It's a tale as old as time in the world of rowing. A big race is approaching, the coach set the line up, and things are coming down to the wire. One question is unanswered. Who will be the coxswain?
Coach spends all week switching two of the top contenders, looking for something to stand out. Maybe they ask the rowers. Who gets the job done? Who has that spark and has an inspiring voice? Whom do you trust to steer straight as an arrow?
These are all metrics that are standard practice in almost every rowing team. Coaches leave nothing to chance: measuring their athletes' strength and technique, rigging the boats down to the centimeter. And yet they leave one of the most important seats in the boat to vague, immeasurable qualities.
If you ask any coach what they want from a coxswain, they will often reply, "I want them to steer straight." But what happens when you have 3 coxswains that all steer straight - at least as far as your eyes can tell you?
Simple geometry shows that a coxswain who steers a straight line but drifts only 3 meters off centerline at the 1000m mark, and then corrects by steering another straight line back to center at the 2000m mark, has added a bow ball of distance to the race. Has any crew lost a race by a bow ball and wondered why? Worse yet, a coxswain that makes repetitive hard rudder course corrections in a race is adding drag to the boat.
In the finals at the Rio Olympics, in both the men's and women's coxed eights, the difference between gold and bronze was 3 seconds. And the difference between a medal or no medal was only 1.5 seconds for the New Zealand women. Hence, all coaches are looking for a coxswain who steers as straight as an arrow. When races can come down to fractions of seconds, every course correction counts. My goal was to be able to find that lost speed and to show rowers, coxswains, and coaches alike that there was a numerical method of judging their coxswains' abilities.
In the summer of 2019, I joined the coaching staff of the Row West Racing Summer camp in Los Angeles. My primary job was to work with the coxswains. In the past, I had used many different methods to help teach young coxswains, but this summer was different. The head coaches Zohar Abramovitz and Tricia Blocher were kind enough to allow me to run a new program with the coxswains.
My goal? I was trying to "Moneyball" the 9th seat.
We were going to subject the performance of the coxswain to a detailed technical analysis. I started with a tool I had used for years with the coxswains at the Marina Aquatic Center Rowing team in Los Angeles. We began by getting GPS data of each coxswain during practice. Using a running app easily accessible on their cell phones, each coxswain would record a map of their boat's course down the river on a standard 1000m piece. The coxswain would send me the map images for analysis. First, I would overlay a straight line (i.e., the ideal course) on their actual course to compare. They received a point against them each time they steered away and needed to correct back from this straight line. A perfect score (a perfectly straight line from start to finish) was a zero.
Below are some maps taken by our coxswains during practice. They were given a map score based on how many times they turned away from the ideal course. The red line is their GPS data tracked throughout the piece, and the white line is the ideal course Photoshopped in afterwards. The lower the points the better! The GPS Maps are accurate within around 5 meters. I primarily use them to gain an idea of how well the coxswain steers as a whole, as well as helping novice coxswains gain spatial awareness and an understanding of their course.
A sample map
While we were working on a reproducible score, we stayed with the classic coxswain tape. We continued to use our standard method for judging tapes. It breaks down each call into one of four categories: Commands, Information, Vocal Quality, and Motivation. Each category is given a grade from a scale of 0-5, giving the coxswain a total out of 20. Coxswains feeling extra motivated were also encouraged to transcribe their tapes and categorize each component. I set the goal that at least 80% of coxswain calls should be informational (i.e., stroke rate, split, time, distance, and boat positioning). These scores were further averaged over the 4 weeks to give each coxswain a tape score.
Things were going well as our summer camp experiment progressed! All of what we were doing was a significant step forward in helping the coxswains learn and grow in their craft. But behind the scenes, there was even more data crunching going on. My ultimate goal was more ambitious than improving skills. I needed to find a way to isolate the performance of the coxswain from the performance of the crew that would allow for coxswains to seat race with the same crew but would account for the variation in performance between pieces.
I broke things down into a classic word problem that you might find on a math test. It's only common sense that if you drove a car in a perfectly straight line at 60 mph, then you should travel 60 miles in one hour. If you drove in that perfectly straight line at 30 mph then you travel 30 miles in one hour. Both drivers drive perfectly straight - even though one car is faster than the other. If you replaced that perfect driver with one that could not steer straight, then even though the car went at 60 mph, it would take the driver longer than an hour to travel a straight line distance of 60 miles. The same would apply to a car with a bad driver going at 30 mph.
In each case, by knowing the speed of the car and the straight line distance it covered, you could determine whether the driver drove perfectly straight (60 miles took an hour) or swerved all over the map (60 miles took longer than an hour). This idea became the key to comparing boat drivers - coxswains!
Therefore, if a coxswain who steers perfectly straight (Super Coxswain) gets in the Varsity 8 and they fly down the 2K course at an average 1:30 split in a perfectly straight line, the 8 should finish with a time of 6:00 minutes. If the same Super Coxswain gets in the novice boat and they row at a 2:00 split, that 8 will finish a 2K in 8:00 minutes. In each case, Super Coxswain did a perfect job.
Now what if you are a novice Coxswain and you oversteer the boat, go off course, correct, and therefore add distance and time to the novice boat? Even though the novice crew still clocks at 2:00 split, the boat will finish a 2K race in over 8:00 minutes because novice coxswain made them travel further. If we put novice coxswain in that Super Varsity 8, she will add time to that boat as well. No matter what boat, and what crew a coxswain has, we can always compare their performance to perfection and thus come up with a way to measure the skill of the boat driver.
That was our key! By using the boat's average speed from the NK data and generating an expected time to cover a known straight line distance, we created a score to reflect how much extra time was added to the ideal by the boat driver. We named this measurement the CAT score (Coxswain Aptitude Test).
A sample comparison grid
Finally, a coach could isolate the Coxswain's effect on the boat! With one week left before we hit the road to Nationals, I furiously collected as much data from our V8 and top two coxswains as I could to see if this score demonstrated some validity with race results and a coach's eye.
Summer came and went fast, as they often do, and the Row West team found themselves in Ohio for US Rowing Nationals. Twenty-eight rowers and five coxswains went out on the course donning their light blue uni's. Through the combined efforts of our entire squad the girls earned themselves an impressive amount of medals and the Women's Points Trophy! We couldn't be more proud of them and as we flew back to LA, sunburnt and exhausted, there were only smiles and fond memories of our 4 week camp.
In my opinion one of the best things that came out of the experiment from Row West was one of our coxswains who had the top 8 by the end of camp. She came from a tiny boy's club team, with not much pedigree or medals to speak of. But through this new system we used, her numbers came out on top again and again. No coach was expecting much out of her when she signed on to the team. And in fact, she told me she didn't expect much from herself! But this process showed us what we were missing! It showed us what no medal or level of seniority could. That this girl could not only steer that boat like no one's business, but that she was also a competitor and a leader.
To say what we are doing is different would be an understatement. And to say that it is complete or perfect would be untrue. But if you take anything away from this, I would only ask that you take a moment to rethink what you thought you knew about the 9th seat and consider a data-driven approach. That free speed you've been looking for may be closer than you think!
Curious to learn more? Want to apply this same scientific approach to your own team? Contact me at firstname.lastname@example.org !