Introducing Pace Partners for FLRC Challenge runners

Last week, we announced the option for groups to have their own FLRC Challenge group pages on the leaderboard, and that remains generally available—just contact me with the name and members to add.

This week we have another new leaderboard feature aimed at helping @Challengers identify compatible running partners. Quite a few people have made new running friends through the Challenge already, and with our new Pace Partners feature, finding people to run with will become even easier. Here’s how it works.

On your Athlete page (click your name anywhere on the leaderboard), click the new Similar Athletes / Pace Partners button to view the list of people who run at similar paces.

Brave Browser.app 2021-07-13 at 16.15.40

Here’s what the top of mine looks like. The order reflects who the algorithm thinks would be the most appropriate running partners, and the color-coding indicates who is likely faster (shades of yellow) or slower (shades of blue). I know most of these people, and the list is pretty accurate for what I’m running right now.

Five of the columns require some explanation:

  • Fastest Pace: This metric averages your fastest times for each course and identifies people whose fastest times are 10% faster or slower.
  • Average Pace: Similarly, this metric takes your average pace for each course and pulls out those whose average paces are within 10% of yours. You’ll see “no data” here if either of you lacks enough runs to calculate an average pace for the courses that you’ve both run. (On your Athlete page, you’ll see “not enough runs” for the Average number.)
  • Similarity: This metric attempts to measure how similar someone is to you based purely on times.
  • Overlap: This metric reflects the percentage of courses for which you and the other person have fastest and average times. The higher the number, the more courses you have in common.
  • Rank: The leftmost column, Rank combines the unweighted Similarity score with a weighted one that also takes Overlap into account. It’s the default sort because it attempts to avoid limitations in each of the other metrics.

What are those limitations, and why might you still want to click the column headers to change the sorting?

  • Fastest Pace limitations: The problem with Fastest Pace is that the leaderboard has no way of knowing whether it reflects your actual fastest pace or if your fastest effort might have been an easy run. In theory, as the Challenge continues and more people record their hardest efforts on more courses, Fastest Pace should become more accurate. However, sorting by Fastest Pace can still give you a sense of who is faster and slower.
  • Average Pace limitations: The main limitation here is that too many people won’t have enough runs to calculate the Average Pace and will thus end up with “no data.” It should improve as more people run more courses such that their Average Pace numbers coalesce.
  • Similarity limitations: This metric can be fooled by people who have very similar fastest times on a course or two but very little overlap.
  • Overlap limitations: With too little overlap, the leaderboard is essentially comparing apples and oranges. The main utility of this metric is to inform the rank and give you a sense of who has run a roughly similar number of courses and efforts.

A vast amount of data and math has gone into these calculations, and it’s not worth diving too deep into the details, since only @steve-desmond can really explain it at a technical level. (But if you’re interested, just ask and he will. :slight_smile: )

The takeaway is simple. If you’re looking for a running partner, consider reaching out to some of the people on your list. When we post a group run here on the FLRC Forum, for instance, you could invite some of your pace partners to join by typing an @ sign and then a few characters of their name to get the autocomplete on their forum nickname, like @adamengst for me. (And if I’m on your list and healthy enough to run, I’m usually happy to run with you.)

Let us know what you think!

1 Like