Right then; it's about time I did part 2 of my North American investigation! And in part 2, I'll be working out... what coaster selections in North America have the strongest top end?

Yes, I'll be focusing solely upon the parks' most highly rated coasters in this question! As I got a few comments about the slightly weird results in part 1, I'm thinking that this question should hopefully correct that and produce results more akin to what you might typically expect, as I'm focusing only upon "the interesting stuff" here!

So how did I work this out?

Well, I used 3 different ways of attempting to measure this.

The first method I used was...

Mean of Top 3

The first method I used was calculating a mean of each park's top 3. For those of you that don't know, the mean is the calculated average of each top 3, and the formula is as follows:

Mean of Top 3 = Sum of All Ratings/3 (as the top 3 is being focused upon here, the count of ratings will always be 3)

When this formula was applied to each park's top 3, the highest rated top 3s came out as follows:

Let's move on to the second measure...

Median of Top 3

The second measure I applied was the median of the top 3. The median is the midpoint of a dataset (I.e. the middle value), so as we're talking about top 3s, the median value is the 2nd highest rated coaster at each park. When I searched for the median of each park's top 3, the top 10 were as follows:

After those two measures, I used one final measure of my own...

Matt N Formula for Top End Strength

The final measure I used was my own formula. You might remember that I used my own formula to denote consistent strength in part 1. The formula for top end strength is slightly adjusted compared to that, only taking into account the highest rating and the upper quartile. The formula I'm using to work out top end strength is as follows:

Matt N Formula for Top End Strength = (Highest Rating + Upper Quartile)/2

This formula does require me to touch upon the entire selection again as opposed to just the top 3, but the use of the upper quartile and highest rating mean that I can hone in exclusively on the more highly rated coasters while also considering the strength of the top end as a whole. For all intents and purposes, it's the average of the highest rating and the upper quartile, so the lowly rated coasters within a park's lineup still aren't considered.

When this formula was applied, the top 10 parks were as follows:

So, what did we learn in this part of the investigation?

Well, I can quite decisively crown a winner for the park with the strongest top end based on my 3 measures, and that is Cedar Point. The park came out on top in 2 of the 3 measures, and even in the measure it didn't win, it still came 3rd. And I'll be honest, Cedar Point was literally a hair away from the top spot; the difference between Cedar Point and the winner in that measure, Busch Gardens Tampa, was only 0.05. (The reason the top 3 are all listed as 9.5 is because I rounded the Matt N Formula Score to 1 decimal place; there was a difference between each of their exact scores, albeit a very small one)

Before I close off, here's my spreadsheet once again, so that you can peruse my workings at your pleasure:

https://docs.google.com/spreadsheets/d/1D_Zv3-Nb3B8oV7WRe3_G34tWDeAde_rJ8cMSeDN6KpM/edit?usp=sharing

Thank you all for reading part 2 of my investigation into North America's major coaster selections, where I attempted to find the coaster selections with the strongest top end. I hope you found it interesting, and I hope the results are more in line with what you were originally expecting than they were in part 1!

TL;DR: I attempted to find the North American coaster selections with the strongest top end. I used 3 different measures to calculate this: the mean of the top 3, a calculated average; the median of the top 3, the midpoint value; and the Matt N Formula for Top End Strength, which calculated the average of the park's highest rating and the upper quartile. The winner was determined to be Cedar Point, which won 2 of the 3 measures and came an extremely close 3rd in the measure it did not win.