Past behavior is a pretty good predictor of future behavior. If you're dating someone whose previous 3 spouses died in mysterious kiln explosions, I'd stay away from the ceramics factory.
Well I think that’s an absurd and extreme example, so the reductio ad absurdum is self-limiting, since serial killers are, hopefully, not the common case.
How about this analogy instead, being more in the realm of the applicable and commonplace: I broke up with many girlfriends, and had many short- and medium-term partners, before meeting my wife of over a decade now, who continues to knock my socks off to this day.
In practice, past performance is not a guide to future outcomes. How many times have we heard this caveat applied, in far more predictable markets than hiring and/or dating? Past performance of employment candidates is a subjective and observational assessment; relying upon it (especially in a negative framing) is to invite not only ones own biases to take center stage, but also those of others; see earlier point re. reinforcement of structural discrimination.
At most, someone’s past successes qualify them for my interview shortlist, but there are many other green flags besides, and I’ll save my type 2 errors for the back half of the hiring funnel, after assessing someone’s potential.
And considering that, someone who has only held long stints at previous jobs probably gets a red flag for not being desirable enough to more companies or just plain lack of curiosity.
Past behavior is a pretty good predictor of future behavior. If you're dating someone whose previous 3 spouses died in mysterious kiln explosions, I'd stay away from the ceramics factory.