but session ends (leaving YouTube) are a factor that gets penalised more.
I've heard that mentioned before, but I frankly do not believe it. I've seen nothing but shoddy research that attempts to prove to what extent such a penalty affects a channel. It's important for us to remember that the algorithm is far more complex than anyone ever seems to like to give it credit for. And it's dynamic, ever changing. They tweak the algorithm more often than people realize. The one thing we know for sure is that Youtube's search and suggested video algorithms give preference to those videos that lead to a longer overall viewing session. We know this because Youtube has told us this (for the first time, way back in 2011 when the changes were announced first in a google products forum thread. There was a lot of questions and a lot of solid, thought out answers from Google employees. It was a very interesting time). Much of the rest, and especially session-ends, is pure conjecture, a mixture of educated speculation and incredibly unscientific testing.
That being said, the only content I can control is my own. The only way for me to assure a decent session length for a viewer that happens upon my channel is to keep them on my channel for as long as I can. This also (obviously) helps me because of the ad revenue from them visiting multiple videos. If someone has a short attention span, them staying on my video for a longer period of time shouldn't have any effect on whether they abandon the site at that point or move on to another video on either my channel or someone else's. If I believed that there was some huge site abandonment penalty, I'd be more concerned and more inclined to agree with you, but I don't. I just don't see any evidence of it, nor do I think it would make much sense for Youtube to weigh that kind of metric so heavily. I mean, people leave Youtube for a huge variety of reasons having nothing to do with the content of the video they were watching when they left. It wouldn't make sense.