Unfortunately it's a case of pick one and use it as your reference. Comparing the two will always result in a headache. The distribution of errors from a clinical test will pretty much follow a normal curve, so there is little you can do about it. Suffice to say that a clinical test would probably show you that your actual blood glucose level would be in the spread between the two.
That's life with fingerprick testing. You just have to accept that it is only ever sort of accurate. In addition, unless you test from the same drop of blood on both machines, you are prone to get different results as blood is not homogenous across even pricks on the same finger.
It's far more important that the results are consistent, and you can live safe in the knowledge that the mean "error" is never more than 15%.
In addition, at higher readings, the variance on the blood meters is generally higher. Sorry to be the bearer of bad news, but it really is a case of choose one and stick with it. Otherwise you are in the same realm as having multiple watches and trying to determine which one is telling the correct time...
