Model accuracy
Measured walk-forward on real history — we don’t cherry-pick winners. The weak numbers are shown as they are.
Market US
| Model | Price toward value in 12 mo. | n | Below floor: excess return, 12 mo. |
|---|---|---|---|
| Lynch Fair Value | 43.6% | 296 | -4.9% (n=58) |
| Sector P/E | 41.2% | 1,475 | -4.8% (n=496) |
| Buffett DCF | 40.4% | 374 | +1.0% (n=48) |
| Shiller CAPE | 38.9% | 496 | -7.0% (n=76) |
| Nash NCAV | 36.6% | 1,830 | +21.2% (n=53) |
| Greenblatt Magic | 35.6% | 1,378 | -1.1% (n=165) |
| Pabrai 10x | 35.5% | 1,517 | +0.4% (n=122) |
| Gordon DDM | 35.0% | 1,091 | -3.5% (n=9) |
| Justified P/E | 34.3% | 1,465 | +5.8% (n=77) |
| Graham Number | 34.1% | 1,379 | +3.6% (n=69) |
| Graham NCAV | 33.8% | 216 | +3.6% (n=2) |
| Schloss Net-Net | 33.3% | 216 | +3.6% (n=2) |
Market Thailand (SET)
| Model | Price toward value in 12 mo. | n | Below floor: excess return, 12 mo. |
|---|---|---|---|
| Schloss Net-Net | 61.7% | 2,599 | +17.0% (n=80) |
| Gordon DDM | 61.4% | 3,154 | +12.2% (n=158) |
| Graham NCAV | 58.9% | 2,599 | +6.0% (n=209) |
| Justified P/E | 55.6% | 4,903 | +6.7% (n=673) |
| Greenblatt Magic | 53.0% | 3,833 | +1.3% (n=802) |
| Buffett DCF | 52.6% | 3,755 | -4.0% (n=1,068) |
| Sector P/E | 52.5% | 3,834 | +2.3% (n=862) |
| Pabrai 10x | 51.9% | 4,041 | -2.9% (n=895) |
| Nash NCAV | 51.1% | 4,827 | +6.8% (n=1,396) |
| Graham Number | 49.2% | 3,832 | +2.6% (n=1,107) |
| Lynch Fair Value | 48.7% | 1,638 | -2.9% (n=668) |
| Shiller CAPE | 45.2% | 4,074 | +1.3% (n=2,268) |
Market US
| Model | Price toward value in 12 mo. | n | Below floor: excess return, 12 mo. |
|---|---|---|---|
| Lynch Fair Value | 62.5% | 8 | +32.0% (n=6) |
| Graham Number | 57.9% | 57 | +16.8% (n=22) |
| Sector P/E | 56.1% | 57 | +2.3% (n=36) |
| Greenblatt Magic | 56.1% | 57 | +25.9% (n=21) |
| Pabrai 10x | 51.7% | 58 | -1.7% (n=16) |
| Justified P/E | 43.9% | 66 | +23.2% (n=15) |
| Shiller CAPE | 42.1% | 19 | +24.7% (n=9) |
| Nash NCAV | 34.5% | 55 | -15.6% (n=4) |
| Graham NCAV | 33.3% | 15 | — |
| Schloss Net-Net | 33.3% | 15 | — |
| Gordon DDM | 25.0% | 32 | — |
| Buffett DCF | 9.1% | 11 | +8.2% (n=2) |
DAVI Score composite · US — v1 (equal vote across models) vs v2 (weighted by measured accuracy)
Measured walk-forward over the same window (2022→now): hit = the 12-month peak after entry ≥ +20% · the numbers show whether a higher score really means a higher hit rate — not a forecast.
v1 — raw votes
| 0 votes | 53.1% | n=1,485 |
| 1-2 votes | 58.7% | n=639 |
| 3-5 votes | 60.2% | n=231 |
| 6-8 votes | 76.9% | n=26 |
| 9+ votes (DAVI Score top) | 88.9% | n=18 |
v2 — accuracy-weighted (live)
| 0% weighted votes | 53.5% | n=1,731 |
| 1-25% weighted votes | 60.8% | n=507 |
| 26-50% weighted votes | 61.7% | n=115 |
| 51-75% weighted votes | 70.0% | n=20 |
| 76-100% weighted votes (top) | 69.6% | n=23 |
DAVI Score composite · TH — v1 (equal vote across models) vs v2 (weighted by measured accuracy)
Measured walk-forward over the same window (2022→now): hit = the 12-month peak after entry ≥ +20% · the numbers show whether a higher score really means a higher hit rate — not a forecast.
v1 — raw votes
| 0 votes | 29.4% | n=780 |
| 1-2 votes | 28.9% | n=1,078 |
| 3-5 votes | 38.1% | n=767 |
| 6-8 votes | 40.7% | n=472 |
| 9+ votes (DAVI Score top) | 48.1% | n=243 |
v2 — accuracy-weighted (live)
| 0% weighted votes | 27.5% | n=1,555 |
| 1-25% weighted votes | 35.6% | n=582 |
| 26-50% weighted votes | 38.9% | n=463 |
| 51-75% weighted votes | 41.6% | n=385 |
| 76-100% weighted votes (top) | 47.5% | n=326 |
DAVI Score composite · KR — v1 (equal vote across models) vs v2 (weighted by measured accuracy)
Measured walk-forward over the same window (2022→now): hit = the 12-month peak after entry ≥ +20% · the numbers show whether a higher score really means a higher hit rate — not a forecast.
v1 — raw votes
| 0 votes | 60.0% | n=20 |
| 1-2 votes | 40.0% | n=20 |
| 3-5 votes | 94.4% | n=18 |
| 6-8 votes | 100.0% | n=4 |
v2 — accuracy-weighted (live)
ML Ranking Model (XGBoost v1)
Went live May 2026 — the live history is still too short to judge 12-month forecast accuracy. First 2–5 week rank-IC (Spearman): US -0.019 (a value near 0 = no forecast signal yet in the short window). We’ll keep reporting this every month as it is, good or not.
Backtested results are not investment advice and do not guarantee future outcomes · a hit-rate near 50% = no different from random.
Past accuracy does not guarantee future results · DAVI is an educational calculation tool, not investment advice.