I was bored last night and thought it would be fun to try to objectively score the Frozen Throne cards and their impact on the meta and then compare that to the scores we collectively rated them prior to the set release. So I went to HSReplay and pulled all the relevant data and dropped it into a spreadsheet. Through some trial-and-error of various formulae, I was able to come up with a 1-5 score for each card, based on its deck winrate, its played winrate and the amount of times its played. I've called it Meta Rating and, while it is almost totally arbitrary, it does seem to match pretty well the subjective scores that most would apply to each card. I've added all this data to our ratings spreadsheet on the tab called "HSReplay Stats".
https://docs.google.com/spreadsheets/d/1LLHGCAPCvMjlzw3bXx8emy87nYm7v2s17d1igF2Cf7M/edit?usp=sharing
On the "Scores" tab, I included the Meta Rating and then did some calculations to see how close our collective ratings were to each card's Meta Rating. You can sort the list by each column by clicking the Filter button in the toolbar or hovering over the column header and selecting your sort via the dropdown. The "Difference" column shows the difference between our rating and the Meta Rating; a positive number means we overestimated the card's meta impact, a negative number means we underestimated it. The "Accuracy" column shows how close our rating was to the card's Meta Rating in percent. So we were waaaaay off on Prince Keleseth, but pretty close on most cards, with over 90% accuracy on close to half the set. Our total accuracy score on the whole set was 86%. Not bad!
Anyway, I had fun messing with the data and hopefully someone else besides me finds it interesting. I might next try to compare the Meta Rating to each individual's card ratings and then give an accuracy score for each person, so we can see who is the best card predictor amongst our group. Might be good to know while we sit here arguing about the upcoming cards!