资本不是弹药；首先要保住舰体、选择射界，然后才是开火。
FY25/26 · capital flight recorder · after-action report
I made capital move fast.
The drawdown taught me
what speed costs.
A private FY2025–26 trading record retaining only model bounds, cycle annualization, holding duration and risk methodology; position outcomes and underlying financial data stay private.
open FY25/26 flight record
bearing · singular throne
One motif, two compositions.
Compare the existing Portfolio backdrop with the home cover. These are artistic views, not astronomical observations or financial data. Both complete images remain available below; enable the overlay to compare their composition.
Every return is a jump through the dark.
Annual return · upper bound
41.4%
modeled · capital-time reconstruction
Sharpe · upper bound
0.85
estimated · risk-free input 4.50%
Max drawdown · upper bound
−22%
estimated · 2026-06-16—25 window
Beta · upper bound
1.85
estimated · technology concentration vs SPX
02 · FY25/26 capital black box
Return is not one number.
It is a chain of assumptions.
Three telemetry layers answer three useful questions: closed-cycle capital velocity, thrust relative to supplied benchmarks, and the holding-time geometry of five completed routes.
Annualized cycle efficiency
+261.2%
CAPITAL VELOCITY
- Weighted capital days
- 7.8
- DAYS / CLOSED CYCLES
- Flight paths
- 05
- PROFITABLE / CLOSED
- Annual volatility
- 45%
- ESTIMATED UPPER BOUND
- Reporting basis
- CLOSED
- LOT-LEVEL WINDOW
How fast did capital turn through the five disclosed closed cycles? 261.2% is a mechanical 365-day efficiency projection, not an achieved account return. Unit: %; capital occupation: days. FY2025–26 · method dated 2026-08-08 · source: disclosed model summary; risk measures are estimates.
VECTOR 01MODEL MAX / QQQ
05×
2.94×
VECTOR 02MODEL MAX / SPX
05×
4.40×
How large is the model maximum relative to each supplied benchmark? QQQ ○ 2.94×; SPX □ 4.40×. Unit: ratio, zero baseline. FY2025–26 · method dated 2026-08-08. Source: disclosed model summary; original benchmark values and exact comparison dates are not disclosed.
Five closed-cycle observations
Which cycles combine high annualized efficiency with short holding periods? Bars: modeled efficiency (%); circles: disclosed holding days. Shared zero-based scales, separate units. FY2025–26 · method dated 2026-08-08 · source: disclosed closed-cycle summary.
+260.4%annualized cycle efficiency0300%
17.5 days0250 days
Two entry clusters in May and June; two exits completed by 16 June.
+208.5%annualized cycle efficiency0300%
3.0 days0250 days
A three-day disclosed holding window; year-end open lots remain outside this cycle.
+257.9%annualized cycle efficiency0300%
11.2 days0250 days
Four rolling entries and four rolling exits from late January through 18 March.
+32.6%annualized cycle efficiency0300%
246.0 days0250 days
The longest disclosed holding window. Operational settlement detail remains private.
+85.6%annualized cycle efficiency0300%
5.6 days0250 days
A short disclosed holding window with several March exits; settlement values remain private.
03 · top 10 allocations · usa
十个 AI 引力源，
组成一座实时运行的产业太阳系。
2026-08-07 08:29 ET · 非农发布前快照：最高权重 NVDA 是太阳，其余九支映射为水星至冥王星。星体按当前历元公转并依据真实自转方向动态演化；点击任意代码，可展开对应产业天体。
How is the research allocation distributed? Each bar uses the same 0–100% scale; NVDA leads at 18%. Subjective research weights, not verified account holdings. As of 2026-08-07 08:29 ET · source: published allocation framework. Celestial sizes are illustrative.
- NVDANVIDIA · Sun · compute core18%
- AVGOBroadcom · Mercury · custom silicon & network15%
- AMDAMD · Venus · second compute platform13%
- ORCLOracle · Earth · AI cloud & database10%
- AMZNAmazon · Mars · cloud & custom silicon9%
- MSFTMicrosoft · Jupiter · enterprise cloud & agents9%
- TSMTSMC · Saturn · advanced manufacturing8%
- GOOGLAlphabet · Uranus · model distribution7%
- MUMicron · Neptune · high-bandwidth memory6%
- VRTVertiv · Pluto · power & liquid cooling5%
这是本人截至 2026-08-07 08:29 ET、美国 7 月非农发布前的主观研究与配置框架，不构成投资建议，也不承诺未来表现。宏观数据、财报、估值与风险条件会变化；研究你自己的航线，守住你自己的舱门。