Private User I did not say that it is worthless, which it is definetely not. But for chasing after distant origins it is not usable at all. Its intended usage is for finding and confirming much closer relatives.
Theres absolutely no way to proove such a distant connection with autosomal-DNA, the only way to use DNA to proove anything this distant is through Y-DNA and mtDNA.
You do inherit DNA from all your 32 third-greatgrandparents, on average you share 3,2% from each of them, but in reality you could share under 1% from some and 6% from others. Just one more generation and theres actually 4 ancestors that you dont share any DNA with and if you go back 18 generations to about 1450, you would have 131.072 15th-greatgrandparents and only 1090 of them would have DNA that you have inherited. On average you would share about 0,09% with each of them.
To even try to proove that the DNA shared is through any of these 1090 ancestors in 1450 is not possible, it is hard enough to proove which of your 3rd-greatgrandparents your shared DNA is from. You would share DNA with all your 2nd cousins but not with all 3rd cousins and its not very probable to share anything with so distant relatives as mentioned.
In addition you would need a full ancestry on both matches at all lines going back to the perod that you try to proove, not many has full ancestry going back beyond maby 8 generations or so and because of that many jump to the conclusion that a distant connection must be the right one, while in reality the connection is through one of the much closer unknown ancestors.
For those who can read swedish this is a good article on the subject: https://petersjolund.se/kan-en-dna-match-bevisa-ett-slaktskap-pa-14...
If it was the ethnicity estimations you where refering to, theres no way to use that to proove that you have scandinavian ancestors. Many norwegians does also have british in theyr results, withouth having any british ancestry. It is just a estimation on what reference populations you have some similar DNA with. The genetic origin of many of these referance groups are overlapping and shares much of the same genetic ancestors who inhabited the areas after the ice-age. It is not possible to really distinguish 100% between any of these groups, but they are considered good to distinguish groups of endogamous groups from the rest and also good at distinguising on a continental level, but not that good on detailed geographical areas.
The sagas are surely not fictional or fairy-tales, but many of them where comissioned by someone with a political agenda to proove a connection or inheritance and it is not that easy to distinguish between what is really historically accurate or not in the sagas. I will let the historians and medieval genealogists try to solve that as I am not that experienced on that yet..