I don't know well the Ancestry.com test or how they calculate similarities by region, but it seems that they just compare with other members in the database, a bit like the ancestry composition in 23andMe.
Dodecad is completely different. It doesn't take look at individual as monolithic blocks but rather as entities already composed of various historical admixtures. The admixture represent ancient populations that usually no longer exist in their pure state, which is why admixtures never reach 100%, and some don't even go above 60%. That's why I find Dodecad so interesting since these "zombie admixtures" of ancient populations can be more easily equated with the spread of Y-DNA haplogroups and actual historical ethnicities. The difficulty is to find the right admixture, since the number of combinations possible is extremely high.
In contrast, the ancestry composition in 23andMe can give you 100% for one country or region if all your ancestors are from the same country, even if that country is heavily mixed ethnically. It looks like Ancestry.com is somewhere in between, looking only at modern population, but dividing more the various chromosomes while 23andMe looks more at the global picture.
Please have a look at the
distribution maps I made for the various Dodecad admixtures.
Regarding your results, it would be surprising to find such high West European admixture if you are of 100% Slovak ancestry. But since you have 25% of "White American" (presumably mostly from the British Isles), it is normal.