← Home

Market Lab Methodology


Purpose


Market Lab publishes aggregated, anonymized research on early Solana launch market behavior. It helps founders understand how strategy ideas perform under naive (paper), market-aware (shadow), and stricter (realistic) assumptions.


Data Pipeline


1. Internal benchmark runs are stored in a private research database (pumpfun-bot).

2. A public export script produces `public_benchmark_summary.json`.

3. The web app reads this file via `/api/benchmark/summary`.


Private database credentials and raw trade rows are never exposed to the web.


Metric Labels


| Internal | Public label |

|----------|--------------|

| C_net | Net Result |

| C_fills | Number of Trades |

| market_count | Markets Tested |

| C_avg_pnl | Average Per Trade |

| pos_mkt% | Consistency |

| top5 concentration | Dependency Risk |

| net_ex_top5 | Robust Result |

| healthy_pct | Data Quality |


Test Modes


  • **Paper** — Ideal test without full execution friction.
  • **Shadow** — Market-aware test with timing and liquidity constraints modeled.
  • **Realistic** — Stricter test with slippage, fees, and failed exit modeling.

    • Status Definitions


  • **Rejected** — Failed or depends on unreliable conditions.
  • **Needs More Data** — Sample too small to trust.
  • **Candidate** — Shows promise under current assumptions.
  • **Research-Ready** — Deserves deeper testing, not live deployment.

    • Simulation Engine


    The parameter playground on `/strategy-lab` applies adjustments to published family baselines. Full custom simulation engine integration with live backend runs is pending — see internal runbook for benchmark pipeline updates.


    Disclaimer


    Market Lab is for research and education. It does not provide financial advice, guaranteed profit, or automated live trading execution.