But experiments test 1 variable.

I don't view hard forks as experiments. Ideally, the experimenting comes before the fork with testing and simulations (which I've been let to believe the steemit devs already did). To hard fork on a regular basis just trying things out would not make for a pleasant user experience.

Concentrating power of votes when distribution is already at 93% is reckless.

Is that based on evidence or opinion? Yes, a whale could upvote or downvote with more power now than before. That could lead to disproportionally higher payouts for some posts. Or it could not. We have the blockchain data to evaluate the voting patterns of whales and simulate things. If they vote for just a handful of things, their share of influence over the reward distribution is still the same, the difference being all the others who only vote occasionally now have more influence. That, combined wth a linear rewards pool, makes it work. I see the overall desired benefit working when these changes are together and (again) from what I've been led to believe they have been tested together.

If SP concentration in the hands of too few is a bad thing, they we should be campaigning for more people to purchase Steem and power up. That's the best long term solution. Other solutions, IMO, create a disincentive for people to accumulate Steem Power to do with as they please.