Common questions

Before you submit.

Does the game use real market data?

No. Instruments, hidden values, news, signals, and order flow are synthetic. The competition is about decision-making in a market mechanism, not predicting a real security or extracting a trading strategy from entrants.

Who owns my bot?

You do. Entrants retain their intellectual property and grant the limited permission needed to run the submitted artifact and publish competition results and public replays.

Can another entrant download my submission?

No. Public replays show match-only pseudonyms, public market events, and post-match statistics. They do not contain submission source code or private signal streams.

Why is the order book anonymous if trades are attributed?

Resting ownership would reveal intent before anyone trades. Match-only attribution on completed trades lets strategies learn from observed behavior after risk has actually changed hands. Pseudonyms reset between matches.

Am I told how good my signal is?

No. A FACTOR seat sees a noisy estimate of the common component that moves both stocks. An IDIO seat sees noisy estimates of the two stock-specific components. You learn which family your seat has, but its noise level is hidden and redrawn each match. See the private-signal explanation.

Why reveal the seed after completion?

The seed is needed to re-execute and verify the result. It is withheld before and during the match so public knowledge of the engine does not reveal future news, signals, same-tick ordering, or the hidden end tick.

Will a local match behave like an official match?

Given the same engine version, parameter manifest, seed, seat configuration, and bot messages, FAST and STANDARD reproduce the same result. Official results name those versions in their replay.

What happens when my bot is slow?

It falls behind. Messages queue and later decisions use older packets. More than five seconds of cumulative overrun suspends the seat for that match, cancels its resting orders, and lets its position settle.

Can I compete in both leagues?

Yes. You may keep one active submission in FAST and one in STANDARD. They may use the same code or different code, but each has a separate rating.

Can I resubmit?

Iteration is expected. The staging interface allows replacement submissions, but its ratings are test-only. Rating carry-over and uncertainty changes for a public season will be defined in that season's published terms.

What determines my leaderboard position?

Matches produce six-way competitor placements. OpenSkill updates μ and σ from those placements, and the leaderboard displays the conservative rating μ − 2σ.

Where should I start?

Install continual-exchange-sdk and follow Your first bot to write and package a bot. Local matches and replay verification become available when the engine source repository opens. Once they do, start in STANDARD, run many seeds, and inspect the replays rather than judging a policy from one match.