I used the respective averages from G25 (Greek Peloponnese, Greek Macedonia, Albania, Polish, BGR_IA, Ingria_IA) and made a 2-way model like the paper but changing between Ingria_IA and Polish -> BGR_IA+Ingria_IA, BGR_IA+Polish.
If you run these models you will see that while in any case (and for the rest of the Slavic impacted Greek pops too ) the "Slavic" increases by using Polish, the percentage increases by that much for the Greek Macedonia, Thessaly and Albanian averages only (Albanian from 22% Ingrian_IA and the rest BGR_IA to 30.2% Polish for example, an increase of 8.2%) . The increase is less severe for the Peloponnesian average and then less than 5% for the rest of the Greek pops (Greek Izmir from 11% Ingrian_IA to 15.2% Polish, an increase of 4.2% etc... ). Not huge differences between them proportionally but I was wondering if this more "dramatic" ? increase of Polish percentage in the Albanians and Northern Greeks is due to extra central/northern european influence on them (celtic, germanic) which the Polish pops should already have in their ancestral composition which the Ingria_IA average lacks.
TLDR; Polish increases the "Slavic" component on the non-Slavic populations when used and not only that but it improves the fits on a 2-way model. This increase is more dramatic on the northernmost non-Slavic Balkan populations. Could one of the reasons be, in addition to the Slavic influence, a possibly increased amount of central/northern european influence on them better captured by the Polish average ?