Hello,

I have a problem with applying custom factor which should select stocks with minimum 5% of mean intraday price range over some period. Below part of code:

class IntraDayMeanDifference(CustomFactor):  
    inputs = [USEquityPricing.high, USEquityPricing.low, USEquityPricing.close]  
    window_length = 20  
    def compute(self, today, asset_ids, out, high, low, close):  
        diff = (high - low)/close*100  
        diff = np.sort(diff) #sort results  
        diff = diff[4:-2] #filter some border values  
        out[:] = np.mean(diff, axis=0)  

The rest please find in attached backtest.
The problem is that based on my custom factor, pipeline returns stocks like Google or Paypal, which we know that are not so volatile.

2016-07-05  PRINT       dollar_volume   high_low_diff  latest_close  
Equity(45971 [AAL])    3.535842e+08       5.198106       29.33  
Equity(45993 [HLT])    2.024517e+08       5.217377         22.88  
Equity(46631 [GOOG])   8.382134e+08       5.683298        699.28  
Equity(46979 [JD])     2.828034e+08       6.038528         21.30  
Equity(47415 [SYF])    2.765042e+08       6.650303         25.33  
Equity(47430 [MBLY])   2.106279e+08       6.687589         47.12  
Equity(47740 [BABA])   6.974509e+08       7.033304         79.67  
Equity(49073 [LABU])   1.510457e+08      11.622539         30.88  
Equity(49229 [KHC])    2.018926e+08      15.681655         88.22  
Equity(49242 [PYPL])   3.059107e+08      16.513557         36.35  

On the other hand, research platform gives rational results when calculating this intraday price range.
Any help/explanation will be appreciated, thanks.