Other than Olympic medals and world championships, it’s arguably American swimming’s most notable achievement. It has blossomed over the last 15 years into a veritable gold mine of information. Like Facebook and Google, it has grown from its American roots to worldwide ubiquity. In fact, it has become so commonplace that within the swimming community that it resembles air – noticeable only in its absence.

SWIMS, the database launched in 2003 to quell the complaints of swim coaches, is now used for everything from determining a 10-year-old’s rank in the LSC in the 50-yard butterfly to settling on qualifying standards. As for the future, well, who can predict with any confidence the outcome of machine learning, where databases are mined electronically for patterns that humans have neither the capacity nor the patience to perceive?

A Whopper

And SWIMS is one whopper of a data base. Any timed swim done just about anywhere in the world is in there. The answer to nearly any question one could imagine about competitive swimming over the last two decades is just a few keystrokes away.

For example, who was the second best American 18-year-old woman in the world in 2017 in the 400-meter freestyle?

Easy. Courtney Harnish – 10^th with a time of 4:12.05, which she did at the Arena Pro Swim in Atlanta.

Hold on, you are undoubtedly thinking, hadn’t she swum better in that event?

She had, which you can easily find out by checking the individual times search, a portion of SWIMS that drives some of the highest number of visits to the USA Swimming website.  At the U.S. Nationals in 2015, the University of Georgia sophomore and former Junior National Team member did a 4:08.22, which ranked her first in the world for 16-year-olds.

And for those with an insatiable curiosity about all things swimming, she negative split the race – out in 2:04.53, back in 2:03.69. Her fastest 50 other than the first one (29.37) was the last one (30.11). You need to get up very - very - early in the morning to put anything over on SWIMS or its offshoots, such as Deck Pass and LSC Portal.

Machine Learning

Next up for SWIMS is to go from answering questions about what has happened in swimming to what will happen. A rather soft touch of this was shown in September with the announcement of 2020 Olympic Trials standards. The qualifying times, which should lower participation by some 300 swimmers from the 2016 Trials, were determined largely by analyzing the progression of times in the U.S. over the last two Olympiads.

A more intricate use of SWIMS and machine learning, according to Larry Herr, USA Swimming’s director of times and statistics and one of the creators of SWIMS, will be analyzing the backgrounds of current and recent Olympians and discerning patterns that can be used to help the next generation of swimmers and their coaches determine the best courses to follow, such as how many meets per year did they compete in when they were younger? How many attended a Zone Select Camp? When did they first make a junior national standard? When did they start to focus on their strongest events? Etc., etc. The questions are endless.

Any Swim, Any Time

Herr finds it hard to believe that what he and others at USA Swimming started working on 20 years ago has evolved into ground zero for swimming’s statistics and trends. The concept arose out of complaints by coaches in the 1990s over standing for hours in lines at major championships meets waiting to verify the times of their swimmers. The obvious solution was an easily accessible library of meet results, but what seems so simple now was fraught with hurdles in a time when most organizations were ill-equipped technically to handle mountains of data.

Herr recalls 30 different databases within USA Swimming, each with unique requirements and formats. Merging them into a cohesive whole that could be tapped internally and externally would be a Herculean task.

Then came the additional challenge offered up by an Olympic coach.

“You need to collect not only the times done in the top meets across the country, but every time in every meet,” he said to Herr and others at an informal meeting at USA Swimming headquarters in Colorado Springs in 2002. “When Michael Phelps wins a lot of gold medals, you are going to want to know the times he and other Olympians did when they were 10, 11, and 12 years old. You have to think bigger picture.”

Just like that, the vision for SWIMS expanded from quickly verifying the times of swimmers at championship meets to collecting times of all swimmers from the beginning to the end of their competitive careers. Herr sensed immediately the challenge would extend beyond the staff at USA Swimming. Dedicated volunteers in each LSC would have to collect and verify the times. Third-party vendors would have to set aside their instinct to compete and collaborate instead to ensure efficient and accurate transmission of data. The chain of custody would have to be inviolable. If one brick falls out of the wall, the integrity of the whole system could be compromised.