Outlier in DNA matches

[neelawson]

My mother had 100% British heritage and my father had Germanic-Polish heritage. I've recently had a new 100cM match from the UK at MyHeritage. I assumed the match was maternal based on our shared matches as all seemed to have British heritage based on their surnames and their shared matches. However one who shared about 35cM with both me and my new match is Polish and so is everyone in that person's public tree. What could be the explanation for that? Frustratingly neither my new match or the Polish match have responded to my messages so I don't know if they have an explanation.

[Millmoor]

It could be a false positive. A lot of MH matches under 50cM from what I gather can be placed in this category. ( For example I have a match on MH where we clearly share  Scottish gg grandparents. One of our shared matches is Gideon Japhet, the father of the owner of MH. I am as sure as I can be that this is a false positive).

Added : I have 2 kits on MH, one is a kit transferred from Ancestry, the other being from MH. I only match with Gideon Japhet on the kit transferred from Ancestry, perhaps adding credence to it being a false positive.

William

[Glen in Tinsel Kni]

False matches on MH at 50cM are well known but there are documented examples at around 70cM,  the ethnicity results between sites are even less comparable than the shared cM figures.

I manage my kit along with a paternal half sibling but look at the results on the two sites and you would think they belonged to four different people. MH claims the shared cM to be much higher and predicts we are full siblings.
 

[Zaphod99]

Could you quote some sources for those statements about false matches please?

Zaph

False matches on MH at 50cM are well known but there are documented examples at around 70cM,  the ethnicity results between sites are even less comparable than the shared cM figures.

I manage my kit along with a paternal half sibling but look at the results on the two sites and you would think they belonged to four different people. MH claims the shared cM to be much higher and predicts we are full siblings.

[Glen in Tinsel Kni]

Several blogs and a couple of youtube channels have shown examples from their managed test results but they are a few years back in my history now and I don't have links to hand