Join by "Largest Overlap" doesn't work as expected

[Scoth]

Maybe I'm missing something, but "Largest Overlap" doesn't make sense. Sometimes, joining features get joined to multiple target features, or multiple joining features get joined to a single target feature.

If it's based on the LARGEST overlap, shouldn't a feature only get joined once... to the target feature it has the largest overlap with? "Largest" means the most, not just a lot of overlap.

[Libra66]

I'm having a similar problem. The target polygons are getting joined with the smallest overlapping join polygons instead of the largest! Did you ever get a solution?

[James_Whitacre_PGC]

Trying to revive this topic...

I have experienced the same issue as @Libra66 when my data is in a geographic coordinate system. Once I projected it to a projected coordinate system, it identified the correct largest overlap. Is there a bug here? Maybe the tool needs to be adjusted for geodesic area? 

[RileyAnderson]

2025 update: join is still wildly wrong if the projections are not the exact same between the target and input layer. It will assign the smallest overlap or a neighboring feature. After ensuring projection is the same, tool works correctly.