Signal performance, measured honestly

Hypothetical and backward looking. No guarantee of future results. Not investment advice. This page shows what following the top signals would have returned against the S&P 500, under rules that were fixed in advance and never change automatically.

The rules, fixed in advance

  1. Selection: every signal above the fixed score threshold in the measurement period. No selection after the fact, no threshold changes after seeing results.
  2. Entry: Closing price of the first trading day after publication of the signal (publication = filing_date for congress, filed date for 13f consensus).
  3. Horizons: 30, 90, 180 days.
  4. Benchmark: S&P 500 total return over the same periods (^SP500TR).
  5. Integrity: these rules are frozen in config/performance_rules.json since 2026-07-05 and are excluded from every automatic update path.

Data history in this measurement: 2026-06-03 to 2026-08-20, updated 2026-08-20. The history is shorter than 6 months, so treat every number here as an early, partial track record.

Every day the full published signal set is hashed into an append-only chain and anchored in two independent public places. See how to verify the track record yourself.

The underlying report data reaches further back than this measured track record: the 13F consensus draws on quarterly filing history back to 2024. The measurement above starts on the date the rules were frozen, so the track record grows forward from there and is never backfilled.

Congress Stock Trades

Selection: Every House purchase with a ticker and an impact score of 70 or higher, from the filing date onward. Signals measured: 20.

HorizonSignals avgS&P 500 TRn
30 days-0.8%+1.3%8
90 daysno signals old enough yet
180 daysno signals old enough yet

Smart-Money 13F Consensus

Selection: Every consensus pick with a conviction score of 50 or higher in the published consensus table. Signals measured: 6.

HorizonSignals avgS&P 500 TRn
30 days+6.9%+2.7%5
90 daysno signals old enough yet
180 daysno signals old enough yet

Startup Capital Raises (Form D)

Form D issuers are private companies without a public share price. No price backtest is possible. The block shows pipeline facts only and says so.

Current pipeline: 60 qualifying raises in the latest edition, $2.3B reported as raised. Because these issuers have no public share price, this block shows pipeline facts, not returns.

Track record: signals that came due

Per individual archived signal (not a report average): 10 signal/horizon combinations have reached their measurement horizon so far. 5 of 10 showed a positive return, 5 beat the frozen benchmark over the same horizon. This list only grows as more horizons elapse; nothing here is projected ahead of time or removed after the fact.

TickerReportSignal date HorizonSignal returnBenchmarkResult
AMZN13f2026-07-0530d+13.6%+2.7%hit · beat benchmark
GOOGL13f2026-07-0530d+3.0%+2.7%hit · beat benchmark
MSFT13f2026-07-0530d+27.4%+2.7%hit · beat benchmark
NVDA13f2026-07-0530d+8.4%+2.7%hit · beat benchmark
SNDK13f2026-07-0530d-18.2%+2.7%miss
INTCcongress2026-06-2330d-23.9%+0.8%miss
UBERcongress2026-06-2330d-6.7%+0.8%miss
MSFTcongress2026-06-0830d-5.0%+1.4%miss
FMAOcongress2026-06-0530d+9.6%+1.9%hit · beat benchmark
MSFTcongress2026-06-0330d-9.7%-0.5%miss

Definition and measurement code: scripts/trackrecord.py (same entry rule, horizons and benchmark as config/performance_rules.json, reused not duplicated).

Hypothetical, backward looking, no guarantee of future returns, not investment advice. Measurement code: scripts/performance_build.py. Rules: config/performance_rules.json.

DataSignals Lab publishes data and research. Informational only, not investment advice, not an offer or solicitation. Every figure links back to the official filing so you can verify it yourself.