Everything I know about connections I learned at a cabin two hours north, where the line is old and the weather is worse and nothing that works at home works the same way there.
At home it is fibre and nothing has ever been a problem. At the cabin it is whatever was installed a long time ago, and the difference is not a small degradation — it is a different set of rules.
Three things I have worked out over four winters.
The number quoted on the bill is not the number you have in February. Cold affects old lines, and a connection that tests at a respectable figure in September can be visibly worse in January. I stopped treating the advertised speed as a fact and started testing at the time of year I care about.
Stability matters more than speed on a line like that. A connection that holds a modest rate steadily is far better for watching anything live than one that averages higher and dips. The averages look better; the evening is worse. Nothing on any speed test tells you this, which is why the test is a poor guide.
And the wireless in an old building is doing more damage than the line is. Thick walls, one router at the wrong end, and half the problem was inside the house rather than outside it. A long ethernet cable run along the skirting board is inelegant and it fixed more than anything else I tried.
The fourth thing, which is not technical: lower the resolution deliberately. A steady lower-quality picture on a bad line is a better evening than a high one that keeps stopping, and most players will let you fix this rather than leaving it automatic.
How connections vary across the country, and what actually copes, is set out here.
Four winters, one cable along the skirting board.