In most cases, I don't even try to counter, as it's usually pointless, but in those cases where I do make a counter-offer, I'm doing it intending to keep them in the fold "permanently" (3+ years, ideally longer).
Anyone who responds on HN claims to behave in an ideal way, but if you took this for an indicator of the industry you'd be surprised almost immediately - whether because these claims are often untrue or simply because the people making them are so unrepresentative.
Suppose Abel is working for Manny; Abel gets a better offer and tells Manny. Manny decides to counter.
In order for it to make sense for Manny to counter and plan-to-fire Abel, he has to be planning to hire Baker to do Abel's job, but wants to keep Abel around a while for stability and to on-board Baker.
That's enormously expensive and risky compared to just paying Abel more in the first place or to just counter-and-keep Abel. If the job is that critical, pay Abel. If it's not that critical, just let the job stay open until you hire Baker.
Taking on the severance costs to fire Abel later (and pay higher UE insurance) just to "get back at him" is self-defeating more than it could possibly hurt an employee who has just today proven he could get a higher-paying job. So, where's the "win" for any rational Manny?
That would be a big motivation for me, especially if his job is critical.
The epilogue to his story is that most of the team he built worked for Sam, not the company. Within six months the team had disintegrated, with almost all of the senior staff having moved on after seeing the way Sam was treated. Frankly, any management worth their salt would have seen that one coming way before cutting Sam loose.
The lesson here is that even though its against their own self-interests, the counter-and-fire thing happens, and probably more frequently than not (though a single vague anecdote does not evidence make, I admit.) I think any engineer who thinks they ought to be making more should ask for it. But they should also be prepared to move on if they don't get what they want, unless they really trust their managers.
And never take the counter.
That would manifest itself as me being more aware of the risk, but I'm also rational enough to know that unless I find a way to rebuild that trust, I might as well have not made the initial counter-offer. (I'm also writing from a pretty big [by typical HN standards] company perspective. On a sub-10 person technical team, the equation is a lot different.)
Deadlines.
Having you walk out this month may be a big deal while next quarter, not so much. Or I may have found your replacement by then.
It's not about "getting back at" someone. It's about minimizing my risk as a manager. I would rather have 80% the performance with an 80% chance someone will be around in a year rather than 100% the performance and a 50% chance someone will be around in a year.