Take two 50 year old engineers. The one who has had 10 jobs over the past 25 years is considered infinitely more employable than the one with 25 years at the same department of the same company.
This case isn't ageism, it's the expectation of career stagnation, yet many of those engineers with that 25 year tenure think it is. It's a leftover from the days when loyalty was tied to tenure, and loyalty was valued highly in the hiring criteria. Those days are over. Loyalty is nice, but loyalty that negatively impacts your career options (no learning or improving, inability to use marketable tech skills, etc) is foolish.
When you say 'people like me', keep in mind that just because you have had three different roles at one company in itself doesn't mean you didn't stagnate, just as the person with ten jobs may have stagnated.
All assumptions have some danger, and I'm not condoning the behavior. I think it's valuable to explain to people who feel they are being discriminated against based on age may be more likely associated with stagnation (actual or assumed) based on extended tenure.
I'm certainly not saying that someone who has held multiple positions has not stagnated, either, simply that fundamentally the only difference between me and someone who worked for three different companies but had the same roles as I did is merely the company nameplate. To make different assumptions based solely on that is illogical.
Are you arguing that in that scenario it's more likely that the one-workplace worker really has better experience, or are you arguing that it's possible that he has better experience?
CV-sorting is not a maximization problem, it's a satisficing problem of getting a few good candidates to the interviews at a reasonable cost; which generally requires throwing out 90% of CVs without spending the resources to seriously investigate those people. As long as you have enough qualified candidates, factors that "often are bad, but sometimes are good" equal an automatic forward to the circular folder under HR desk.