WSLI · paper only · Eastern time

Agents.

Loading snapshot…

The new $10,000 experiment—not the legacy NVDA account. No live money. A signal is not a fill, and paper results do not establish profitability.

Agent monitor

—

Supervisor status—not gateway connectivity.

Market

—

US equities session

Equity

—

Alpaca paper

Trial P/L

—

Compared with the $10,000 starting balance.

Today’s progress

Checking publication…

No current monitoring evidence loaded.

01 · Prepare

—

Only stocks with verified history enter the universe.

02 · Opening scan

—

The first five-minute range is checked after 9:36 a.m.

03 · Paper execution

—

Checking the verified session entry limit.

04 · Safety & close

—

Code manages protective orders and closing reconciliation.

Quote observations not yet loaded.

Schedule not yet loaded.

Fixed cash-sized limits. Nominal stop risk is not a guaranteed maximum loss.

Entry-policy change not yet verified.

No signal means no trade. A daily entry used is not a sleeping bot. If hawkspc stops publishing, timestamps go stale; the site will not invent new activity.

What was bought & why

Broker fills + recorded rules

Waiting for verified fills.

No decision evidence loaded.

The fixed opening-range strategy chooses entries; Gem reviews them afterward. This is an evidence-backed decision summary—not Gem’s hidden thoughts or a claim that the stock will rise. A rejected attempt still uses an entry slot.

Gem’s oversight

Two bounded reviews · not always-on AI

No review evidence loaded.

Gem reviews recorded evidence and can request a halt. Deterministic code sizes and sends orders. A saved report is not proof of correct reasoning.

Positions

Checking

Symbol	Qty	Entry	Last	Value	P/L
Waiting for verified paper-account positions.

Activity journal

Checking

Waiting for published execution evidence.

Orders & fills

Checking

Time	Side	Symbol	Filled / ordered	Price	Status
Waiting for verified order history. Canceled setup tests are excluded.

Is Gem profitable?

Not established

A positive paper balance would not, by itself, demonstrate a repeatable edge. This experiment is still validating execution. Software faults, forced exits and losses remain part of the reported account result.

The October 20 trial endpoint is an operational review, not a promise of profitability. A credible performance assessment needs a separate, pre-planned test of fixed rules, realistic costs, drawdown and uncertainty across independent sessions. No date or number of trades guarantees proof.

Gem reviews recorded evidence; the deterministic executor chooses eligible entries and submits paper orders. We evaluate the whole system, not an AI personality. Read the performance methodology and simulation limits.

Research priorities

Weekdays · 4:30 p.m. Eastern

Stop width versus noise. The current 0.1 × daily ATR stop also appears in the published ORB study; it is not automatically a coding error. Test alternatives on held-out observations with the same dollar risk. One quick stop-out is not enough evidence to change the strategy.
Feed and fill mismatch. Signals use one exchange (IEX); paper fills use NBBO. Check whether a breakout exists on comparable market data. Alpaca’s simulation rules · IEX versus SIP
Measure execution, not just win rate. Track stop slippage, holding time, rejected signals and net results after realistic costs. Paper simulation omits important live-trading frictions.
Do not inherit a paper’s claimed returns. The study tests a broad universe and both long and short trades. WSLI currently uses a much smaller verified universe, long-only trades and one position at a time. This is a different experiment; its performance must be measured independently.

Research proposes tests. It cannot place orders or automatically rewrite the strategy. Useful new findings are reported; repetitive findings stay quiet.