Is it possible that whatever gains DA makes among formers ANC voters are offset by defections to the PA, and that the party ends up with a similar vote share to 2019?
Certainly could be the case. I personally suspect the answer lies in differential turnout of whites vs coloured, and both vs the rest. That won't really be resolved until the count can be decently modeled. That said, the DA turning off former coloured voters was expected, as you can see further back in the thread. A bit of further evidence:
Spruitdrift Primary School, Matzikama, Western Cape, mostly coloured voters, but also some white voters:
DA 54% (-12)
ANC 22% (-1)
PA 14% (+14)
FF 6% (+1)
That -12% for the DA is the level that begins to threaten the provincial majority. But pending many other variables.
And some rural Western Cape coming as well:
Klaarstroom community hall, Prince Albert, coloured and white voters.
DA 51% (-12)
PA 46% (+46)
ANC 2% (-28)
Turnout: 94%