To do a weight loss challenge with friends, agree the rules before day one. Score each person's weight loss percentage, not kilograms. Compare the average of the first week with the average of the last week, not two single mornings. Run it for four weeks or more, and score habits such as days logged beside the scale.
TL;DR
- Scored in kilograms, the heavier friend wins on the same effort: losing 0.5% of body weight a week for four weeks is 1.89 kg at 95 kg and 1.19 kg at 60 kg, both 1.99%.
- In 100,000 simulated 28-day challenges, when two friends' true losses are 0.4 kg apart, the one who lost less wins about 25% of the time if the score compares the first morning with the last.
- Comparing the first week's average with the last week's cuts that to about 8%, or from 31% to 16% if mornings are noisier.
- In two studies (1999 and 2012), people who did it with friends or teammates lost more, though neither shows competing was the cause.
- Score each person against their own target, not the lowest number: a rule that rewards the biggest drop rewards under-eating.
How do you do a weight loss challenge with a friend?
A weight loss challenge is a fixed number of weeks in which a few people weigh themselves and compare results. An accountability partner, or weight loss buddy, is someone who sees your numbers regularly; a challenge adds a score. The arguing comes at the end, so settle five things first:
- When to weigh. Every morning, or the same days for everyone: on waking, before food, on the same scale.
- The score. Percent of starting weight, never kilograms.
- Which readings count. The average of the first seven mornings against the average of the last seven. A week average is just that: seven readings added up and divided by seven.
- How long. Four weeks or more, and not across a festive week.
- What else scores. Days logged and days within your own target, beside the scale.
This post is about keeping score. It says nothing about how much anyone should lose, or how fast.
Should a weight loss challenge be scored in kg or percent?
In percent. Your weight loss percentage is your starting weight minus your current weight, divided by your starting weight, times 100. Kilograms favour whoever starts heaviest, because the same effort moves a bigger body further.
Take two friends, one at 95 kg and one at 60 kg, who each lose 0.5% of their current weight every week for four weeks. That pace is an arithmetic example, not a recommendation. The first ends at 93.11 kg, down 1.89 kg. The second ends at 58.81 kg, down 1.19 kg. Both lost 1.99% of their starting weight, so on percent it is a tie; on kilograms the heavier friend wins by 0.69 kg for the same effort.
In an office weight loss challenge or any weight loss competition, a kilogram scoreboard is mostly a list of who started heaviest.
How often does the scale pick the wrong winner?
More often than most groups would accept. A scale moves for reasons that are not fat: water, salt, food still in transit. In one man's 9,521 standardised morning weigh-ins, the change from one morning to the next had a standard deviation of 0.53% of body weight (Schneditz, Renal Failure, 2023), the same figure I used in how to find your maintenance calories. If mornings are independent of each other, that puts a single morning about 0.375% off the true weight: 0.3 kg at 80 kg.
I ran 100,000 simulated 28-day challenges per case. Each friend's true weight falls in a straight line, each morning adds random noise of that size, and the challenge is scored two ways. Figures in brackets use the noisier 0.53% per morning.
- First morning against last morning
- Two friends whose true losses are 0.4 kg apart at 80 kg (0.5 percentage points): the one who lost less wins 25% of the time (31%).
- 0.2 kg apart: 37% (40%). 0.8 kg apart: 9% (17%).
- Six friends with true losses of 1.0, 1.5, 2.0, 2.5, 3.0 and 3.5%: the true leader wins 72% of the time (61%), and the whole order comes out right 16% of the time (7%).
- First week's average against last week's average
- Two friends 0.4 kg apart: the one who lost less wins 8% of the time (16%).
- 0.2 kg apart: 24% (31%). 0.8 kg apart: 0.2% (2.4%).
- The same six friends: the true leader wins 92% of the time (83%), and the whole order is right 64% of the time (36%).
Length matters as much as the rule. With the same small difference between two friends, 0.125 percentage points a week, the wrong friend wins on week averages 32% of the time over two weeks, 8% over four and 1% over six (36%, 16% and 5% with noisier mornings). Two weeks cannot separate people who are close.
The week average does not fix everything. Schneditz's spread grows from 0.53% over one day to 0.69% over seven, so part of the wobble lasts for days and mornings are not truly independent: hence the noisier figures. Weight also swings about 0.35% within the week, higher after the weekend (Turicchi and colleagues, 2020, covered in the same post), so everyone should weigh on the same days. And in Malaysia, do not start or end a challenge across Hari Raya, Chinese New Year or Deepavali open-house weeks, which move everyone's scale by different amounts.
Does losing weight with friends actually help?
Two studies point that way; neither proves that competing is the reason. In a 1999 trial, 166 people were recruited alone or with three friends or family members, then randomly assigned to a standard behavioural programme or the same with social support strategies. Those recruited with friends lost more weight by the end of the 4-month programme and at month 10. Recruited alone on the standard programme, 76% completed and 24% kept their whole loss from month 4 to 10; recruited with friends and given support, 95% and 66% (Wing and Jeffery, Journal of Consulting and Clinical Psychology, 1999, abstract). Those are the trial's two ends; the abstract does not split the gap between the friends and the support.
