How Many Checkpoints Does an Event Need?

"How many should we put out?" is the first question every organizer asks, and for a long time our answer was a shrug and a guess based on venue size. We now have six finished deployments with full on-chain records, ranging from an afternoon with 20 players to a three-day fair with 351. Enough to stop guessing.

Here is every event we have run, with the two numbers that matter: how many checkpoints went out, and how hard each one worked.

The six deployments

• NOVA IMS Spring Bootcamp, April 7, one afternoon: 42 checkpoints, 20 players, 220 finds. 5.2 finds per checkpoint.
• ETHDenver, February 18 to 21, four days: 59 checkpoints, 207 players, 992 finds. 16.8 finds per checkpoint.
• Future Maker at NOVA IMS, March 17 to 19, three days: 115 checkpoints, 265 players, 2,591 finds. 22.5 finds per checkpoint.
• Data with Purpose Summit, June 25, one day: 25 checkpoints, 58 players, 777 finds. 31.1 finds per checkpoint.
• Festival do Cadaval, May 22 to 23, two evenings: 57 checkpoints, 126 players, 2,182 finds. 38.3 finds per checkpoint.
• Portugal Smart Cities Summit, May 12 to 14, three days: 103 checkpoints, 351 players, 8,123 finds. 78.9 finds per checkpoint.

The ratio, not the count

Sort those by checkpoints per player and something falls out. Five of the six land between 0.29 and 0.45 checkpoints per player. ETHDenver and Smart Cities both sit at 0.29, Data Summit and Future Maker at 0.43, Cadaval at 0.45. Different venues, different audiences, different event lengths, same band.

The sixth is Spring Bootcamp at 2.1 checkpoints per player, and it is also the event where each checkpoint worked least: 5.2 finds. 42 tags for 20 people meant most tags were found by a handful of players, and 620 of the 840 possible finds never happened. The tags were not wrong. There were just too many of them for the crowd.

So the useful planning number is not a checkpoint count. It is roughly one checkpoint for every two to three people you expect to play, and then a sanity check on whether the venue can physically hold them that far apart.

Two things that move the number

Venue density. ETHDenver had 59 checkpoints and got 4.8 finds per player across four days. Smart Cities had 103 and got 23.1 per player across three. The difference is not the game, it is the walk. ETHDenver sprawled across a large campus with attendees who had panels to get to. Smart Cities was one exhibition hall where every checkpoint was within a few minutes of the last one. A dense floor lets people collect; a spread-out one turns each find into a trip.

How long people stay. At Smart Cities, 278 of the 351 players came for one day only, 35 came for two, and 7 for all three. If most of your audience is there for one day, the hunt has to be completable in one day, which caps how many checkpoints are worth deploying no matter how big the venue is.

What to do with a small crowd

Spring Bootcamp is the case worth learning from, because small events are the ones most likely to over-provision. With 20 players, fewer checkpoints placed closer together would have produced more finds, more collisions between players, and a tighter race. The final margin was three finds, which was good; but it happened despite the layout, not because of it.

If you are running something under 50 people, start nearer 15 to 20 checkpoints and make them dense. You can always hide more mid-event. Recovering from a floor that feels empty is much harder.

The one number we would check first

If you want a single diagnostic after an event, use finds per checkpoint. Under about 10 and you over-provisioned. Between 20 and 40 is a healthy floor. Above 70, as at Smart Cities, and you probably had room for more checkpoints than you used.

Every figure above comes from published reports: ETHDenver, Future Maker, Portugal Smart Cities Summit, Festival do Cadaval, Data with Purpose Summit and Spring Bootcamp. For how a deployment is planned from the organizer side, see event gamification.