That's comparing Steppe R1b-Z2103* probably dead with modern R1b-Z2103 south of the caucasus. Obviously, they will not show as coming from somewhere else than south of caucasus, that should be obvious. After coming from Steppe, all those " basal branches " had a founder effect south of caucasus. Only ancient dna can explain you where it was, not modern dna.
I'm not so interested in dead lines, only the branch from which modern Z2103 developed. From SNPs, yfull estimates this modern Z2103 branched four ways in 3,500 BC; from STRs, I estimate it started branching sooner (4,700 BC).
You say "after coming from Steppe", as if that were a given:
1. It is important to define
what would have come from the Steppe - do you mean Z2103 in formation, fully formed Z2103 or various subclades of Z2103? (These could be separated by thousands of years)
2. Z2103's uncle clade PF7562 also seems to coalesce south of the Caucasus, indicating a good possibility that the whole family of living M269 derived from that region.
3. When would the founder effect of the South of the Caucasus basal branches have arisen? STR variability suggests this occurred very early (5th millennium BC, or 3,500 BC per yfull), certainly nowhere near the mid 2nd millennium BC when autosomal DNA suggests that EHG-heavy Steppe DNA arrived there.
Ancient DNA explains where ancient people lived, but we have no way of knowing whether any of these people were our ancestors. People actually ancestral to us might have lived somewhere completely different, so ancient DNA (like modern DNA) is just a guide, not the magic bullet that many seem to imagine.
Modern DNA can also act as a guide to where ancient DNA was. We used to realise that, but have lazily forgotten it. For instance, there are many precise branches of Z2103 (including at the basal level) within which the greatest range is exhibited in modern samples from Armenia. The quantity and precision of the data reflecting this is now sufficient to allow us to say with some confidence that this is not coincidence. The other explanations are that (i) living Z2103 developed somewhere other than Armenia and all remained together without any leakage anywhere else for thousands of years before migrating en masse to Armenia (highly implausible, especially as we know that Yamnayan early Z2103 was highly mobile and expansionary), or (ii)
living Z2103 began developing in the general vicinity of Armenia and that only one basal branch of it spread at an early stage into the Steppe and beyond). If scenario (ii) seems to fit most plausibly with the data, then it seems reasonable to hypothesise that it brought autosomal CHG with it as it spread.