Understand the simulation engine.
A practical guide to using SimuFi for crypto scenario analysis, interpreting agent debates, and separating live market context from hypothetical events.
What SimuFi does
SimuFi takes a crypto or onchain scenario and evaluates how different market participants might react. The engine combines available live market context with a multi-agent debate so you can examine competing viewpoints instead of relying on a single model response.
How to write a strong scenario
Be specific about the asset, event, timing, and what you want to test. For example: “What if a whale sells 10% of a token's circulating supply into current liquidity?” Specific scenarios give the agents a clearer decision surface.
- Name the asset when possible.
- Describe the hypothetical event clearly.
- Include a timeframe when it matters.
- Ask about a measurable market reaction.
The agent room
SimuFi uses eight distinct perspectives: Whale, Trader, Researcher, Degen, Panic Seller, Influencer, RWA Investor, and Developer. They are designed to emphasize different incentives and risks rather than produce eight identical answers.
Three-round debate
Reading the verdict
The bullish, neutral, and bearish percentages are a simulated distribution of the agents' final assessment. They should be read as scenario probabilities, not as a promise that an event will happen or a price target.
Live data vs. hypothetical data
Market prices and related indicators are treated as observed context when available. A scenario such as “what if an influencer posts about this token?” remains hypothetical unless the user supplies evidence that the event actually happened. SimuFi should never be treated as a source of verified breaking news.
Market context
Depending on asset coverage and provider availability, SimuFi can use price, recent performance, volume, technical indicators, order-book information, funding and other market signals. Provider outages can affect what is available, so the interface distinguishes live context from unavailable data.
Limitations
- Simulations are model-generated and can be wrong.
- Hypothetical events may have outcomes that differ from the simulated debate.
- Market conditions can change after a simulation is run.
- Data-provider availability and asset coverage can vary.