The second is a team competition. Shape Up Rhode Island 2009 was a 12-week online programme in which teams competed on weight loss or activity. Its 3,330 overweight or obese completers, on 987 teams, reported losing 4.2% of their starting weight on average. Results clustered within teams, and people who reported more influence from teammates lost more and were more likely to reach a 5% loss (odds ratio 1.20) (Leahey and colleagues, Obesity, 2012, abstract). It is observational and self-reported: teammates and results moved together, which is not proof that one caused the other.
What can you score besides the scale?
The habits each person controls. A 2019 trial called STEP UP gave 602 adults a wearable step counter for 24 weeks. Each person picked their own step-goal increase, and the game awarded points and levels for hitting that goal. Compared with the control group, daily steps rose by 920 in the competition arm, 689 in the support arm and 637 in the collaboration arm; in the 12-week follow-up, only competition stayed significantly above control, at 569 (Patel and colleagues, JAMA Internal Medicine, 2019, abstract). It measured steps, not weight. The point is the design: people competed on hitting their own goal, not on who walked most.
Three habit scores borrow that design, for a diet buddy pair or a group:
- Days logged: the share of days you recorded what you ate.
- Calories on target: the share of logged days within 10% of your own calorie target, either side.
- Protein target met: the share of logged days reaching your own protein target.
Score the target, never the lowest number. "Fewest calories" or "biggest deficit" is a rule that rewards under-eating, and a prize for the biggest drop rewards whoever cuts hardest in the final week.
A worked example with three made-up people, Aina, Ben and Chong, over one week, each with a made-up target of 1,800 kcal and 90 g of protein. Within 10% means 1,620 to 1,980 kcal; 1,619 and 1,981 are out. Aina logs all seven days, six in the band: 86% on calories on target. Ben logs all seven at about 40% under target and scores 0%: far under is a miss, not a win. Chong logs three days, all in the band: 100% on calories, 43% on days logged. Without days logged as its own score, whoever logs least can top the calories table.
What I built for this, and what it does not do
I built Circle, part of the Fitness module in Habitus Mind; it went live on 7 October 2026. A circle holds up to six people. The leader invites each with a single-use code, shared by WhatsApp link or email. The member list shows display names and join dates, nothing else. A member can send another a food or exercise line, which lands in their diary only if they accept it.
The Leaderboard tab has four boards: Days logged, Calories on target, Protein target met and Towards my target weight. Each is off until the member switches it on. A board covers a Malaysia week, Monday to Sunday in Malaysia time, or a calendar month, this one or the last. Only percentages cross between members: never food, weights or targets. The scoring is the arithmetic above; I built it so that eating far under your target scores zero, not first place.
The weight board, Towards my target weight, shows how far each member has moved from the period's first weigh-in towards their own target weight. From 90 to 88 kg with a target of 80, and from 62 to 61.2 kg with a target of 58, both show 20%, although one lost 2.0 kg and the other 0.8 kg. Moving away from the target shows a negative number: 70 to 71 kg with a target of 66 is -25%.
Its weak spot, said plainly: it compares the first and the latest weigh-in of the period, which is the first-morning-against-last-morning rule from the simulation. For someone truly losing 0.5% a week and weighing every morning, the weekly board shows them moving away from their target about one week in five (21%, or 28% with noisier mornings); over a month, almost never. Read the weight board over a month, not a week. Circle runs no prize, no money and no start date: its members agree those themselves.
Do I need an app for a weight loss challenge?
No. A shared sheet does it. Give each person one row per week with that week's average weight, and a percent column: the first week's average minus this week's average, divided by the first week's average, times 100. Add a days-logged column if you score habits.
Habitus Mind is free to start. Circle and its four boards live in its Fitness module, in the free plan; to try it, create a free account.
FAQ
How do you do a weight loss challenge with a friend?
Agree the rules first. Everyone weighs on the same mornings, the score is percent of starting weight, and the readings compared are the first week's average and the last week's. Run it for four weeks or more, keep it clear of festive weeks, and score days logged beside the scale.
Is it fairer to score a weight loss challenge in kg or percent?
Percent. Two people who each lose 0.5% of their weight a week for four weeks lose 1.89 kg at 95 kg and 1.19 kg at 60 kg, both 1.99% of their starting weight. Scored in kilograms, the heavier friend wins on the same effort.
How long should a weight loss challenge with friends last?
Four weeks or more. In my simulation, two friends whose true rates differ slightly swap places on week averages 32% of the time over two weeks, 8% over four and 1% over six. Over a shorter challenge, ordinary day-to-day scale noise decides too many results.
Does losing weight with a friend actually help?
Two studies point that way. In a 1999 trial, people recruited with friends lost more weight at 4 and 10 months. In a 2012 study of 3,330 people in a 12-week team competition, more reported teammate influence went with more loss. The second is observational and self-reported.
What should a weight loss challenge score besides weight?
Habits each person controls: days logged, days within 10% of their own calorie target, and days reaching their own protein target. Score the target, never the lowest number, because a rule that rewards the biggest deficit rewards under-eating.
