AGI Scorecard › Invest › Track record · 中文 · English
📊 Public track record
Every explicit judgment is logged in the open, hits and misses shown side by side (misses carry the lesson). Currently 8 scored, 5 hits, hit rate 63%.
Aug 8, 2026 ·
⏳ PendingOptical-transceiver top forming in Jul–Aug 2026: all six top-zone checklist items lit (record volume, retail slogans, sold-out narrative, Innolight's mega HK IPO — the same pattern that marked the memory and space tops). No 10x from here; optics' real window is its own next cycle bottom.
Aug 8, 2026 ·
⏳ PendingMemory's next 10x window projected for late 2027: pre-registered shopping list (SNDK / MU / SKHY) executes only when three triggers fire — contract prices turning negative, capacity-cut announcements, PS≈1x with price stabilization. Leveraged ETFs banned.
Aug 7, 2026 ·
⏳ PendingSector-wide scan: spatial computing / AI glasses is the second open window (one notch weaker than robotics); low-altitude economy a half-window; solid-state batteries only via second-tier equipment names; seven other tracks closed.
Aug 7, 2026 ·
⏳ PendingProject-10x stock ranking: tier 1 = Wuzhou Xinchun (screws+bearings, lowest PS, un-named) and Keli Sensing (six-axis force); tier 2 = Moons' (Unitree's sole coreless-motor supplier) and Orbbec; equipment layer (grinders = 46% of a screw line's capex) added outside the pool.
Aug 7, 2026 ·
⏳ PendingSpace theme topped in May–June: SPCX (history's largest IPO, $1.75T) broke issue and sits ~45% off highs — same structure as SKHY at the memory top. Chasing here = paying up in the consensus phase. Falsified if SPCX reclaims $135 weekly AND makes new highs.
Aug 7, 2026 ·
⏳ PendingRobotics is the only track in early-mid cycle; any 10x structure lives in low-cap bottleneck-component pure plays, not humanoid leaders. Unitree lists ~Aug 19: no chasing for the first month, by rule.
Jul 5, 2026 ·
✅ HitStorage sector in the late stage of its peak zone: risk outweighs opportunity — distribute.
Jul 5, 2026 ·
❌ MissIPO-week base case (40%): peers spike before listing, distribute into the spike. Missed — the crash had already happened; I cited stale June-peak data and mis-weighted the scenario.
Jul 5, 2026 ·
✅ HitCold-reception scenario (25%): pricing discount widens, break-issue (below IPO price) risk rises.
Jul 5, 2026 ·
✅ HitThe 2x-leveraged 7709 suits only short event windows — never hold it long; choppy tape means negative-compounding double drag.
Jul 6, 2026 ·
✅ HitWeekly check: 0 of 3 liquidation triggers fired; hold the core, de-risk around the event this week.
Jul 7, 2026 ·
❌ MissQ3 base case (55%): bottoming in the -20% to -35% band; if Micron's late-September earnings don't cut HBM guidance, a recovery rally follows.
Jul 13, 2026 ·
❌ MissWeekly-check aside: 'trend is up this week, lower-high not confirmed.' Missed — SNDK fell 12.6% that same day and the downtrend was confirmed by 7/17. Lesson: in a discipline week, execute the plan, don't editorialize on direction.
Jul 20, 2026 ·
✅ HitStorage downtrend confirmed: lower-lows in place (SNDK $1,354 / MU $849); SKHY break-issue below $149 raised to the base case.
Jul 20, 2026 ·
⏳ PendingWeekly check #3: 0 of 3 liquidation triggers; hold the 1/3 core unchanged; buy-back clause suspended — a -37.8% drawdown is a trend break, not a healthy dip.
Get specific entry/exit levels & real-time trigger alerts
Research record, not investment advice. Full rationale and price levels are in the Chinese archive and Pro content.