I'm shown as having 477 DNA matches listed under my Discoveries tab from the link in the email.
FTDNA shows me 936 matches
23andme 1719 but the bulk of those are anonymous; only 200 open sharing; 195 previously public, not yet shared with; 359 sharing)
Ancestry too many to count, the VAST majority of them very small matches, although the latter does include some known relatives I've found as a result of a tree search having found them on GEDMatch with reasonable sized segments!
Top of the MyHeritage list are the two known relatives.
Following them the top six matches showed their largest segment match as being over 30cMs, and I only recognized ONE name.
Now I'm a double junkie, dna and genie, and work with the matches on all three companies for myself and a heap of family kits.
So I was (very) surprised.*
The report seems to be ordered by shared DNA in descending order ob cMs within the overall percent (with no immediately obvious way of filtering the report or obtaining a usable summary but I've only just started looking).
More 30cMs+ segments appeared further down the list.
Comparing results with other companies:
My maternal aunt is shown on MyHeritage as
Possible relationship Aunt
FamilyTreeDNA reports my aunt as
Total shared 1642.96
Largest segment 120cMs
Shared segments 49
Largest 124 cM
Looking at the name I recognized
Possible relationship 3C to 3C1R
Size 16cM (Gedmatch 16.5)
Another as I scrolled well down the list for familiar names:
Largest 16.25 (1 segment)
so that's comparable between sites.
So far, this has not imbued me with confidence that this will be a useful site, but it has to be said these are early days, and on the positive side, I could simply not have noticed all those new closer matches - well close by my standards outside the known relatives recruited to test :)
I do find MyHeritage trees completely and utterly unnavigable.
At least Geni has a pedigree view, as does FamilyTreeDNA although the latter does need the DNArboretum Chrome extension to be readable
And there's always WikiTree.com with it's wonderful array of DNA focused charts.
* Updated 7th Sep 2016
I started again at the top looking at the matches I didn't recognize, searching my projects and match data for anyone who may match them (I manage several projects and manage many kits).
First 2, not recognized anywhere.
The third, supposedly a 3C to 3C1R and sharing
with me might be someone who matches some of my matches on 23andme but hasn't appeared in mine unless anonymous there - which they can't be as that data came from the old CoA report with segment data.
Next is the first known match above, 31cM on MyHeritage 16cM on 23andme/GEDMatch
Then 3 not found anywhere, or insufficient information to determine if the same person
Then another 3C - 3C1R
Where a match could be found on FTDNA identified by the trees attached on both MyHeritage and FTDNA. But the match on FTDNA is to one of my distant cousins, not me.
Lowering my search criteria from 10cM to 7 brought up a kit I help with, not related to me, but still no match to me.
Next 2 couldn't find.
The next might be a 26cM match to a distant cousin of mine on FTDNA but no candidates presented themselves for me and no tree was attached to confirm the id.
Next 6 couldn't be positively identified
We're still at 20 cMs and up here.
looked to be found on FTDNA - but again as a match to another person, not known to be related to me.
Next couldn't be found, but the next looked promising.
Name unusual enough to be the one showing up on FTDNA as a match around 13cMs to three of my 2nd cousins. But no match to me there down to 7cMs
No tree on MyHeritage but a match on Ancestry (via GEDMatch) to one of the above cousins.
I gave up checking at this point, rather disillusioned about the discrepancies/accuracy of the "match" criteria on MyHeritage - but on the plus side, will follow up on this last match, even though only 13cM on FTDNA/16cM on GEDMatch, as anyone matching that set of 2nd cousins always shows promise to chip away at some of the brickwalls that got me into DNA testing in the first place.
** Updated 9 Sep 2016
(as I posted to the DNA Newbies list)
as saying that they're taking all data, whatever build, from whichever company, to the same base set of data, imputing the bits that they haven't received from the kit you've uploaded.
(so given a choice you should upload the one from the company with the most tested SNPs)
Surely if you've tested on all three companies, and can't find the top matches in any of your match lists, the imputations of the missing bits to this common standard for comparison of eg your FTDNA test uploaded, can't be correct if it doesn't match the values from your independent test at ancestry or 23andme.
I'll keep looking further down my "match" list to see if I get any better hit rate of recognition.