The last two seasons of the show saw underdogs winning the million dollar prize. Brooke and Scott and Dana and Matt obviously didn’t dominate the race, but they managed to pull it off at the very last minute.
By underdog, I mean a team which didn’t really perform amazingly well in the race but managed to reach the last leg. On the other hand, I call a team dominant if it performs really well throughout the race.
I got curious to see how often an underdog win is the case for the 29 seasons of the show. In this post, I answer this question with some basic data analysis.
This is the second post in my Amazing Race Domination series. The other posts in the series are:
- Who Dominated Amazing Race 29? A Computational Approach
- Do Underdog Teams Always Win the Amazing Race?
- Number-Crunching Amazing Race Seasons 1-29: Who is the Most Dominant Team?
- Predicting the Amazing Race Winner with Random Forests
Definition: Underdog and Dominant
My definitions of underdog and dominant teams are based on the domination index I defined in the last article.
Basically, I set a threshold K such that a team with a final domination index lower than K gets classified as an underdog, while a team with a domination index higher than K gets classified as a dominant team.
In my analysis, I set the threshold to 0.6, which means that a team is dominant if, on average, it places higher than 60% of the other teams. This is purely a subjective criterion which I based on the domination distribution across all finalists in the 29 seasons. In terms of Season 29, Tara and Joey get classified as a dominant team, while Brooke and Scott and London and Logan do not.
Top 3 Underdog and Dominant Teams
We first look at the distribution of underdog and dominant teams constituting the top 3 in the 29 seasons of the show.
We see here that underdogs edge out dominant teams in number in the final leg. A more interesting thing to look at is the constituency of the top 3 teams. That is, how many dominant teams are there in the top 3 for each season?
Top 3 constituency
Let’s look at the dynamics of the top 3 constituency.
We see here that only Seasons 23 and 26 had a top 3 which included all dominant teams. Season 26 is the Blind Dating vs. Currently Dating season, and all three teams in top 3 were indeed quite strong. If fourth-placers Mike and Rochelle (underdogs) were able to outlast another team in the top 3, then only Season 23 would have had the distinction of a top 3 of dominant teams.
The last two seasons of the show had a top 3 of two underdogs and one dominant team. It’s interesting to point out that stronger teams were eliminated in the 4th and 5th places. Surely, luck plays a central role in reaching the top 3, not just pure racing skill.
Looking at the distribution of team constituency, the most common occurrence is a top 3 of two dominant teams and one underdog team, followed by one dominant and two underdog. A far third is the three dominant team case.
I have to point out that there hasn’t been an instance of three underdog teams in the final leg. Boy, wouldn’t that be interesting to watch. If that were the case, however, if all teams were underdogs, wouldn’t that mean that no one is an underdog?
Let’s now answer the question of whether underdog teams always win.
From the win-type dynamics, we see that when dominant wins occur, they tend to occur one after the other. Underdog wins happen only once in a while. That’s probably why they’re so exciting to watch.
A dominant team winning happens almost twice as often as an underdog team winning.
I’m gonna say that I’m a fan of underdog wins. It’s quite boring to see a team be dominant the whole race and basically steamroll to the end. An out-of-the-left-field win is much more interesting. That’s probably I liked the Brooke and Scott win at the end of 29, though I hated the Dana and Matt win of 28. For me, it boils down to personality as well, and although I found Brooke to be very annoying, Dana was flat-out abusive. I really wish Sheri and Cole won 28.
Yes, dominant teams win almost twice as often as underdog teams. If you’re interested to see the Python code I wrote to analyze the data, check out my Github.
In the next article, we identify the best racers in the 29 seasons of the show measured in terms of race domination.