Offline computing

Not long ago I was playing the computer game “V Rising” when I came across a surprising feature. In the game you are able to have followers gather resources after a set period. What’s interesting is how the feature was implemented. You might think that they used in-game time to measure this – but instead it was tied to real-life time. This meant that if the host shut down the server the amount of time was applied as if the followers were still working when the server was off.

When I saw that feature I couldn’t stop thinking about it. I mean: isn’t there something extremely weird here? Normally computer programs aren’t supposed to be able to run when the power is off. This goes against every instinct there is for how software works. Granted, the effect is an illusion – the algorithm isn’t really “running” offline – the server only adjusts the results when it’s back on. But I believe this is still an absolutely whimsical way to think of software.

It means that there are systems for which results don’t have to be given in real time and that further – there are interesting trade-offs. For instance, imagine a novel securities market based on the V Rising gathering design. You might allow algorithms to control resources, defend, attack, gather, and so on. The server would charge a fee on submission which is used to spawn resources in-game. Effectively, a time-simulated game without the need for full server availability.

Benefits

Reduced interaction – Humans don’t have to interact in real time. All strategies are pre-emptively committed to, possibly with markets for different time frames like daily, weekly, etc.

Deterministic – A purely deterministic simulation might not be as interesting, but the server can still inject randomness, e.g., random areas for resources to spawn.

Lower cost – The entire system can be down after strategies are committed to, which reduces costs. Users don’t have to watch charts all day either.

Fairness – Market blackouts make participation fairer. Front-running is strongly mitigated by avoiding order books and keeping strategies private.

Editing – One twist would allow strategies to be edited until execution is ready.

Implication

The essential property of deterministic algorithms is not temporal execution, but mapping from initial conditions to outcomes given the passage of time as a parameter. This opens the door to systems that operate in offline, batch, or delayed modes without losing fidelity, fairness, or strategic complexity. In essence, algorithms can exist conceptually before they are ever executed, and their “running” is a matter of computation convenience, not logical necessity.