http://www.investopedia.com/articles/financial-theory/11/calculating-covariance.asp
如果你只看一天结束时的价格（为简单起见），并且你想将一只股票与所有其他股票进行比较，那么试着将其他股票都视为一个元素，即：

n=number of days to analyze
marketReturn = average(return of each stock that isn't APPLE)
marketAverage = (sum of marketReturn's)/n
appleAverage = (sum of APPLE returns)/n

然后在你分析的天数的样本大小上应用总和。

Answer = (((day 1 APPLE return %)-(appleAverage))*((day 1 marketReturn %)-(marketAverage)) + ((day 2 APPLE return %)-(appleAverage))*((day 2 marketReturn %)-(marketAverage)) + ...)/(n-1)

我想如果你想单独检查每一只股票：

for(x=0;x<numStocksNotAPPLE;++x)//increment through each stock
{
    covariantList[x]=(((day 1 APPLE return percent)-(appleAverage))*((day 1 stock[x] percent)-(stock[x] average)) + ... + (((day n APPLE return %)-(appleAverage))*((day n stock[x] %)-(stock[x] average)))/(n-1)
}//that should give you a covariant APPLE vs each stock

如果你只想要一个数字，你可以取协方差列表的平均值：

foreach(covariantList c)
{
    sum+=c
}

averageCovariant = sum/covariantList.size

请记住，这是所有的psuedo代码，我不声称它是任何功能