I didn't say it was the best for the user. Really the article misses this by comparing a new UI feature to a life-saving drug, but it doesn't matter. The point

I didn't say it was the best for the user. Really the article misses this by comparing a new UI feature to a life-saving drug, but it doesn't matter. The point is, whatever metric you're targeting, do you use this algo or fixed group sizes?