This is a new post entirely because I wanted to make a very different point.
{quoted}
Are these bad things? Because I always play solo Q on new champs i've bought. See, way I see it, if I've played Yasuo mid for like 50 games, and I understand how mid lane works, I can learn another champ there easily. Like, say for instance, Ahri.
Unless it's Leblanc. She's stupid.
Not everything in the list is supposed to be a bad thing, although this would most certainly be frowned upon in ranked.
The idea of my post (and I was not joking about how many things got cutoff and I had to edit.....) is that the LCS is an environment where everyone takes very safe picks, everyone knows their champs EXTREMELY well, everyone works together as a team, and it's generally the highest level of play. Even then, win rates are not necessarily a good indicator, even in LCS, due to the statistical effects of ban rates (i.e. sometimes if a heavily banned champ is not banned, there are specific reasons for that that decrease the statistical value of that event).
Throw in a crap ton of other things that almost certainly DONT happen in the LCS, and the already sketchy importance of win rates goes out the window. In the specifically quoted two examples, the more popular a champ is, the more likely it is that someone who can't play them will try to pick them and do well anyways. 
and 
are good examples of this. Those people are going to have a MUCH different win rate than someone who is REALLY GOOD at the champ at their same elo. But those people are still included in the overall win rates, since it's a pain to put in things like "win rate for people who have played at least 20 games with the champ recently", even before we get to all the really negative shit people do in ranked that just makes you lose.
Effectively, the statistics of this are too prone to error to be worth proving arguments, and this becomes more of an issue the more problematic a champion is. It's also affected by ranks too...
Let's just look at the most popular and least popular champs. For all ranks, in ranked solo queue, the most popular champs are: Graves, Jinx, Thresh, Leona, Morgana, Lee Sin, Blitzcrank, Vi, Caitlyn, and Annie. For master and challenger, this changes to Graves, Thresh, Janna, Lee Sin, Ahri, Morgana, Vi, Jarvan IV, Leblanc, and Lissandra. Half of those changed.....
For win rates at all ranks, the highest are Heimerdinger, Sion, Sejuani, Blitzcrank, Katarina, Vi, Malzahar, Fiddlesticks, Jinx, and Annie. But only 4 of those are over 20% popularity... that's a big difference. So if you wanted a popular champion with a high win rate across all ranks, you might want to pick Blitzcrank, Vi, Jinx, or Annie.
Let's move to master + challenger again. Highest win rates in ranked are Karthus, Miss Fortune, Tryndamere, Shen, Warwick, Aatrox, Elise, Kog'maw, Kayle, and Nunu. Tryndamere is the only one there that breaks 3% popularity. None of them are popular whatsoever, and the list is blatantly different from who is popular by.. a lot.
All of this is just me trying to show that statistics here are pointless after a "well that's interesting" when it comes to trying to prove something. Win rate really isn't that good of an indicator. At all levels, there is only a tiny level of overlap between popularity (and how hated someone is), and their win rates, and at the highest levels these two things are drastically different This is why win rates vs balance isn't very useful in very general terms, especially when the op is quoting older statistics. He is also including the "bad" people at Diamond.
: Top > Mid > Support
: Support > Mid > Top
: Support > Mid
: Top > Mid
: Mid > Top > (Support)
: Top > Mid > Adc
etc...
and 
which are over 20%. We need much more diversity for support champs, but I think we need a support item overhaul for anything like that to happen.