The False Diamond Problem


You finally decided to listen to that Nigerian prince, and as it turns out, you’re the lucky recipient of a true random selection of 10'000 of the worlds diamonds! And you know that 30% of diamonds are fake/synthetic1. Yet, when you look at your brand new collection, only 10% of them are marked as synthetic. About 20% of your collection is lying about being natural diamonds.

No problem, you think! Time to go through the collection, marking the outliers. Problem is, you can’t tell them apart. You know there is a 20% discrepancy, but you can never tell if any individual case belongs to said discrepancy or not.

On a macro level, you know there is a massive error. However, on a micro level, no individual case will reveal itself as part of that error. There is a known discrepancy, but no good way to detect discrepant cases.

This is what I call the False Diamond Problem.

  1. UuUUmmM AchthualLy DiaMONDs caN’t BE FaKE, bEcAuSE dIAMOnDs ArE sTiLl DiamOnDs eVEn iF ThEy WeRE MaDE in A LaB.
    Also, the 30% statistic is completely made up. ↩︎