It's often quite common to see a citation say "BTW, we weren't able to reproduce X's numbers, but we got fairly close number Y, so Table 1 includes that one next to an asterisk."
The difficult part is surfacing that information to readers of the original paper. The semantic scholar people are beginning to do some work in this area.
