Rankings – My Favorite Types of Numbers
I am a grad student in science. I deal with numbers a lot. Every day I have to do some kind of computation, whether it is the tedious work of calculating amounts of compounds needed to make a buffer or analyzing data points. Yet these numbers aren’t exciting at all. Sure, I hope to see that a certain condition is different from the control, since it will lead me closer to graduation, and sometimes I even think about all the implications the results would bring in terms of “furthering science”, though mostly because I have to.
Yet nothing is as exciting to me as numbers that determine rankings. Rankings – how best to compare people and teams?
There are so many variables that can’t be controlled when you do these types of comparisons, so many factors that are evident when you watch sports but don’t show up in the results directly. How is an athlete’s attitude? How did he combat adversity? Was there a matchup issue that led to the “upset”?
Tennis is one of the best sports to analyze the way we rank people. Tracking how professional and college rankings of individuals and teams is interesting, especially since these numbers are put into a computer system and spit out. Things to consider include:
- Is every win the same or are some wins worth more than others?
- How many wins and losses should be kept in the calculation?
Point #1 is very complicated in both the professional and college level. Is a win at the quarterfinal level worth more than a win in the first round? Most people would say yes, since in the quarterfinal the opponent is “tougher” and the “stakes are higher”. But it isn’t the same as having the first race be 100 m and the second 400 m. Winning best of 3 or best of 5 is still the manner of earning the win. No one is given bonus points for taking longer to achieve that win. At the same time, being able to string multiple wins in a row is a sign of strength. If we are using rankings to predict who would beat who, then we would want the “stronger” players to be ranked higher.
Another way in which wins are counted unevenly is the level of the tournament. A first round win at a ATP 1000 or WTA Premier level is more than that at a ATP 250 or WTA Tier 2 (if they still use Tiers…I forget). This also feeds into the measure of “strength”, since the fields at the smaller point events are assumed to have “weaker” players in them.
But what about an up-and-coming player? Should he or she get bonuses for beating a top player? In the professional system of today, the answer is no. Anyone who wins a match at a certain level will receive those ranking points (well at least to some limit which is part of point #2). However, it doesn’t work this way in college. It can’t.
Players in college are ranked on a very short time scale. There is no way that all the Div. 1 players or Div. 2 or Div. 3 players can play in one tournament or play each other. They cannot accurately choose the expected level of their competition since in the team format, the team’s level can vary quite a lot from year to year.
Therefore in college the computer system has a baseline ranking to work off of, put in arbitrarily, and then adjusts as the season goes on. Anyone who complains about professional tennis rankings can look at college rankings and be quite confused. The team rankings seem especially strange. Why is the Stanford women’s team, undefeated on the season, ranked below Florida, who they beat? Why did Ohio State University jump over USC on the men’s side when USC hasn’t suffered any losses since ITA Indoors?
This leads to point #2, which is how long of a time scale should be included? And also, how many wins in that timescale should be included?
The big controversy in the WTA is whether Caroline Wozniacki is a deserving #1 ranked player. She didn’t fudge her results to get that ranking via the computer, but people say she doesn’t deserve this ranking because she hasn’t won a Slam. She has already won 3 titles this year (I think?) and been to a few finals too. Meanwhile, Kim Clijsters has won the last two grand Slams but has suffered through some injuries, and before that, Serena had won 2 out of the 4 slams when others only had 1 and she was also not the top player.
Yet Kim and Serena played much less tennis then Caroline, and also didn’t consistently make as many tournament finals as Caroline. But is it fair to count one’s best result as your true ranking, like they do in sports like gymnastics (at least on the team side) and track and field?
On the other side, the college singles and doubles rankings fluctuate quite madly. I don’t know exactly how they work, but with bonuses for beating top players, they seem to average out results as the season goes along. For example, Martin Kildahl was unranked coming into the season. He jumped into the rankings after defeating then #13 Sanam Singh. He has fallen almost out of the rankings after losing a bunch of matches in Pac-10 play to opponents who were relatively close in ranking. In my opinion, he is probably deservedly in the low 100’s/unranked, but there was a time in the season where he was ranked higher than UW’s #1 and #2 singles players. But college coaches know their players and they know the opponents’ players. If UW had put Kildahl at #1 singles over McMorrow or Neduchezhiyan, eyebrows would have been raised. I’m not really sure if the opposing coach would argue too much on it, since there would be numbers to his case.
Anyway, rankings. I love discussion about rankings because you get to see how each person evaluates athletes. What matters most to them? Consistency, hot streaks, fighting ability, etc. are all things to consider. I don’t have the time to really work on my own ranking system, so I go off of what is published and decide for myself, based on what I have seen and heard, whether one person will defeat the other.
Anonymous Sports Fangirling 
No trackbacks yet.