Published 2026-10-09 · Data as of shown per asset
Daily index and equity scenarios
Public price acquisition · local analysis, selection and writing · no external inference service
Index and equity scenarios
Each report explains market conditions at the data date, upward, corrective and downward paths, and response conditions. The 5 companies per market are selected from internal market-cap records, not certified current whole-market rankings.
| Market | Asset | Data as of | Last 5 days | Last 20 days |
|---|---|---|---|---|
| Korean market · KR | KOSPI | 2026-10-08 | -3.10% | -5.28% |
| US market · US | S&P 500 | 2026-10-08 | +1.29% | +2.29% |
| Korean market · KR | Samsung Electronics | 2026-10-08 | -2.41% | -2.59% |
| Korean market · KR | SK hynix | 2026-10-08 | -5.44% | -5.44% |
| Korean market · KR | SK Square | 2026-10-08 | -6.45% | -5.87% |
| Korean market · KR | Samsung Electro-Mechanics | 2026-10-08 | +2.97% | +7.22% |
| Korean market · KR | LG Energy Solution | 2026-10-08 | +13.36% | +11.17% |
| US market · US | NVIDIA | 2026-10-08 | -0.16% | +5.55% |
| US market · US | Apple | 2026-10-08 | +3.06% | +4.24% |
| US market · US | Alphabet | 2026-10-08 | +2.97% | +4.72% |
| US market · US | Microsoft | 2026-10-08 | +1.91% | +6.13% |
| US market · US | Amazon | 2026-10-08 | +2.35% | +0.86% |
Previously validated scenarios · including B3
Retrospective evidence, separate from outcomes of the current outlook.
View periods, Top-2 and comparisons
| Model | Horizon | Top-2 | Reference | Period |
|---|---|---|---|---|
| S&P 500 · B3-1 | h5 | 50.81% | B2 44.90% | 2024-01-02 – 2026-09-14 |
| S&P 500 · B3-1 | h20 | 55.95% | B2 47.30% | 2024-01-02 – 2026-07-31 |
| KOSPI · B3-style | h5 | 48.91% | Separately trained | 2024-01-03 – 2026-08-28 |
| KOSPI · B3-style | h20 | 47.06% | Separately trained | 2024-01-03 – 2026-08-06 |
Top-2 is not a profit probability. Details include five-report outcome checks, trading results and rejected B3-2/B3-3 combinations.
Appendix · MiDM/B3 usage and internal/external processing
MiDM: Applied · B3: Shown per asset
| Stage | Location | Usage |
|---|---|---|
| Market data acquisition | External → internal | Public prices are downloaded and stored locally. Requests contain public symbols and date ranges only. |
| Numerical and scenario analysis | Internal | Indicators and historical analogues are calculated from stored prices. B3: Shown per asset. |
| Decision and selection | Internal | MiDM selects buy, wait, hold or sell candidates. |
| Text, charts and languages | Internal | Calculated values and selections populate templates, charts and tables. No external writing or translation service. |
| Report publication | Internal → external | Reviewed reports, aggregates and documentation are published to GitHub and Hugging Face. |
| Country-based language | Visitor → external | Automatic URLs query Country.is using the visitor IP. No financial inputs or report body are sent. Direct language URLs skip this lookup. |
Internal calculations, prompts and raw data are not sent to external inference services. Public report content is uploaded for publication. The performance below is historical KOSPI evaluation, not the success probability of this asset.
MiDM · historical accuracy 54.37% · balanced accuracy 49.24%