Oberth Capital · an autonomous fund family, operating around the clock

Autonomous
capital.

A family of self-directing strategies that research, decide, and trade without a human at the desk — inside guardrails written as code, where no argument, however persuasive, can cross them.

Explore the strategies Read the discipline
2
Strategies
11
Autonomous agents
88
Codified risk checks
0
Manual orders
I

Two temperaments. One constitution.

Strategy 01 — Core

The Steady Book

Own what the news hasn’t finished pricing.

A catalyst-driven long book. Specialist scouts sweep filings, approvals, awards, and earnings around the clock; a strategist distills their work into a written thesis each morning; an executor trades only inside it. Fully funded by cash — leverage is not in its vocabulary.

  • MandateLong equities, cash-only
  • EdgeConfirmed catalysts, early
  • Sizing≤ 5% per name
  • BenchmarkBeat the index, or fold
Strategy 02 — Escape Velocity

Income at the Limit

Paid to enter. Paid to wait. Paid to exit.

The leveraged wheel, run relentlessly. Quality names held in round lots with calls written against every one; entries financed by fully cash-secured puts; premiums laddered so a third of the book renews each Friday. Margin to a hard 2× ceiling — never an inch past it.

  • MandateCovered options income
  • EdgePremium, compounded weekly
  • Exposure≤ 2.0× — enforced in code
  • BenchmarkAbsolute return
II

The part no one can override.

Every order — every single one — passes through a deterministic risk engine before it can reach a market. It is not another intelligence with better judgment. It is arithmetic. It cannot be persuaded, and nothing trades around it.

R-01
One gate. The engine is the only path to the broker; the strategies never touch credentials.
R-02
Nothing naked, by construction. Every call written is backed by owned shares; every put is secured by reserved cash — not buying power.
R-03
Hard ceilings. Position caps, order caps, and the 2× exposure limit are arithmetic checks, not policies.
R-04
Circuit breakers. A bad day stops new risk automatically; a deep drawdown halts everything until a human turns the key.
R-05
Journal before trade. Every decision — approved or refused, with its full written rationale — is recorded before any order exists. If it can’t be recorded, it doesn’t trade.
R-06
The kill switch. One file. Everything stops.
III

Routines, not resolutions.

The scouts sweepFour specialists — generalist, biotech, energy, contrarian — hunt confirmed catalysts three times every trading day, and file every pick as a falsifiable claim with a deadline.
The morning thesisThe steady book's strategist reads every overnight note, the macro tape, and last week's audit — then writes the day's plan: what may be held, at what size, and when to walk away.
The ladder is setEscape Velocity's navigator prices the option ladder, picks strikes worth being assigned at, and sets the day's leverage stance against the tape.
Execution heartbeatsHourly passes on the steady book: orders proposed with written rationale, priced off live markets, judged by the engine. Declining to trade is always a valid outcome.
The wheel turnsEscape Velocity covers every uncovered lot, harvests calls at target premium capture, sells entry puts for names it wants — and logs why, every time.
The weekly auditsIndependent risk officers grade each strategy's week in writing — coverage, leverage path, premium capture, benchmark honesty — with questions that must be answered.
The research reviewA researcher reads the week's new academic work on markets and files adoptable, falsifiable ideas — most are rejected, which is the point.

Every routine runs unattended · every output is written down · weekends are for research, not trading

IV

It gets better while it runs.

Most operations improve when someone remembers to improve them. This one is built as a loop: every agent's work becomes another agent's evidence, and the loop closes every single week — whether or not anyone is watching.

Act in writing

Every scout pick, thesis, and order carries its full rationale. Nothing happens as a hunch; everything happens as a claim.

Face the record

Picks are graded hit-or-miss against the market on their stated deadlines. A scoreboard, not a memory, decides which instincts deserve capital.

Answer for it

The weekly audits name what worked, what bled, and what was luck. Their findings aren't suggestions — the strategists must adopt or rebut each one, in writing.

Fold it back in

The next morning's thesis begins from the graded record of the last one. Good instincts compound; bad ones lose their sizing. Then the loop starts again.

V

An organization that writes itself.

The fund isn't a fixed program — it's a living organization defined as a versioned document. Its agents, their charters, their schedules, and their tools can all grow. One thing can't: the guardrails don't evolve from the inside.

E-01
Specialists on demand. A new sector scout — defense, semiconductors, anything — is commissioned in an afternoon and inherits the full constitution: the caps, the journal, the scoreboard.
E-02
Capabilities as small tools. New data sources arrive as compact, auditable instruments — regulatory filings, trial registries, federal awards — never as opaque dependencies.
E-03
Strategies earn their seat. The second strategy had to justify itself before it traded a dollar; the third will have to beat the scoreboard, not a pitch meeting. Expansion is gated on evidence, not enthusiasm.
E-04
Ideas are audited too. The research desk feeds the fund from published work — and its suggestions face the same fate as trades: adopted with reasons, or rejected with reasons.
E-05
The one immutable part. Agents can propose anything except changes to the risk engine. Any attempt to touch it is reverted automatically — evolution happens inside the guardrails, never to them.
No trade is a valid outcome.
— from the execution charter