{"id":51234,"date":"2024-08-22T12:11:10","date_gmt":"2024-08-22T12:11:10","guid":{"rendered":""},"modified":"-0001-11-30T00:00:00","modified_gmt":"-0001-11-30T00:00:00","slug":"how-non-runners-skew-weighted-statistics","status":"publish","type":"post","link":"https:\/\/turnstonebookkeeping.co.uk\/TeaforFuel\/?p=51234","title":{"rendered":"How Non-Runners Skew Weighted Statistics"},"content":{"rendered":"<h2>The core problem<\/h2>\n<p>When you crunch numbers for a running cohort, you assume every data point reflects a stride, a heartbeat, a pace. Throw a handful of non\u2011runners into that pool and the whole thing starts wobbling like a loose shoe on a treadmill. The weighted averages you brag about? They become a mirage, a statistical illusion that flirts with reality but never quite lands.<\/p>\n<h2>Weighting without a filter is a recipe for disaster<\/h2>\n<p>Look: weighting is supposed to give the most active athletes a louder voice. You assign a factor\u2014say, a runner\u2019s weekly mileage\u2014to each participant, then multiply their performance by that factor. The math sounds clean. The reality? If you slip in a couch\u2011potato who logs zero miles, the weight collapses to zero, dragging the whole sum down. It\u2019s like adding a dead weight to a barbell\u2014nothing lifts, everything stalls.<\/p>\n<h2>Where non\u2011runners sneak in<\/h2>\n<p>First, data collection portals rarely ask \u201cDo you run?\u201d They ask \u201cWhat\u2019s your distance this week?\u201d A blank field or a zero entry gets interpreted as a valid data point. Second, social\u2011media challenges pull in followers who join for the badge, not the run. Their participation skews the denominator, inflating the participant count while deflating the true activity metric.<\/p>\n<h2>The ripple effect on key metrics<\/h2>\n<p>Average pace, for instance, drops dramatically. A 6\u2011minute mile group mixed with a bunch of zero\u2011minute entries spits out a 4\u2011minute average that no runner can achieve. VO\u2082 max calculations, which rely on mileage and heart\u2011rate zones, become nonsense. Even injury\u2011rate percentages get messier\u2014a year with one sprain looks harmless when a hundred non\u2011runners dilute the sample.<\/p>\n<h2>Statistical sanity checks you need<\/h2>\n<p>Here\u2019s the deal: before you apply any weighting, filter out entries with a mileage of less than 1\u202fkm. Use a binary flag\u2014runner vs. non\u2011runner\u2014and exclude the latter from the weighted formula. If you must keep them for demographic reasons, compute a separate \u201cinactive cohort\u201d metric and report it alongside the active stats. Transparency beats a polished but hollow figure every time.<\/p>\n<h2>Impact on predictive models<\/h2>\n<p>Machine\u2011learning pipelines love clean data. Feed them a mixed crowd and the model learns patterns that don\u2019t exist in the real running world. It predicts \u201caverage weekly distance\u201d based on a dataset where 30\u202f% of users never lace up. The output? A forecast that tells sponsors the community is \u201clow\u2011intensity,\u201d jeopardizing funding.<\/p>\n<h2>Real\u2011world example from <a href=\"https:\/\/nonrunnerstomorrow.com\">nonrunnerstomorrow.com<\/a><\/h2>\n<p>A 10\u2011km race organizer posted a post\u2011event report showing a 12\u202f% increase in average speed year over year. Dive deeper: the previous year had 200 participants, 50 of whom logged zero miles during registration. The weighted calculation ignored the zero\u2011mile entries, falsely boosting the average. The truth? The active runners actually ran slower than the prior year.<\/p>\n<h2>Take action now<\/h2>\n<p>Audit your dataset. Strip out anyone whose activity flag is off. Re\u2011run the weighted stats. You\u2019ll see the numbers settle into a realistic range. No more ghost runners. No more misleading headlines. Cut the noise, let the real pace shine. Start with a quick script today.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>The core problem When you crunch numbers for a running cohort, you assume every data point reflects a stride, a heartbeat, a pace. Throw a handful of non\u2011runners into that pool and the whole thing starts wobbling like a loose shoe on a treadmill. The weighted averages you brag about? They become a mirage, a [&hellip;]<\/p>\n","protected":false},"author":86,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[],"tags":[],"class_list":["post-51234","post","type-post","status-publish","format-standard","hentry"],"_links":{"self":[{"href":"https:\/\/turnstonebookkeeping.co.uk\/TeaforFuel\/index.php?rest_route=\/wp\/v2\/posts\/51234","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/turnstonebookkeeping.co.uk\/TeaforFuel\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/turnstonebookkeeping.co.uk\/TeaforFuel\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/turnstonebookkeeping.co.uk\/TeaforFuel\/index.php?rest_route=\/wp\/v2\/users\/86"}],"replies":[{"embeddable":true,"href":"https:\/\/turnstonebookkeeping.co.uk\/TeaforFuel\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=51234"}],"version-history":[{"count":0,"href":"https:\/\/turnstonebookkeeping.co.uk\/TeaforFuel\/index.php?rest_route=\/wp\/v2\/posts\/51234\/revisions"}],"wp:attachment":[{"href":"https:\/\/turnstonebookkeeping.co.uk\/TeaforFuel\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=51234"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/turnstonebookkeeping.co.uk\/TeaforFuel\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=51234"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/turnstonebookkeeping.co.uk\/TeaforFuel\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=51234"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}