Spot the irony: Speed identifying, quality controls, 4th, 5th, 6th, IDs: community perspective?
The general consensus (with which I tend to agree) is that leaderboard climbing (and gamification in general) is mostly harmless. It’s just one of many legitimate ways to engage with the site. And the potential pros and cons of extra supporting IDs were quite nicely summed up by a post in the thread linked above, so there’s no point re-hashing it all again here.
The specific question about “a good percentage of supporting versus leading/improving IDs”, depends on how on “good” is interpreted. A good percentage for an individual wishing to climb the leaderboard as fast as possible might be 100%; whilst a good percentage for the community identification effort as a whole might be 60% (or less).
It was pointed out earlier that there are currently 108 million observations in the needs-id pool - to which I will add that 77 million are more than a year old. That’s roughly 38% and 27% of the total, respectively. Clearly, piling on umpteenth supporting IDs for observations that are in little danger of losing their RG status is not going to help reduce that enormous backlog.
When I started identifying, I chose to apply the golden rule. I would generally much rather have my oldest needs-id observations reviewed first, so that’s what I try to focus on myself when identifying: i.e. do as you would be done by. But even if I never uploaded any observations, I think I would still take this approach. I find individual targets somewhat motivating, but competitions quite boring - I would much prefer to gauge my contributions relative the collective effort as a whole. If the needs-id backlog was ever reduced to a few percent, I would of course change my approach accordingly - but I don’t expect that to ever happen in my lifetime. The vast majority of my supporting agreements took the observation to RG: about 57% out of 104k identifications (along with 52k annotations). That doesn’t even get me onto the bottom of the global leaderboard (which currently requires over 118k), so I suppose I’ll never be classed as a “power identifier”. I’m pretty sure I’m going to be able to cope without this particular tin badge, though 
The leaderboards and ID stats are currently a bit of a curate’s egg. It’s not easy to interpret them in a way that gives due recognition to all the differing contributions of our community, and this sometimes leads to speculation about the motives of those that happen to end up at the top. I don’t blame people for being curious about this: it’s just human nature to speculate when not in possession of all the facts. I have always felt that the site might benefit from giving less prominence to individual rankings, and emphasising our collective accomplishments (and failings) instead. The year in review statistics go some way towards that, but it seems like a bit of a sideshow, and it’s not clear how they should be interpreted in relation to one’s individual approach as a contributor to the site.