← Coil home
LIVE RECORD

The live engine record

What the engine actually did with real money, next to SPY and QQQ, published every session — win or lose. This is the live account. It is not the research backtest.

Every figure on this page is read from /api/perf and /api/board/proof when the page loads. None of it is typed by hand.

The record

Three returns over one window, from one start date. The engine line is the account it trades; both benchmarks cover the same window on the same basis. Δ is the engine minus that benchmark, in percentage points — negative means the engine is behind.

COIL ENGINE
LIVE
From —
Sessions
SPY
BENCHMARK
S&P 500 · same window
Engine Δ
QQQ
BENCHMARK
Nasdaq-100 · same window
Engine Δ

Reading the record from /api/perf.

This is the live record. It is not the backtest.

Two different kinds of number live on this site, and reading them as one would be the worst mistake you could make here.

The figures above are the live engine: real orders in a real brokerage account, funding-adjusted, from a single start date forward. The sample is thin — the session count on the card says how thin — and a thin sample is not evidence of an edge in either direction. It is a receipt, not a result.

The other number is a simulation. The decade-replay figures elsewhere on this site — the 2017–26 research backtest of the ranking backbone, on point-in-time index membership with modeled fills — are a different thing entirely. Different period, different capital, no orders. They validate the ranking brain; they are not this account and not a forecast. The replay ships with its own honesty rider in how Coil works, and if you want the questions to ask of any backtest, including ours, start with how to read a backtest before you trust it.

So: when the engine trails its benchmark, this page says so at the same size as anything favourable. The summary above is written by the engine, not for this page, and it is rendered exactly as served.

What the number includes, and what it leaves out

Both paragraphs below are fields on the API. They travel with the number wherever it is served, so the frame cannot be dropped on the way to a marketing page.

Proof of prior

A record is only worth reading if it could not have been edited afterwards. So every market morning, the exact bytes of the board the engine published are hashed with SHA-256, and that digest is appended to a public log with the date and the number of names scored that day.

The point is timing. A hash committed before the market resolved cannot be back-fitted to the outcome. If a past call were quietly rewritten to look better, its bytes would no longer reproduce the digest already sitting in the log — and the log is append-only, so the old entry does not go away.

Reading the commitment log from /api/board/proof.

How you would check one

Here is the honest version of that instruction, cost included.

Two things to know before you try it. The commitment log is free, unlimited, and needs no account or key. The archived board it hashes is a paid read — a small per-read fee in USDC, priced on the agents page — so reproducing a digest end to end is cheap but not free. We would rather say that than imply a check nobody can run.

For machines

Both records are public JSON. No key, no account, no email.

  • /api/perf — the record above, including every field this page renders.
  • /api/board/proof — the full commitment log, the canonical-JSON rules, and reference implementations in Python and JavaScript.

If you are wiring an agent to the board itself rather than to the record, the endpoints and the free MCP server are on the agents page.

Past performance does not predict future results. Percentages describe realized structure over a thin sample. Research publication, not investment advice.

Read the board, or run the engine

The Scanner publishes the same board this record is built on, every market morning ($12/mo, free preview). The Trading Bot is the engine itself — $29 one-time, and it ships disarmed.

See the Scanner Get the Trading Bot

Coil is software you install and run yourself, with your own brokerage credentials and capital. It is long-only and not investment advice, not a managed account, and not a signal service. Leveraged ETFs, where the engine uses them, can lose value rapidly, including total loss. The figures on this page are the live engine record on a real brokerage account, rendered as served; every other performance figure on this site is a research backtest — point-in-time and survivorship-free, not live or client returns. Past performance does not predict future results.