While I agree that we need better metrics (and, more in general, a better academic system) I think that publishing fake research (committing fraud) is still way

While I agree that we need better metrics (and, more in general, a better academic system) I think that publishing fake research (committing fraud) is still way more severe than committing H-index hacking. The latter mostly translates into publishing lots of papers that present small advancements and does encourage scientists to work on topics where there is very little risk of failing (rather than chase challenging topics). These are have long-term negative implications for the systems and for our society as a whole, but, at the end of the day, it's still real research.