The most accurate measure of a champion's strength is impact.

Now, I see a lot of people here arguing about whether winrates or pick rate determines how strong a champion is. So I wanted to explain this concept in a way that helps clarify determining what champions are really good, and why pros seemingly only play 45% winrate champs.

What is impact? Impact is like zone pressure. Zone pressure is when you have a specific skill on CD (say Thresh hook), and so if you were to walk up to a person, they have to back up because they know you're going to hook them. Impact is one's ability to control the game based on the champion's skillset, level of consistency, current strength, and output.

Skillset - how one's skills correlate with your options in the matchup. For all intents and purposes, we're going to include base stats here as well. This is how you determine if a champion is strong early, mid, or late-game.

Level of consistency - how well a champion performs under the meta. By this, I mean how well they do with the best items, strategies, or synergies with top tier champions. This is far more important in competitive games than in solo queue, because solo queue by definition is inconsistent. Meanwhile, competitive games, as well as high elo games are much lower variance. Look at pro games. In solo queue you can see Malzahar or Anivia mid but in competitive games the only variance you really have is Khan picking whatever and Huhi playing ASol.

Current strength - how good the champion is on an absolute level. I use it to refer to how easily a champion can perform his/her task.

Output - what happens when they perform their task well. For simplicity's sake, let's just call this winrate.

So how does this tie into impact? Look at a champion like Yasuo vs. a champion like Syndra. In terms of skillset and strength, Yasuo should beat Syndra in almost any situation. He has better output and a countering skillset. He's cancer to play against at all stages in the game. However, he requires constant perfect play (low current strength) and his level of consistency is incredibly low. Meanwhile, Syndra still has a respectable output, her skillset scales well into the game, and if you master her, she's got very high current strength, and her relative safety as well as her ability to easily set up picks keeps her high on a consistency level, even in an AD heavy meta. Thus, 46% winrate Syndra, once mastered and used to her full potential, is currently better than Yasuo.

Bear in mind this is not an intuitively easy thing to measure. But the best way to do it is to ask yourself these questions;

1. What currently is the meta? What does it favor? (keep in mind the meta is simply what gives you the best results in the most general scenarios of high level play. aka highest impact).
2. Where does my champion rank on an overall scale?
3. Does my champion's overall rank and contribution (output) fit what is needed?
4. Does my skillset mesh well with my teamcomp or against the enemy?

And here's an example from 2-3 weeks ago that i used regarding picking Vladimir.

1. The meta was currently heavy protection for squishies. Multiple lockets, ardent censer, and heal support/barrier AD meta made it very difficult to burst down enemy ADs.
2. Vlad is very good compared to other midlaners. Overall he ranks higher than average and contributes a lot of damage and %damage to teamfights.
3. However, it's more important to have utility. Shielding for ADs or CC to get picks/win teamfights is far more important than having a lot of damage.
4. They have a Lulu/Kog. And thus, regardless of winrate, 2-3 weeks ago, Vladimir was not a good champion to play. However, with the overall reduction of tankiness, nerfs to Locket stacking, Vlad is now a stronger champion with a better impact on the course of the game.

Hope this helps.