Frequency is NOT a good indicator of origin. Diversity is a better indicator as well as presence of earlier branching subclades.
That's what we see in Anatolia, the Balkans, the Caucasus and the Near East.
I agree with you. I think it's highly probable that R1b appeared somewhere just South or East of the Black Sea. And yes, frequencies change with time. What I find strange, however, is the occurrence of a founder effect with peaks of 90% R1b (Ireland, Wales, Basque Country) in Western Europe. Why don't we find "fellow" Anatolian J2, G2a, J1, E, etc, in these areas, in at least remotely similar frequencies to R1b, as it happens in West Asia?
Drawing a parallel, it's as if Latin America (which has Y-DNA haplogroup frequencies very similar and proportional to Iberia's) had 50% J2, with regional peaks of 80%. With J2 being about 10% of Iberian lineages, what are the odds of such unbalanced founder effect? Unless we assume, returning to the R1b topic, that Western Europe was overtaken by one random R1b man with his sons and brothers, and by their descendants. Which, I guess, isn't impossible, as a gradual process.
Nothing is 100% absolute, but I would not let the exception (the small minority of R1b in Europe speaking non-IE) confuse the vast correlation of R1b, specifically P312 and U106, with IE languages across almost all of Western and Central Europe.
We don't know. We are just speculating, but I would not focus too much on the anomalies.
The thing is, non IE-languages were not spoken exclusively in the Basque homeland, but in a much broader area, including all of Mediterranean Iberia as well. And as linguistic evidence suggests, those regions spoke not one, but two apparently unrelated non-IE languages (Iberian in the Mediterranean and Aquitanian-Basque in the Pyrenees). So it's fair to say that in R1b-saturated Iberia, non-IE was not just some small anomaly, but almost the "co-rule", so to speak.