Decline in number of SNPs tested

[jonga]

Can anyone explain the declining number of SNPs being tested, please?

When I had my first batch of Family Finder tests done in 2011 a total of 726183 SNPs were tested. The next batch in 2013 were Build 36 with 714549 SNPs and now my last in 2015, still apparently Build 36, only has 708193 SNPs, although curiously, it has two SNPs that weren't tested before, one at the start of Chr 1 (72017) and one in the X (93154922). This represents a loss of 17,990 SNPs, or 2.5% of the original total.

Does this loss mean reduced precision regarding matching or have the 'lost' SNPs proved unreliable or showed too little variation to be useful, or is it a cost saving reduction?

[dna]

Different equipment is being used. First Family Finder tests were on a different brand altogether (Affymetrix microarray). Then they moved to Illumina (and I think the tests were re-run), then to a newer Illumina...

These two changes were widely announced, but from your SNP counts it appears that the newest Illumina equipment is slightly different again.

It would be very nice to have these at least announced in one place by FTDNA.

W. (Mr.)

P.S.
There were other changes in the technology used by FTDNA. And I understand they reluctance to discuss such details, as after all almost everybody just wants her or his results. On the other hand, many look at their own SNPs and having different ones available depending on the time of the test is rather unnerving...

[jonga]

Thanks for the reply. I only discovered this because I import the SNP data into spreadsheets to make my own comparisons, especially on segments where there appear to be clashes between matches of known cousins, and found I had to add in the missing SNPs to make the rows correspond.