello Jeff,

    thank you very much for your great concepts here. You wrote,
    “the biggest thing missing is probably the stop loss”.
    Isn’t it typical for this kind of mean reversion strategies,
    that the MAE is a big multiple of the average profit? If
    so it seems very hard to find a adequate stop position.

    Thanks

    Thomas

    • This can be a problem. Some people will use other means to hedge such as buying options. You can limit your risk by trading only a percentage of your equity, as done in the example code for this article. Yet it’s true, stops often hurt this type of system. Exploring the MAE and setting a hard catastrophic stop may give provide some peace of mind.

  • John says:

    Jeff – Thank you for sharing this analysis. I found your site today and your info is very interesting! May I ask…what securities did you use for your backtest to 1983? VXI or VXO? S&P 500 Index?

    Thank you.

    • That would be the S&P cash market, and VIX. But that 1983 date is a bit confusing as you may notice the first trade is not until 1991 when the VIX data became available. I should probably update the text to match when the first trade was taken. Thanks!

  • Marco says:

    Hello Jeff! I think adding maybe a 2,5*ATR5 stop loss would definitely put the last improvement to the system and make it tradable with a good leverage.
    Could you make a follow through post?

  • Marco says:

    I said 2,5 but it could be fine also 3 or 2, we would need to explore the range.
    http://nightlypatterns.wordpress.com/

  • Alex Argyros says:

    I look forward to your answer to Marco’s questions. In addition to that, could you clarify when entrances and exits are made in this model: intraday or eod?

    Thanks for the excellent work.

    • Hello Alex! The entry and exists happen on EOD on the close.

      • Alex Argyros says:

        Thanks for the reply, Jeff.

        Could you tell me if, with a reasonable stop in place, you would consider trading this strategy? If yes, well, that would be it. But, if no, could you explain why?

        • Hello Alex, I think with a reasonable stop you could turn this into a trading system for SPY.

  • James says:

    Great article Jeff.

    Finally got round to replicating this myself and get similar results, except for the 2 sell threshold test charts — where you test different RSI values at which to exit. I’m getting the exact opposite…better overall profit and profit-per-trade when sell threshold is low. It makes sense (to me anyway!) that an exit on a low Vix RSI (bullish S&P) would be a better exit.

    Would you mind checking and confirming to me that I’ve got this somehow wrong? Cheers.

    • Alex Argyros says:

      James,

      The sell threshold is based on the RSI of price (i.e., ES), not of the VIX.

  • James says:

    Thanks Alex – I was using the rule from an older version of this article where the exit rule states “Exit when RSI(2) of the VIX falls below 30”. It’s since changed to become “Exit when RSI(2) of the price rises above 65”.
    Testing both versions give pretty much the same results as one would expect.

    It’s also good to see that this strategy has continued to perform well up to the present day.

  • rafael says:

    i trying cfd and have afraid by stop loss on the other hand i read rsi2 the other article you wrote i have doubs wich one is better in your experience i need to knwo a litle more for aurora pro trading
    thanks

    • I really can’t say which one is better. They both have their pluses and minuses. You best bet is to test each one and determine what best fits your personality.

  • Jarrod says:

    Jeff, I love all the work you have done testing Connor’s strategies. Thanks very much for sharing. I have just done some testing with this RSI VIX strategy myself. Although I have done it all manually. The results are really nice. I have been using RSI2 in my trading for some years now on individual stocks, but trading the SPX is becoming attractive for me. Just a point I have noticed; you haven’t mentioned in your rules about only taking trades when the VIX opens higher than yesterday’s close. Even if you haven’t included it in your rules, I doubt the results would drastically change anyway.

    Keep up the fantastic work!

    Jarrod

  • Mash says:

    Really liked this idea. I was wondering if I can use this to Trade Emini futures? Like Nasdaq, Dow Jones or Russell emini

    Regards,
    Mash

    • I would guess so. I’ve not look at it on those markets in a while but in general, I would think it would be a workable concept.

  • >