This is essentially what we think of as L23xL51, or the old L23* people. The remnant of L23+ L51- people appear to Z2103+ Z2105+. I'm using Z2103 as the "lead with" SNP because it is listed first by ISOGG, it is ordered as often as Z2105 and the R1b Ht35 (new) project admins use it as their "lead with" label. They are two separate SNPs though, and one day we might find out one is older than the other.
Today, we know of two good sized subclades under Z2103, L277 and L584. However, there also is a paragroup of true Z2103* people. In the R1b-Early_Haplotypes spreadsheet I label haplotypes that seem to fit into Z2103 with labels starting with "mml2303". The 23 is of course L23 and the "03" is just a short designator for Z2103. This is sorting purposes primarily.
Unlike M269*, we do have a lot of Z2103 like haplotypes out there. Z2103, frequency-wise, dominates the "eastern" R1b, if such a thing can be said fairly. Many people associate its modal with the Ht35 haplotype, or also call it the Armenian Haplotype. Perhaps it is the lesser (#s wise) equivalent of WAMH. Maybe it should have been EBMMH for R1b Eastern Black Sea/Mediterranean Modal Haplotype...
There are a couple of hundred long haplotypes for Z2103 so I think variance calculations on it are fair game. All I can say it is appears significantly older than nephew L11 (the WAMH folks) and probably, I think, than brother L51. One anomaly that I intend to dig into further is that part of what drives Z2103 variance is up is people associated with Armenia. I'm not sure if this means the mix is well balanced there and this is an origin or if this a crossroads geography... or maybe its just that our Z2103 everywhere else is too limited. I wish we had Greece and Calabria better tested.
One thing I've noticed is that 406s1 splits Z2103 fairly nicely. Has anyone else noticed this? L277 appears to be a subset of the 406s1<=10 side of Z2103 and L584 appears to be a subset of the 406s1>=11 side of Z2103.