[troppl]

This is actually not such a good example I think. A better example would be that you have a small improvement to an existing algorithm, but you know that another group across the globe is also working on/with that algorithm and maybe they also found out the same improvement as you did. If that is the case, then it will be a race on who publishes a paper about that first.

If you're not first, there's no reason to publish a paper anymore. And you just lost a good paper... (And paper count is really almost all there is in academia, at least that is what it seems to me sometimes)