Shared autosomal DNA and relationship probabilities

[Tom Tramp]

I have tentatively identified 2 DNA matches as, respectively, a 6th cousin and a 5th cousin twice removed. According to my identification of them these two matches are not related to each other but they are related to me through the mother and father, respectively, of my great great great grandfather. What is surprising is that both matches seem to be far better than would be expected.
The first match (6th cousin) is estimated to be:
Shared DNA 0.4% (27.8‎ cM), Shared segments 2, Largest segment 17.3‎ cM
The other (5th cousin twice removed) is:
Shared DNA 0.7% (47.5‎ cM), Shared segments 3, Largest segment 28.5‎ cM
I understand that the average DNA shared for both 6th cousins and 5th cousins twice removed would be about 0.01%, and also read (on isogg.org/wiki/Cousin_statistics) that the probability of there being no detectable shared DNA at all would be about 90%. So my question is, can such remote family relationships possible contribute to so much shared DNA? Can anybody advise on how to work out the likelihood?
Thanks.

[Craclyn]

Plug the number of cM into the Shared cM tool to see the possible relationhships and their probabilities. https://dnapainter.com/tools/sharedcmv4

[Tom Tramp]

Thank you for the link. It is an interesting tool