VALUE DECENTRALIZATIONPowered by Frontier ProtocolRestoring session…VALUE DECENTRALIZATIONPowered by Frontier ProtocolRestoring session…Five boats share one fishery and cannot see it. Write a mission and a wallet policy, then watch agents sound the grounds, read each other, and pay for restraint under escrow.
01 · THE MISSION
Five boats share one fishery and none of them can see it. A ground is known only if somebody worked it, and what they learned ages from the moment they looked away. Fuel is issued for the whole season, not the round — and nobody is told how long the season is.
RULES IN 30 SECONDS
02 · THE VALUE MARKET
The median solvent boat's cash at the end. Cares that people were paid, not that the fishery was left pretty.
Of the fish you gave up against a boat that takes everything, how many were still in the water at the end. Leaving a ground a greedy rival is steaming towards buys nothing.
What your contracts changed, per DemoUSD you paid for it — measured by replaying the same season with your agreements removed.
03 · WRITE A MISSION
An entry is a name, a standing instruction, the wallet the match loop enforces, and how the boat works a ground without a model. The three tabs edit one and the same thing, so a dial moved here shows up in the JSON and a file dropped in moves the dials. The form alone reaches 98,762,112 distinct entries; the JSON editor is not bound by the sliders' steps, so it reaches more.
A model-backed boat reads this text and nothing else about your intent. One season has measured between three and six minutes — the model answers twice a round, once to negotiate and once to fish, and how long it thinks is its own business. The page stays usable while it runs.
The scripted preview below is instant and costs nothing, so it shows the world reacting. Sailing your mission shows your words being interpreted.
04 · WATCH THE SEASON
Putting to sea…
05 · SOLUTION LANDSCAPE
Sailing your entry and the three references…
Each approach sails the same 7 seasons; the figure is its median and the line beneath counts the seasons it took that axis. An approach stays on the frontier unless another beats it on all three at once, and more than one usually survives — there is no single best season, only seasons that are good at different things. Watch the season counts rather than the medians: a lead held in one season out of seven is the draw, and a lead held in five is a habit.
06 · VALUE ALLOCATIONS
Fishing families and the crews they employ.
Sailing…
Everyone who will fish this water next season.
Sailing…
Those who fund restraint rather than perform it.
Sailing…
No pool is funded, because one season decides very little. Measured across 7,200 scripted seasons, the draw accounts for 20–38% of an outcome and a fixed policy for only 3–8%; the rest — 54–77% — is which policy suited which season, and that is the part only a skipper who reads the water can reach. Picking the right approach per season, with hindsight, beats the best single approach by 36% to 286%. So a single result here is closer to one hand of poker than to a race, and nothing is ranked on one.
07 · SUBMIT
Everything above this line runs in your browser on a season you picked, which is practice and worth exactly what practice is worth. A submission is scored on the server over 12 published seeds, all of them, with the wallet policy enforced by the match loop rather than by the boat. The three outcomes come back separately and are hashed together; nothing adds them up.
Nothing submitted yet.
The sandbox identifies an entrant by wallet only, which is not proof of personhood, and keeps nothing past a restart. A revision never overwrites the one before it: every submission keeps its own result hash, so a worse one cannot be quietly buried.
08 · WHAT YOU HAVE TRIED
Nothing sailed yet.
Nothing here is saved beyond this visit. Change the approach or the season above and the run is added to this list, so you can see whether an approach holds up across seasons or only won the one you happened to look at.
09 · EVIDENCE
Four reference entries sailed the 12 scored seeds. The contract was told what each one scored, worked out the frontier itself, and kept 2 of 4 — nothing beats either on all three at once. Then each pool paid its own axis.
Fill the hold10,000 FDT
Landed the most fish, and gave up nothing to do it.
0x1ad10a1df3… ↗Work the season10,000 FDT
Left fish in the water and they were still there at the end.
0x40d69c91b3… ↗Work the season10,000 FDT
Its agreements changed the season; the closed wallet's could not.
0x639d79dbac… ↗Two pools backed the same entry and one did not, which is the whole point: the boat that landed the most fish gave up nothing and signed nothing, so it served one axis and no other. Had the contract been able to rank, that disagreement would have been averaged away into a single winner and the result would have said less than it does.
The settlement contract has no function that ranks entries and none that adds the three outcomes together. It records what each entry scored, marks everything nothing beats on all three at once, and lets each of the three pools back whoever served its own outcome — so three funders can and do support three different skippers. It also refuses to record a result from fewer than three seasons, because one season is mostly the draw and paying out on it would be paying out on the weather.
FDT are demonstration credits on a test network and make no monetary claim. The engine resolves the sea; an agent only chooses where to fish and what to offer.