See below.

Russ Maurer
Pepper Pike, OH

Itzhak wrote

I was looking for FTDNA FF database size data but could not find it.....

The link is near the top of the FT welcome page. As of now, the total
number is 643,537. They break out certain categories (which add up to
more than the total because of people taking multiple tests and so on.)
They have 480,915 Y tests of all sorts, but they do not break out the
number of autosomal test records.

I will test with 23&M soon and will compare the results. I
have 1,940+ matches which is typical for a fully eastern Ashkenazi
Jew. When I see the equivalent 23&M report, I will be able to
compare more fully.....

I am also fully Ashkenazi. I have 1019 matches on 23 and 1573 on FT,
but I'm not sure the comparison is meaningful. 23 seems to cut off
the list around a thousand for everyone, taking off some old matches
as new ones are added. I have not seen this acknowledged by 23, but
it has been mentioned on some forums and is certainly true in my case.
I had something like 994 when I first got my results 6 months ago, and
way more than 26 new ones have been added to my list since then. I
certainly hope the removed ones were the weakest matches, in which
case I don't really mind, but I have no way to be sure about that.

Also -- as I read in other forums, FTDNA customers are more into
genealogy while most 23&M customers are mostly interested in the
medical reports. People found FTDNA matches more likely to respond
to genealogical correspondence.....

I can't speak to FT, but it has been my experience that the response
rate to my genealogical inquiries on 23 has been less than 25%.
Undoubtedly some of this is due to people whose interest is health
rather than genealogy, but users who don't care about genealogy can
opt out the family finder. So I have assumed that anyone whose name
appears on my relatives list is interested in genealogy. It may be
that some people have adjusted their settings so they don't receive
email notifications of inquiries, and then they forget to log into
the site to check on their profile. Despite the large number of my
matches, I have received few inquiries >from other users. Also, 23
seems to attract people looking for biological parents (adoptees,
donor children), for which autosomal DNA analysis is clearly best.

One significant difference in the 23 vs. FT experience is that FT
provides you both the name and the email address of all matches as
the default. 23 provides the name only if the user chooses to make
it public, which a small percentage do, and does not provide any
email addresses. The only way to initiate contact with a match is
via a message through the 23 messaging system. If the match doesn't
respond, you are pretty much stuck. I think there are some trade-
offs here between utility and privacy which each person will view
differently.

Autosomal tests (FTDNA'a at least) overestimate the relation
proximity for Ashkenazi Jews. mtDNA tests are almost useless for
genealogical research. Y-DNA tests, while limited to one line, are
most reliable....

I disagree with you on this. Each kind of testing has its place. Y
DNA (also mtDNA) works by a different principle than autosomal DNA as
far as analyzing ancestry. Y DNA is passed along, essentially
unaltered, >from father to son. Occasionally, mutations (changes in
the sequence) take place, which serve as a sort of generational clock.
The more differences between two Y chromosomes, the more generations
separating the two individuals, on average. However, it's a clock
that ticks irregularly and unpredictably. If two males are a perfect
match on a Y-test, they could be brothers, or father and son. They
could also be distant cousins if it happened that none of the analyzed
sites mutated in that family for several generations. Similarly, if
two males had only one difference on the test, you couldn't tell if
they are first cousins or more distant. Obviously, as you say, it
gets easier if more relatives get tested, but few families have more
than 3 generations alive to test - and most of the interesting genetic
action probably took place in the inaccessible past. So it seems to
me that Y testing is a poor tool for certain genealogical purposes,
especially resolving recent family relationships.

With autosomal DNA, only half of it gets passed >from parent to child,
and in each generation, the segments get scrambled by recombination.
Thus, in contrast to Y-DNA, whose "signal" remains strong over many
generations (allowing you to plausibly finger someone as a "sixth or
seventh cousin or thereabouts"), the autosomal signal of relatedness
gets diluted with each generation, so that by the time we're talking
4th or 5th cousins, the signal is faint or vanishes entirely. Yet
that's a big part of the basis of its utility for sorting out close
relations--1st, 2nd, 3rd cousins, aunts and uncles. The "clock" for
autosomal DNA is the dilution of the matching (and also the shortening
of matching segments due to recombination), not the occurrence of new
mutations. For someone trying to fill in 2nd and 3rd cousins on a
family tree, autosomal makes the most sense to me. As always, it must
be backed up with conventional sleuthing. The proximity overestimation
you refer to due to inbreeding does indeed occur, but allowance can be
made for it and it does not affect the basic utility of autosomal
testing.

BTW, Russ. How many relatives did you discover via 23&M?....

Over a thousand matches, as mentioned above. In terms of the number
that I've been able to place on my family tree, it's currently zero,
but I'm very close on a probable 3rd cousin. I know which of my great
grandparents he relates to, we just don't have the common ancestor.