Here is the TradeStation code if anyone is interested. I am getting similar results to mcshort so I assume that I got the code right for the most part. If there are any errors, please feel free to point them out and I can repost it. I will also post the trades since Jan 2000 for SPY and QQQ. I'd be interested to know if I got the same ones as you mcshort. Since it doesn't trade very often, the sample size is really too small to judge robustness. The test was run using 100 share lots and deducted 1 cent per share commissions. Due to the small quantity of signals, the returns are small, but of course they beat the heck out of buy and hold since 2000. Also, I am only taking 1 position with this code, so the exit is coming at the 1 ATR target level. You would need to add the multiple exit portion to get the same trades that mcshort is getting.
Trades in the QQQ since 01/01/2000. (Oh, BTW, technically the trades start 50 bars into 2000 since that is required to get the 50 EMA...my data only goes back to January 2000.)
Yes for the most of the SPY trades they are the same except I got a short trade on 7/9/02 and a long trade on 9/16/03 and on 12/4/01 I didn't get a signal. All the rest were the same though. I agree just going back to 2000 really isn't going to make it robust. I went back to 1980 for the SPX and the winning % is about the same and it came up with 90+ trades. Still not a heck of a lot though. I think if you let the 2nd part run is when the return will get bumped up. Eric
Ok, glad its close. Looking at those dates, the long trade in 12/01 qualified with only a slight dip of the RSI (its probably just small differences in our data that caused that), and the two others that mine missed would have triggered if I took the signals after the RSI crossed back inside the BB's. As it is coded, if the breakout occurs more than 1 bar after the RSI retraces back inside the bands, then no trade is taken. But for the most part this doesn't come into play very often.
LOL...for only $325 you too can have the secret of the bands. Sheesh... Here's a freebie to save you 6 and a half ES points. Take the part of my code with the plot statements inside the squiggly brackets and use them in an indicator, that will give you the "new" RSI (RSI with BollingerBands from the looks of that chart you posted). Then plot some bands like moving average bands or Keltner channels, set them up using some funky colors or plot them as "crosses" and viola, you have the magical band charts. I am constantly amazed at the crap people will pay for...maybe I have been on the wrong side of the business all this time. :eek:
Wow that strategy from Band Traders sure is hard to understand. Although, I guess they just sell when price hits a top Bollinger Band and buy when it hits a bottom band... with a simple exit strategy. I wonder if using something like Keltner Bands, or moving average bands instead of Bollinger Bands would make shorty's strategy better over intra-day. Or maybe instead of buying above the 50MA and selling below it, we could try using the 50MA to determine direction and then only entry in the same direction as the 50MA.