Skip to main content

Metrics Methodology

Investors' technical reviewers discount self-reported performance numbers by default -- correctly, since most trading-platform marketing sites do not disclose methodology. This page states how ChimeraMiND's numbers are produced so they can be checked rather than taken on faith.

Live status and uptime figures​

chimeramind.com/status and the platform's uptime figures are pulled at render time from /monitoring/health and /monitoring/overview, which read from the same Redis-backed health state the backend uses internally -- these are not static marketing copy, they are live reads and will show STBY/Unavailable rather than a fabricated number if the backend is unreachable (see Architecture Overview).

Capability counts (Service Catalog)​

Every count in the Service Catalog -- 60 MCP tools, 67 REST route files, 15 strategy modules, 5 ML models, 7 exchange gateways, 14 analytics streams -- is a direct count of code artifacts (registered tools, files, classes), reproducible by anyone with repository access via the same grep/glob commands used to produce them. These are structural counts, not performance claims. Endpoint-handler counts (not file counts) vary by methodology -- decorator grep, OpenAPI schema, and mounted-route enumeration each return a different number in the 450-580 range depending on what they include (see Architecture Overview) -- so this page cites the file count, which is unambiguous, rather than a single endpoint number presented as more precise than it is.

Trading performance numbers​

Status: not currently published on the public site. No Sharpe ratio, win-rate, or similar backtested/live performance figure is shown on chimeramind.com today -- deliberately: an unverifiable performance number is worse than no number. Any performance figures published in the future will state the measurement window, whether it is backtested or live/paper, slippage and fee assumptions, and survivorship handling -- consistent with how a quant reviewer would expect a Sharpe ratio to be sourced.

Early traction​

As of this writing, 26 users have registered (Supabase auth.users, excluding internal load-test/monitoring accounts and the founder's own accounts). All onboarding is currently paper-mode; no user has submitted a testimonial or case-study review yet, so none is published -- consistent with the no-fabrication policy on this page. This number will be updated as it changes rather than left stale.

What "verified" will mean when performance numbers return​

The platform's own paper-trading account (_system, isolated from any real tenant, see Architecture Overview) is the intended source for any republished performance claim, since it runs the identical decision/risk pipeline as live accounts (see Exchange Gateways) without requiring disclosure of any individual user's real trading results. A future independent-verification mechanism (e.g., a read-only exchange API key held by a third party) is a reasonable diligence ask and is not yet built.