Trading books, in the open
Each book is one strategy with its own capital, holdings and complete ledger. Results are net of trading costs, shown with their benchmark, and kept — losses, pauses and stopped books included.
This record shows 12 of 17 forward books Greenyear runs, selected by the operator. Selected books are published with their complete history; individual trades are never selected. Books scored week by week are on the Weekly page.
Live — actual capital
No live books. Trading actual capital has not been authorized.
Signal — hypothetical portfolios
Signal · Hypothetical portfolio Running
Fully invested in SPY; dividends reinvested. The benchmark for the agent cohort, started on the same terms.
Current holdings of SPY buy and hold (agent cohort)
| Holding | Weight | Return |
| SPY | 99.3% | −10.3% |
| Cash | 0.7% | — |
- Net return
- −10.2%
- Max drawdown
- −11.3%
- Track record
- 50 days
Since 1 Oct 2026 · valued 20 Nov 2026, 21:00 UTC
Signal · Hypothetical portfolio Running
Holds uninvested cash, earning nothing: the floor every agent should beat after costs.
Current holdings of Cash (agent cohort)
| Holding | Weight | Return |
| Cash | 100.0% | — |
- Net return
- 0.0%
- Max drawdown
- 0.0%
- Track record
- 50 days
Since 1 Oct 2026 · valued 20 Nov 2026, 21:00 UTC
Signal · Hypothetical portfolio Running
The agents' whole universe in equal weights, rebalanced monthly: what choosing nothing in particular earns.
Current holdings of Equal weight, 45 instruments
| Holding | Weight | Return |
| WMT | 2.8% | +23.2% |
| PG | 2.7% | +17.7% |
| XLP | 2.5% | +12.5% |
| V | 2.5% | +12.1% |
| HD | 2.5% | +20.1% |
| Cash | 7.2% | — |
- Net return
- −4.1%
- Max drawdown
- −6.8%
- Track record
- 50 days
Since 1 Oct 2026 · valued 20 Nov 2026, 21:00 UTC
Signal · Hypothetical portfolio Running
Long-only, up to 10 positions of at most 20% each, chosen from 45 liquid US ETFs and large-cap stocks after every close by a language model running in Claude Code on the operator's desktop, from Greenyear's evidence packet and its own web research. Reviews it does not answer before the next open are missed.
Current holdings of Claude Code agent
| Holding | Weight | Return |
| XLV | 12.0% | −2.2% |
| SPY | 9.4% | −6.4% |
| XLF | 8.7% | 0.0% |
| LLY | 8.4% | +17.2% |
| JNJ | 6.8% | −5.0% |
| Cash | 35.0% | — |
- Net return
- −4.2%
- Max drawdown
- −5.7%
- Track record
- 50 days
Since 1 Oct 2026 · valued 20 Nov 2026, 21:00 UTC
Signal · Hypothetical portfolio Running
Long-only, up to 10 positions of at most 20% each, chosen from 45 liquid US ETFs and large-cap stocks after every close by a language model running in Codex on the operator's desktop, from Greenyear's evidence packet and its own web research. Reviews it does not answer before the next open are missed.
Current holdings of Codex agent
| Holding | Weight | Return |
| LLY | 5.5% | −3.7% |
| XLV | 5.3% | +2.7% |
| COST | 3.4% | −0.1% |
| MSFT | 2.5% | −4.1% |
| Cash | 83.3% | — |
- Net return
- −2.3%
- Max drawdown
- −3.4%
- Track record
- 50 days
Since 1 Oct 2026 · valued 20 Nov 2026, 21:00 UTC
Signal · Hypothetical portfolio Running
Long-only, up to 10 positions of at most 20% each, chosen from 45 liquid US ETFs and large-cap stocks after every close by xAI's Grok 4.7 (via OpenRouter, on the server), from Greenyear's evidence packet and native web search, within a $1 daily model budget.
Current holdings of Grok 4.7 agent
| Holding | Weight | Return |
| AAPL | 15.3% | +3.2% |
| Cash | 84.7% | — |
- Net return
- −1.2%
- Max drawdown
- −3.7%
- Track record
- 50 days
Since 1 Oct 2026 · valued 20 Nov 2026, 21:00 UTC
Paper — broker simulation
Broker paper trading has not launched.
Historical research
Backtests on stored history are kept separate from forward results. See research.