It is not a completely general problem. In the field I work in (computational physics) we very often reproduce results published by other research groups. It's

It is not a completely general problem. In the field I work in (computational physics) we very often reproduce results published by other research groups. It's very common in our field to see plots in a new paper overlaid with data from old papers showing where our results agree (where they disagree is not a reproducability issue but is instead due to our ability to do more accurate calculations). Here is a paper where some colleagues of mine wrote a benchmark paper explicitly to check various methods against each other (with good agreement): https://journals.aps.org/prx/abstract/10.1103/PhysRevX.5.041041 https://journals.aps.org/prx/abstract/10.1103/PhysRevX.5.041...