Abu Dhabi Pickleball Club · a case for public courts
Pickleball is a paddle sport, usually played in doubles on a court about a quarter the size of a tennis court. It arrived in Abu Dhabi as one Tuesday morning pickup game in April 2025. Sixteen months later, the club has 505 players, coordinating play across five borrowed venues, none of them built for the sport and none of them the club's own. The group's scheduling data shows the demand already here.
Evidence 01 · Growth
Members in the group, every day since it opened on 20 April 2025. Dots mark each hundredth member; the brackets underneath show how many days each hundred took to arrive.
In sixteen months membership has never once shrunk on a month-over-month basis: people join, a few leave, but every month has ended higher than it started. Growth is getting faster. The next 200 people joined in about half the time:
Net members added each month: joined minus left. A taller bar means more members were added.
The slower months, from September to December 2025 and in February and March 2026, were also the months with the fewest organised sessions. This points to a shortage of court time, rather than a lack of interest.
Evidence 02 · Utilisation
Four terms are used below. Each one is a count of something in the message record, not an estimate.
Reading the dip
Sessions fall away between November 2025 and March 2026 while the group itself stayed busy, with 561 to 1,192 messages a month in that period. Play was being arranged through recurring pinned posts rather than fresh polls, so the poll count misses that period rather than showing a real lull in play.
The schedule grew 40 times over. One organised session in June 2025; 46 in May 2026; 39 in each of the three months since. That is better than one session a day, every day, on courts borrowed from other sports.
It isn't just more members. Measured against the size of the group, the schedule still grew six times over: 1.3 sessions per 100 members in June 2025, 7.9 per 100 in July 2026. Members joined fast; play per member rose faster.
Each session is a crowd, not a casual hit. The average session collects 12.6 sign-ups. A pickleball court holds four players. On average, that is enough people for three courts at the venues the group can borrow.
The limit is already binding. Of 129 session slots that stated a player cap, 26 filled it exactly and 15 went over it. In those cases, people signed up after the organiser had announced the limit. Waitlists are mentioned 116 times and published as numbered lists 68 times.
Membership numbers alone do not show who wants to play. The group uses player caps, numbered sign-up sheets, and waitlists “in case someone drops out.” These are clear signs that the number of available courts limits play.
Evidence 03 · The limit
368 session slots with a stated start time, from 20 April 2025 – 8 August 2026, were read from 253 polls. Some polls offer several slots at once. How that's counted ↓.
This counts every hour a session was actually running, not just the hour it started. For example, a 7:30–9:30pm slot counts toward both 7pm and 8pm. That shows how many sessions were happening at once, through the day. A slot with no stated end is assumed to run 2 hours.
The community mainly plays in two short time periods. 320 of 368 slots (87%) start at 5pm or later, and another 44 start between 5 and 9 in the morning. Just 4 fall anywhere between 10am and 4pm, and that is forced by the weather, not a preference: from May to September an uncovered court in Abu Dhabi is unplayable in the middle of the day. Both of the windows the group actually uses are windows in which an unlit, unshaded court is worth very little.
Evidence 04 · The venues
Sessions naming each venue in its scheduling poll. A relocated session can name two venues, so these do not sum to 253.
Two venues carry almost all of it, and neither was built for pickleball. Sadim and Al Masar between them cover the large majority of named sessions; Al Masar’s court is a tennis court used for pickleball instead, and other courts have been found by searching out unused park courts and taping pickleball lines onto them.
The lights aren’t the club’s to keep. Sadim is floodlit, and games there have run past midnight under the lights. The floodlights belong to the park, not the club. When officials switched off lighting across Khalifa City’s parks for over a week (24 March–1 April 2026), evening sessions at Sadim and Al Masar went dark together regardless of what the poll said, and organisers spent the week routing players to whichever park still had power.
Weather closes the same two venues just as fast. Rain has taken out Sadim and Al Masar together too, damaging the court surface and forcing a same-day move between them. Neither closure is something the club has any say over.
Nothing is permanent. Nets and paddles are hand-carried in and packed up every session, and play gets relocated whenever the borrowed court is already occupied by other park users.
Lighting the club controls. 87% of session slots start after 5pm, and the single most common start time is 7pm, in the dark. Borrowed floodlights work until a park operator switches them off. This happened across Khalifa City for over a week in March 2026, and the club could not change it.
Shaded or covered play. Four session slots out of 368 in the 10am–4pm window is not a preference, it is heat. Shading existing courts converts six dead hours a day into usable ones without building a new facility.
Enough courts to stop capping sessions. Organisers are already turning players away at 12 per slot, with published waitlists. Play is spread across five borrowed venues; a single permanent home would bring together a community that has grown from 13 members at the start to 505 in sixteen months.
Everything here is drawn from the exported message history of a single community WhatsApp group, from 20 April 2025 to 8 August 2026. It contains 38,166 lines, 13,609 messages written by members, and 943 membership events. No message content and no member’s identity appears anywhere on this page. Every figure is a count, not a quote. Membership is built up from every add, invite, leave and removal event; this method comes out five above the 505 members WhatsApp itself reports, a difference in how the group’s first members are counted, and 505 is the figure used here.
Sessions are counted as scheduling polls posted in the group, which misses play arranged by other means and misses more of it between November 2025 and March 2026. Sign-ups are votes cast on those polls; a vote is a stated plan to attend, not a count of who actually showed up. Session start times show when play was planned, which is the closest available measure of when it happened. See below for how times are read. Venues are counted when one of five known place names, the parks this group has used, appears in a session’s poll text. A session can name more than one when play was moved.
How session start times are read: most polls give one time for the whole session. But 80 of the 253 polls, almost all from 30 April 2026, when evening polls began listing 5:30pm, 7:30pm and 9:30pm as separate choices in the same poll, give more than one time. Each choice is counted as its own slot with its own sign-ups, instead of the whole poll being counted once. Of the 368 slots with a stated time: 163 give a clock time with am or pm (“7–9:30pm”) and are used exactly as written; 205 give a plain range such as “7 to 9.30” together with the word morning or night, often in the poll’s own title (“SADIM NIGHT”) rather than in the choice itself. This shows which one it is. A further 17 polls name a time of day only (“Sunday morning”, “under the lights”) with no clock time, and 39 give no time at all; neither group appears on the chart. Numbers that are not times are taken out first: dates, scores, and player limits (“MAX 12 MEMBERS”). An earlier version of this count included those numbers and wrongly showed games starting at 3 and 4 in the morning. No session slot in the record actually starts later than 9pm or before 5am; the 9pm cluster is real and repeats often, not a mistake. The chart above counts a slot in every hour it runs, not just the hour it starts: when a poll states an end time that's used directly (“7–9:30pm” runs through the 7 and 8 o’clock hours); when only a start is given, the slot is assumed to run 2 hours, matching all but a handful of the sessions that do state both ends. Back to the chart ↑
This is one group, not a full count of everyone who plays the sport in Abu Dhabi.
Other clubs and informal play are not included, so these figures should be
read as the least demand could be, not the whole picture. August 2026 covers eight
days and is drawn
dashed. Every figure on this page is produced by analysis/extract.py
in the Space repository, run against the export.