Finding The Challenge GOAT: Eliminations

What’s better than America’s fourth favorite sport - Real World/Road Rules Challenge? Not many things I can think of.

To start the search for the True Statistical Challenge GOAT, looking at the elimination challenges is a good(easy) place to go. While not the be-all and end-all, they are straightforward one-on-one(usually) competitions unlike many daily and final challenges. Ignoring social and political games, these are mostly a measure of physical skill.

Rather than looking at basic will-loss records of the eliminations, it’d much more useful to make a rating system, which can be tested in its accuracy.

So instead of wondering if a 14-7 record is better than an 8-4, a ratings system would take both competitors' opponents into account and return a single number for each, which are more accurately compared.

After extensive testing over lunch, the rating system I’ve decided to use is Elo with a K-factor of 350. Much more volatile than chess, isn’t it. But it seemed to have predicted 75% of elimination challenges correctly.

Filtering out the competitors with less than 5 eliminations, we’re left with 66 names.

Men

Women

As previously stated, the ratings are very volatile, which means if two competitors have close enough numbers, you shouldn’t probably think one is much better than the other.

Something else to keep in mind is that while we may know who got lucky where, the ratings do not - they only measure results, so adjust accordingly.

Overall, though, clear patterns emerge and those who’d you expect to do well in physical competitions are near the top.

Or is this all just a convoluted way of taking a shot at Bananas?