While calculating a series of molecules comprising P=O, P=S and P=Se bonds at the BP86/def2-TZVP level, I have found that the Wiberg bond orders are increasing with the increase in bond length from P=O to P=S to P=Se. While the Mayer bond order is showing expected decreasing trend with the increase in the bond length. What might be the reason for the anomalous wiberg bond order data?