Not just heavy seas.
Want to see how little means a lot?
Take a bit of scrap paper, draw a vertical line, oh 5" long thereupon. Now, draw a horizontal line 5" long centered on the top of the vertical line, giving a T.
Find a suitable substrate and stick a pin at the base of the T.
Now rotate the T just 5º either way. Tick off on the substrate where either the low or high end of the crossbar on the T gets to. Now, rotate 5º the other way. Note how big that difference is.
If we just arbitrarily call that a scale of 1" = 10'-0", for a 50' long yard 50' up, you can now scale how high the vertical motion at the horse would be with a 10º roll.
Ok, this is not to proper scale. Ships roll around their Center of Buoyancy in all three axes. So, that T probably ought be pinned about 6.5" down. Except we then need to subtract for the moment arm coming from the CG(t) (thrust Center of gravity), so, scootch up an 1/8. Also, to average the pitch out, we probably need to come up another 3/8" - 1/2" too. While we are at it, square -rigged ships only roll about 6-7º, but, that's skewed into a heel of 3-6º
The above is why such examples use simple 5" diagrams (which can be executed on folded bits of 8.5x11 paper). Fur a test of land-based vertigo, get a 8.5x11 copy of an elevation of a clipper mast , and pivot that through 5-6º--pay attention to those t'gallants & royals).