Research and trading, on the record.
Research collection and historical live trades, reported separately. These results do not establish a profitable strategy.
Read-only queries against the database opened by the running DigitalOcean bot. These are dated audit results, not live counters. A recent feed refresh does not mean a recent trade.
What Sayf is doing now
The platform provides planning, journaling, risk tools, and Razayel, an AI learning coach. The bot is collecting market and strategy-research observations. Sayf for Developers is an early-access offering in development; paid API plans and datasets are not available yet. Sayf Banking remains a concept.
Research collection at the audit
Each count below is a number of stored rows. Repeated observations are not independent experiments. A simulated quote is not a live order, and a funding observation is not a funding payment or trading return.
| Dataset | Rows | Latest observation |
|---|---|---|
| Maker simulation quotes | 164,373 | Oct 15, 2026 |
| Crypto funding observations | 61,950 | Oct 15, 2026 |
| Cross-venue observations | 30,253 | Oct 15, 2026 |
| Crypto price snapshots | 69,743 | Sep 26, 2026 |
Collection freshness differs by dataset. The crypto price snapshots were not current at the audit, even though funding collection continued. These counts show collection activity, not validated predictive edge.
Historical live trading results
The running bot’s trade table contains 102 live Kalshi trades, all settled, in strategy epoch v3.0: 37 wins and 65 losses. The latest recorded live trade was August 7, 2026 at 3:12 p.m. Eastern. This is historical activity, not an ongoing live trading track record.
Two of the 102 live trades did not have a valid pre-event CLV value and are excluded from this mean. The 37 wins and 65 losses describe all 102 settled live trades; they are not the win/loss counts of the 100-trade scored subset.
CLV compares the entry price with a pre-event closing benchmark, signed toward the position. This is an execution measure, not realized profit. We are not presenting a P&L or return figure from these aggregates.
Correction to the earlier public score
The earlier B− / −0.83¢ / 1,255 trades headline combined live trades, paper trades, and a separate experimental epoch. Its latest included trade was August 17, 2026. It should not have been described as the founder’s current live trading score.
The exporter also stamped the aggregate with the time it was generated. That timestamp showed the feed was running; it did not show when the underlying trades occurred. The mixed headline has been removed from the Company and Record pages. This dated audit replaces it.
Breakdown of the previously combined 1,255 trades
| Mode and epoch | Scored trades |
|---|---|
| LIVE / v3.0 | 100 |
| PAPER / v3.0 | 1,151 |
| PAPER / v3.0-inv | 4 |
Each subset had settled outcomes, a non-null CLV, and a pre-event method label. They belong to different modes or experiments and should remain separate.
Earlier measurement error
In August, reported 87% and 90.5% “edge” figures were retracted after we found that some closing snapshots had been captured after market resolution. That leaked outcomes into the measurement. The audit identified 527 observations labeled with the retired leaked_v1 method. They remain quarantined rather than deleted.
Historical results using that retired method are excluded from the pre-event CLV figures above. The research archive preserves the explanation of the original failure.
What the evidence supports
Sayf has trading tools, a historical live trading record, and ongoing research collection. This record does not demonstrate a validated positive predictive edge.
Future performance updates need a defined strategy, actual trade dates, separate live and paper results, a documented methodology, and results after relevant costs. More stored rows alone will not establish profitability.
Questions about the record or the data?
Ask about the methodology, a correction, or a developer research use case.
Contact SayfUpdate policy. This page changes when a new database audit is verified. It does not turn exporter timestamps into new trading activity.
Past performance does not establish future results. This page is not investment advice or an offer of a fund, pool, or managed account.