Tag: robson
RBX Journal articles filed under this tag.
- August 8, 2026
Too much rigor is also a bug
This morning we published the story of the reviewer that rejected its own text. In the afternoon came the next chapter: an entire data exporter built in one day, proven byte by byte against reference vectors, with more than a hundred green tests. And the most valuable finding of the day came from a second reviewer, a twin of the implementer, which found the defect no test in the suite could catch: the code was too strict for the real world.
- August 8, 2026
The reviewer who rejected itself
Two days ago we told the story of a data plan that shrank until only the essential remained. What came next was more interesting: building what survived under continuous adversarial review, with two models in opposing roles. A contract went through three verdicts, a wrongly numbered migration nearly broke the live database, an SDK was hiding the winner of a write race, and in the middle of it all the reviewer found a defect in text it had itself demanded. A story about how certainty is bought by the pass, and about knowing when to stop buying.
- August 6, 2026
The plan that shrank
We wrote a data platform plan for Robson: sixteen metrics, eight analytical tables, a warehouse, dashboards. We handed it to a second model with a one-line instruction: protect production. The verdict was NO-GO, and every cut it proposed survived verification against the code. What remained fits in five invariants, a two-table contract, and a rehearsal with numbers. The story of how a good plan gets better by shrinking.
- August 5, 2026
The first span
Yesterday we told the decision: one ruler for all of risk. Today it debuted in production, on a real position. This post checks the geometry number by number, walks the two seconds of events between the signal and the protective stop, and explains why the position size went down when the span went up.
- August 4, 2026
One ruler
The operator looked at a screen and asked a one-line question: why is the stop three times farther than the first target? The answer exposed two rulers living in the same system, an environment variable that turned one binary into two products, and ended in the largest engineering effort on Robson since the Rust rewrite. Including the day the fix was nearly worse than the defect.
- August 4, 2026
The missing slot
A month with zero governed losses, and the dashboard showed 3 slots out of 4. The number was right and the rule was wrong. How one operator question became the formal supersession of an ADR, why the obvious fix was unsafe, and what a second model found in our own text before approval.
- August 1, 2026
An audit trail is not a telemetry sink
Our trading system database weighed 17 GB. The genuine audit trail fit in under 1 MB. How 96% of an event log became noise, how we cleaned it without losing a single line of evidence, and the rule we adopted so it never happens again.
- July 12, 2026
Volatile logs are not evidence
Recovery in critical systems must emit an auditable event, metric, or state. Logs help diagnostics, but they are not enough as proof.
- July 7, 2026
Engineering update: the brief a stranger could run
A reconciliation fix led to a second bug nobody asked to find, and the next chapter shipped as a brief written for an agent with zero memory of the day. It ran anyway.
- July 6, 2026
Engineering update: the fallback debuts in production
Less than 24 hours after merge, the market-data fallback protected a real entry. And the root-cause hunt ended somewhere unexpected. The full update.
- July 5, 2026
The stop gain that audited the system
A profitable stop that fired during a deploy exposed two invisible failure modes and changed Robson risk policy. The story and the decisions are here.
- July 5, 2026
Saved risk becomes an extra operation
Robson guaranteed 4 operations per month. Now each entry is charged its actual risk, and leftover budget becomes another chance to trade. Here is the math.
- July 4, 2026
The retest that did not stop us: live-fire validation of the stop buffer
Two shorts on the same symbol, hours apart. One exited above the recent high exactly as designed. The other watched price tag 63,450 — past the level that used to be the stop — and stayed alive. Between them we found a number that lied and a risk rule nobody had ever adopted.
- July 4, 2026
One entry, four bugs: a night of production debugging on Robson
We shipped a new protective layer for the stop engine and used live entries to validate it. The first order was rejected by our own risk gate, by a margin of 0.0000000000000000000002. Three more bugs were hiding behind that one.
- July 3, 2026
Stops that survive failures: redesigning the execution layer of Robson
A trade closed 48% above the risk cap while the daemon was down. The investigation produced two ADRs, five production fixes, and a Binance API surprise. Here are the numbers.
- April 22, 2026
Robson executes its first real-capital trade
After a formal two-phase validation process on testnet, Robson opened its first live BTCUSDT perpetual futures position with real capital. This post covers the migration, the architecture, and what the validation process actually confirmed.
- April 16, 2026
How we validate Robson before enabling real capital
Before any real capital is enabled, Robson runs a formal 5-phase E2E validation on testnet. Here is what that process looks like, what we found, and what the architecture learned.
- April 5, 2026
Robson v2 finds its missing execution glue
Over the last three days, Robson v2 gained the governed layer it was missing between risk, approval, audit, projections, and recovery. That is laying real groundwork for v3.