PostEaseIt is a prediction system, not a ranking of "good" videos
YouTube's recommendation system is built to predict what a specific viewer is likely to watch and keep watching, based on that viewer's own history and what has performed well with similar viewers. It is not judging a video's quality in the abstract. It is estimating whether a particular person, right now, is likely to click and then stay.
Retention is the signal that carries the most weight
How long people watch, and whether they watch to the end, tends to matter more than the view count itself. A video that gets fewer clicks but holds attention through most of its length is a stronger signal than one that gets clicked often and abandoned quickly. The system is optimizing for watch time, not click count alone.
The thumbnail and title do a different job than the video itself
A thumbnail and title's entire job is earning the click. Whether that click turns into real watch time is a separate question the content itself has to answer. A title that overpromises can earn short-term clicks and still hurt a channel, because the resulting low retention signals the system to recommend it less.
Consistency helps, but not for a mystical reason
Uploading on a predictable schedule does not earn algorithmic favor by itself. What it does is give an audience a reliable reason to come back, and a channel that earns return viewers produces the retention and engagement signals the system responds to. The benefit of consistency is indirect, through audience behavior, rather than a reward from the system for showing up on schedule.