Track Record
The goal
Beat the S&P 500 with disciplined value stock-picking. Concretely: if you took our BUY and STRONG BUY calls at any point in time (time T) and held them, you should end up ahead of simply buying an S&P 500 index fund at that same moment. This page measures exactly that — every week, against reality — and shows it whether the model is winning or losing.
Track record to date
CalibratingCollecting cohorts. No horizon has enough matured, independent observations to mean anything yet.
8
cohorts issued
3
matured (countable)
Apr 18, 2026
since
Model version: v10-2026-06 — undecided
No version has 12+ matured cohorts yet — performance is not yet judgeable. Current version v10 ships as a challenger on trust, NOT proven better. Verdict pending matured data.
All recommendations vs SPY
BUY+ basket return minus the cap-weighted SPY ETF, averaged across every matured cohort — all model versions pooled. This is the headline: did following our calls beat buying the ETF?
| Horizon | Matured cohorts | Avg vs SPY | Hit rate |
|---|---|---|---|
| 30 days | 2 | +7.52pp | 100% |
| 90 days | 3 | +0.15pp | 67% |
| 180 days | 0 | — | — |
| 365 days | 0 | — | — |
By model version
The same cohorts, grouped by which model version issued them. History is never discarded when the model changes — the version is a label so we can see whether newer logic actually does better.
v10-2026-06current · since Jun 22, 2026 · 5 cohorts
| Horizon | Matured cohorts | Avg vs SPY | Hit rate |
|---|---|---|---|
| 30 days | 2 | +7.52pp | 100% |
| 90 days | 0 | — | — |
| 180 days | 0 | — | — |
| 365 days | 0 | — | — |
v0-pre-versioning · since Apr 18, 2026 · 3 cohorts
| Horizon | Matured cohorts | Avg vs SPY | Hit rate |
|---|---|---|---|
| 30 days | 0 | — | — |
| 90 days | 3 | +0.15pp | 67% |
| 180 days | 0 | — | — |
| 365 days | 0 | — | — |
Methodology
Cohorts, not "since inception"
Each weekly snapshot of our recommendations is a cohort — a dated vintage of calls. Measuring an old snapshot against today's price conflates skill with elapsed time (an old call looks good just because the market drifted up). Instead we measure each cohort over fixed forward horizons (30, 90, 180, 365 days), so every cohort is judged on the same yardstick.
When does a result start counting?
A cohort's horizon only counts once that many days have actually elapsed (it has "matured"). A recommendation made last week tells us nothing about its 90-day performance until 90 days pass. Pending horizons are excluded, not guessed.
The model evolves — but no history is thrown away
When the valuation or scoring logic changes, we bump the model version and tag each cohort with the version that issued it. We do not discard older cohorts — every past recommendation really did beat or lose to SPY over its window, and that's genuine performance data. The headline pools all cohorts across versions; a separate per-version breakdown lets us check whether newer logic actually improved on the old.
When is it "stabilized"?
Confidence grows with the number of independent, matured cohorts spanning different market conditions:
- • Calibrating — under 12 cohorts (~3 months). Noise.
- • Early signal — 12+ cohorts. Directional only.
- • Emerging — 52+ cohorts (~1 year). One regime.
- • Stabilized — 150+ cohorts (~3 years) across bull and bear markets. Trustworthy.
Benchmark & known limitations
The benchmark is the actual SPY ETF — the cap-weighted S&P 500 fund a real investor would buy as the alternative to our picks — measured on a total-return basis (dividends reinvested), so the comparison gives the ETF full credit for its payouts. Its return is measured over each cohort's exact window. One asymmetry remains: our BUY+ basket return is currently price-only and equal-weighted within itself, so the ETF side is, if anything, given a slight edge — we'd rather understate our own result than overstate it.
Personal analysis tool, not financial advice. Generated Jul 13, 2026 from 8 weekly snapshots.