The Conviction
A descriptive basket of 28 tickers that 30 tracked superinvestors most aggressively accumulate — drawn from a universe of 29 positive-ConvictionScore positions across 9 quarters of 13F data (time-decayed, 6 signal layers), sector-capped at 8 per sector, and conviction-weighted so each slot scales with the intensity of cross-manager accumulation.
This is pure data-display, not investment advice. Not an “optimal” or “ultimate” portfolio. Not a buy recommendation. HoldLens has no licensed financial-advisor credentials. The Conviction reflects aggregate institutional positioning per the public-record 13F snapshot — already 45 days lagged per SEC rules. Rebalances quarterly when new filings land, never daily or hourly. Read the methodology + how this is composed below.
How this is composed
Intentionally simple: conviction-weighted + sector-count cap. No active optimization, no factor tilts, no leverage, no derivatives, no shorts, no “alpha generation” claims. The discipline is the diversification cap; sizing reflects intensity of cross-manager accumulation.
Sector breakdown
Sector-count cap (max 8 positions per sector ≈ 27% by count) forces baseline diversification. Bars below show aggregate weight per sector under the conviction-weighted sizing — actual share varies by the ConvictionScore intensity of each sector's positions.
All 28 positions
Ordered by ConvictionScore. Weight scales with score — top position 10.70%, bottom 0.26%. Click ticker for the per-stock dossier.
Buyer CAGR column shows the average 10-year historical compound annual return of the public-record managers who built each position (sources: Berkshire annual reports, PSH NAV, public mutual funds, partner letters). Backward-looking public data only — not a forecast, not a recommendation.
| # | Ticker | Company | Score | Buyers | Weight |
|---|---|---|---|---|---|
| 1 | GE | GE Aerospace | +41 | 9 | 10.70% |
| 2 | FICO | Fair Isaac | +31 | 10 | 8.09% |
| 3 | BABA | Alibaba | +26 | 22 | 6.79% |
| 4 | SCHW | Charles Schwab | +25 | 16 | 6.53% |
| 5 | COHR | Coherent | +23 | 7 | 6.01% |
| 6 | MU | Micron Technology | +22 | 13 | 5.74% |
| 7 | OXY | Occidental Petroleum | +21 | 15 | 5.48% |
| 8 | KHC | Kraft Heinz | +21 | 5 | 5.48% |
| 9 | UBER | Uber | +17 | 24 | 4.44% |
| 10 | CB | Chubb | +16 | 14 | 4.18% |
| 11 | HEI | HEICO Corp | +16 | 9 | 4.18% |
| 12 | PLTR | PLTR | +14 | 9 | 3.66% |
| 13 | TRMD | Torm plc | +14 | 7 | 3.66% |
| 14 | GRBK | Green Brick Partners | +14 | 11 | 3.66% |
| 15 | MCO | Moody's Corp | +14 | 9 | 3.66% |
| 16 | BN | Brookfield Corp | +11 | 36 | 2.87% |
| 17 | MSCI | MSCI Inc | +11 | 10 | 2.87% |
| 18 | NOW | ServiceNow | +8 | 11 | 2.09% |
| 19 | NVDA | NVIDIA | +7 | 34 | 1.83% |
| 20 | JD | JD.com | +7 | 10 | 1.83% |
| 21 | DPZ | DPZ | +6 | 10 | 1.57% |
| 22 | NOVO | Novo Nordisk | +6 | 1 | 1.57% |
| 23 | GOLD | GOLD | +3 | 2 | 0.78% |
| 24 | META | Meta Platforms | +3 | 65 | 0.78% |
| 25 | ARM | ARM | +3 | 3 | 0.78% |
| 26 | VST | Vistra Corp | +1 | 26 | 0.26% |
| 27 | INTC | INTC | +1 | 12 | 0.26% |
| 28 | CVI | CVR Energy | +1 | 5 | 0.26% |
Why these specific rules
Why positive ConvictionScore (no arbitrary floors)
ConvictionScore already aggregates 9 time-decayed quarters of 13F data across 6 signal layers — smart-money consensus, contrarian flow, insider cross-check, trend streak, sector context, and prior-score persistence. That composite is the gate. Layering an extra “≥+20 score” or “≥2 buyers” floor on top would double-count what the score already encodes and discard real signal. The universe is every ticker with score ≥ +1 (net-accumulation positive). Currently 29 candidates pass before the sector cap.
Why 8-per-sector count cap
Without a cap, the composite would historically pile into whichever sector the smartest investors are most concentrated in (often technology). Capping by count (≤8 positions per sector ≈ 27% by count) forces baseline diversification independent of how heavy any individual position is. Count-based is simpler and more transparent than a weight-based cap — you can see at a glance how many positions a sector holds; the weight then varies by ConvictionScore intensity.
Why conviction-weighted (not equal-weight)
Methodology v1 used flat equal-weight (1/30 ≈ 3.33% each). v2 (live since 2026-05-19) uses ALL available signal: a position's slot scales with its ConvictionScore (slot = score ÷ sum-of-selected-scores × 100). Higher conviction = larger weight. The score already encodes the multi-quarter, multi-signal evidence — equal-weight would discard that information. Top position currently 10.70%, bottom 0.26%. Sizing reflects intensity of cross-manager accumulation, not a uniform rule.
Why quarterly (not daily/hourly) rebalance
13F filings only emit every 90 days (SEC filing deadline = 45 days after quarter end). Daily/hourly rebalancing would be theater — no new data is available between quarters. Real-time rebalancing also requires registered investment-advisor (RIA) credentials, which HoldLens does not hold. The cadence is determined by the SEC's 13F rules, not by us.
What this composite is NOT
Not an “optimal portfolio.” Not a recommendation to buy or sell any security. Not a backtested-and-validated index. Not a registered investment product. Not real-time. ConvictionScore correlation with forward returns is r = −0.12 across 221 ticker-quarter pairs — no predictive signal. The composite is a positioning tracker: what tracked managers already did, aggregated and described. Anyone considering acting on it should consult a licensed financial advisor first.
Related
- → ConvictionScore rankings — every tracked stock
- → Most-bought + most-sold this quarter
- → Big bets — top 100 by conviction × size
- → Consensus — widely held + net buying
- → Concentration — managers ranked by top-N weight
- → Q1 2026 13F signal summary
- → Snapshot JSON (LLM/dev)
- → How ConvictionScore is computed
Not investment advice. The Conviction is a descriptive aggregate of public SEC 13F filings — what tracked managers already did, not what they will do, not what you should do. Position sizing is conviction-weighted (slot ∝ ConvictionScore); sector-count cap is the only risk discipline applied. 45-day filing lag applies to all underlying data. HoldLens holds no licensed-advisor credentials. Consult a licensed financial advisor before acting on any of this. See methodology + disclaimer.