Hesper Atlas

The signal ledger, in numbers.

Aggregate statistics of every closed BUY-to-SELL signal on the Hesper Atlas public track record: win rates by year and by theme, the full return distribution, holding periods, and the best and worst signals ever produced. Regenerated after every US close from the same ledger, losers included.

As of 2026-08-20 03:31, the Hesper Atlas public ledger contains 570 closed BUY-to-SELL signals across 155 tracked names since 2022-05-25. 63.0% ended in profit; the average closed signal returned +38.8% over 134 days. The average winner made +66% while the average loser cost −7.4%, and 8.6% of all closed signals at least doubled.

Source: every figure on this page is computed from the machine-readable JSON ledger (field stats_ext), regenerated after every US close, losing signals included. Educational tool, not financial advice.

570
closed signals since 2022
63.0%
ended in profit
+38.8%
average return per closed signal
+66% / −7.4%
avg win vs avg loss: the asymmetry is the edge
67
median days held
9.5%
of signals ran longer than a year

Results by year

Every closed signal, grouped by the year it exited. Trend engines earn their keep in trending years and get whipsawed in choppy ones; both show up here.

YearSignalsEnded in profitAvg returnMedian
20225024.0%−4.6%−4.5%
202313060.8%+10.2%+2.2%
202413067.7%+35%+10.2%
202515772.6%+52.6%+9.1%
202610364.1%+79.7%+4.7%

The current year contains only the signals that already closed; open positions join it when they exit.

What a closed signal looks like

The distribution behind the averages. A trend engine's math is asymmetry: it takes many small losses to stay in the game for the few very large wins.

8.6%
at least doubled (+100% or more)
14.6%
returned +50% or more
27.7%
returned +20% or more
63.0%
closed positive at all
9.6%
lost 10% or more
2.3%
lost 20% or more

Results by theme

Closed signals grouped by the theme of the underlying name, largest first.

ThemeSignalsEnded in profitAvg returnMedian
Genomics & AI-Health9657.3%+10.9%+2.6%
AI Software6360.3%+32.9%+3.3%
AI Power & Grid5856.9%+44.2%+3.3%
AI Chips5453.7%+29.4%+1.8%
Consumer & Internet5170.6%+14.3%+3%
Fintech & Crypto3860.5%+43.5%+9.3%
HBM & Packaging2770.4%+42.1%+9.3%
Quantum2454.2%+84.4%+16.2%
AI Servers & Systems1872.2%+78.8%+15.7%
Energy1668.8%+5.5%+3.8%
Industrials1560.0%+30.2%+2.6%
AI Cloud1566.7%+103%+9.4%

Ten best closed signals

SymbolEntryExitReturnDays
AXTI2025-08-282026-06-10+2820.9%286
PLTR2023-05-112026-01-28+1492.6%993
IREN2025-05-012025-11-05+1099.5%188
SMCI2022-08-122024-07-24+989.5%712
RGTI2024-09-132025-02-26+987.9%166
POWL2022-10-252025-02-18+809.8%847
LITE2025-06-252026-07-06+696.8%376
VRT2023-04-262025-01-27+656.3%642
NVDA2022-12-092025-03-14+616.3%826
RGTI2025-04-212025-10-17+471.9%179

Ten worst closed signals

SymbolEntryExitReturnDays
RGTI2024-07-152024-08-05−33.1%21
MDB2022-08-112022-09-08−32.9%28
TER2025-02-262025-04-25−32.8%58
TEM2025-06-182026-03-20−32%275
IREN2024-07-262024-08-07−29.1%12
IONQ2023-09-112023-09-27−27.6%16
NU2024-09-192025-02-28−27.4%162
TER2022-08-122022-10-24−25.7%73
AXTI2023-01-272023-02-24−22.9%28
ON2024-02-222024-04-22−22.6%60

Held-out evaluation, separate from this ledger

The ledger above replays today's rules over history, so it inherits hindsight (the caveats below spell this out). The held-out evaluation is the harder test: each engine is tuned on the earlier part of a name's history and scored on the last 40% it never trained on, after costs, with next-day fills. As of the July 2026 evaluation across all 152 evaluated names: median worst drawdown -30% with the engine versus -57% buying and holding; 150 of 152 names had a shallower worst drop; median return +22.7% per year versus +18.9% holding. For scale over the same January 2022 to July 2026 stretch: cash returned about +4.1% a year, the S&P 500 about +11.9% and the Nasdaq 100 about +14.2% (total return with dividends). Backtested, not live. Method on the methodology page.

Read this before quoting the numbers: the honest fine print.
Generated 2026-08-20 03:31 across 155 tracked names from end-of-day data · Source data: the machine-readable JSON ledger (field stats_ext) · Educational tool, not investment advice · Past performance (measured or backtested) does not predict future results.
Open Hesper Atlas →