Ranking Feature
Forecasting the USTA National Championships
by Dallas Oliver, 1 August 2015
| Share: |  |
|  | |
|
|
This next week is the highlight of the American junior tennis calendar. The qualifiers are all done, and the USTA National Championships take place in Kalamazoo, San Diego, Atlanta, San Antonio, and Little Rock.
Tennis Recruiting is onhand once again to forecast the tournaments for every age division. Let's take a look...
USTA Nationals Heat Maps
We use the Tennis Recruiting Power Ratings to produce Heat Maps for these elite junior tournaments. You can view the Heat Maps by clicking here. We plan to provide heat maps and forecasts for all eight age divisions - adding tables for the various age groups as draws become available.
These Heat Maps use our power ratings to show probabilities of how far we expect players to progress through the draws. (You can see power ratings for all players in the Nationals by
clicking here.) We will update the heat maps after each round of the tournaments to reflect the new probabilities based on actual wins and losses.
Explore the data by choosing different age divisions using the top tabs - and then click on the top links and the table header columns for different sorts of the data.
One of the most powerful features of these Heat Maps is the prediction of upsets. Given that we have probabilities for these match-ups, we can look broadly across the tournament to estimate the number of expected upsets. If five seeds all have 80% probabilities of winning their first matches, they are certainly the favorites, but we would collectively expect them to win only 80% of their matches - or four out of five. Our system fully expects one of those seeds to lose.
Predictions
Scattershooting through the heat maps, there are lots of interesting predictions you can make based on the probabilities...
- Players with the best chance of advancing to the semifinals:
- The field in the Girls 16s looks to be wide open. All the players have tough draws - with no player having better than a 41% chance of reaching the semifinals. Our forecast likes the chances for unseeded Abigail Desiatnikov to make a deep run, but it does not think any player has better than a 17% chance of winning the title.
- Top seeds with the toughest roads to the semifinals:
- The draw with the most interesting first-round matchups is the Boys 16s. That draw features twelve matches between closely-ranked players where the probability is 60%-40% or tighter. We also project 17 upsets in the 64 first-round matches on Saturday.
These are just a few of the interesting predictions you can explore using different sorts of the data.
Frequently Asked Questions
Since running our first heat map during the Clay Court Nationals last month, we have received lots of questions about how the forecasts work - and about our Power Ratings that are behind the heat maps.
Here are three of the more common questions...
(1) What rating numbers did you use to make these predictions? How do you compare players from different graduating classes? One of the most common misconceptions about our rankings is that we rank each graduation class separately. This is not the case. Our rankings start out as single lists with all players of the same gender. The final listings presented on our website are broken down by graduation year because our lists are intended for coaches and players to use in college recruiting - and kids are competing with others with the same graduation year for roster spots and scholarships.
Our ranking system first uses all match results for all players to come up with a Power Rating for each player regardless of age and graduation year. We then order all players by Power Rating to come up with a master list before filtering it by graduation year to produce our rankings.
The Power Ratings described above are used to produce the predictions in our heat maps. You can see the Power Ratings for all players in the USTA Nationals by clicking here.
(2) When I sort by the WIN column, why doesn't the player with the highest ranking/rating appear at the top?
The predictions take into account (1) the Power Rating of each player and (2) the opponents they will face on their path to the championship. For example, the No. 1 and No. 2 seeds are separated into different halves of the draw. Their paths to the finals include completely different players. All unseeded players have different strengths, and the ease or difficulty of a player's path is often the luck of the draw. For example, the No. 1 seed might draw the highest-rated unseeded player in the first round, and it is possible (although highly unlikely) that the second- or third-highest rated unseeded players could be on the draw lines just below. For all of these reasons, the highest-rated player does not necessarily have the best chance of winning the tournament.
When you look at the next match to be played, however, the higher-rated player will always be predicted to win and will have the darker shade of green.
(3) You seem to do a pretty good job predicting how many upsets will occur in the tournament. Why don't you adjust your ratings so that you expect fewer upsets?
There is a big difference between knowing how many upsets to expect in a given round and knowing where those upsets will occur. A good way to think about this is rolling a 6-sided die. I want you to predict whether or not I am going to roll a 6. If you want to get the answer right most of the time, you will guess "not a 6". If we do this experiment 1,000 times, you would expect to be wrong one sixth of the time - or about 167 times. But you do not know which individual rolls will come up 6 - and so guessing "not a 6" is the best answer to give on every die roll.
We treat our tournament match-ups similarly to the die roll. With the probabilities from our ranking system, we can predict how often we expect to be wrong without knowing where the upsets will come.
Being able to accurately estimate the number of upsets shows that a system has a pretty good understanding of the relative quality of players.
Follow Along!
As the tournaments move along, we will update our probabilities and predictions. Check out the upsets - and cheer on the players who move further than their power ratings might suggest. We expect there to be quite a few upsets - that's why they play the game!
Leave a Comment
More Ranking Articles
19-Sep-2017
Comparing Rating Algorithms
During the U.S. Open we posted information on social media about matchups involving American players - previewing matches by displaying rating and ranking differences and highlighting wins by underdogs. You might have noticed that the various rating systems often agreed on their favorites. But what about the close matches where they disagreed? Let's take a look ...
13-Jul-2016
Behind the Tennis Recruiting Rankings
This week marks the 569th consecutive week that Tennis Recruiting has
put out graduation-based rankings of American junior boys and girls.
Rankings are front and center on Tennis Recruiting, and people often ask
us how our rankings work. Today we describe our ranking system ...
19-May-2016
What Is An Upset?
Tennis Recruiting is a website that rates and ranks junior tennis
players. One of the questions we get most often from our users is,
"What exactly is an upset?" There are many possible
definitions of an upset - this article explores the question and
puts forward an answer.
The lifeblood of Tennis Recruiting is its rankings - and our team has been ranking tennis players for more than two decades. Learn about tennis rankings in general - as well as our best-of-breed ranking system.
