Bet on a capper's all-time grade and you might be following yesterday's story. We compared every graded capper's all-time grade against their last-30-days grade: 71.5% changed, and 57.3% moved two or more letter grades. The matrix below is the reason your "A+ lock" is on a losing streak.
The data: every figure below is computed from the CappersTracked dataset — 1,460 tracked handicappers, 170,170 graded picks, 372 currently graded (protected variant). Snapshot: 2026-07-29. Simulations are Monte Carlo runs with a fixed seed — every chart in this series is reproducible.
1. The Transition Matrix
Each row is the capper's all-time grade; each column is where they landed over the last 30 days. 239 cappers had both grades available:
| All-time \ L30 | A+ | A | B+ | B | C+ | C | D | F |
|---|---|---|---|---|---|---|---|---|
| A+ | 5 | 2 | 1 | 0 | 2 | 0 | 0 | 3 |
| A | 5 | 3 | 0 | 2 | 3 | 1 | 2 | 4 |
| B+ | 9 | 2 | 2 | 0 | 2 | 3 | 3 | 4 |
| B | 1 | 1 | 0 | 1 | 2 | 3 | 7 | 16 |
| C+ | 5 | 0 | 2 | 2 | 4 | 3 | 3 | 19 |
| C | 1 | 2 | 1 | 5 | 1 | 5 | 3 | 10 |
| D | 0 | 2 | 1 | 3 | 3 | 3 | 2 | 9 |
| F | 4 | 1 | 2 | 0 | 5 | 1 | 2 | 46 |
2. Which Grades Hold? Which Collapse?
For each all-time grade with enough data, here is what happened over the last 30 days — the share that held, improved, or dropped:
| All-time grade | Cohort | Held | Improved | Dropped |
|---|---|---|---|---|
| A+ | 13 | 38% | 0% | 62% |
| A | 20 | 15% | 25% | 60% |
| B+ | 25 | 8% | 44% | 48% |
| B | 31 | 3% | 6% | 90% |
| C+ | 38 | 11% | 24% | 66% |
| C | 28 | 18% | 36% | 46% |
| D | 23 | 9% | 52% | 39% |
| F | 61 | 75% | 25% | 0% |
3. The F Column: Nobody Is Immune
The most useful column in the matrix is F — it is the sink every grade feeds. Here is who fell into it over the last 30 days:
| All-time grade | Cohort | Fell to F in 30 days | Fell 2+ letters |
|---|---|---|---|
| A+ | 13 | 23% | 46% |
| A | 20 | 20% | 60% |
| B+ | 25 | 16% | 48% |
| B | 31 | 52% | 84% |
| C+ | 38 | 50% | 58% |
| C | 28 | 36% | 36% |
| D | 23 | 39% | 0% |
| F | 61 | 75% | 0% |
The surprise is not that mid-pack cappers fall — it is that the A+ and A cohorts fall too. A grade trails a small sample of recent bets; one cold 10–15 bet stretch is enough to drag an A+ down to a D or F. That is variance doing what variance does — not, in most cases, a capper suddenly forgetting how to pick.
4. Why Grades Move So Much
Three reasons, in order of size:
- Variance. As our trust curves study shows, a confidence band of ±20%+ at small samples means a 10-pick stretch can swing a grade several letters. Most grade changes are noise, not regime change.
- Volume weighting. The L30 grade is computed on a fresh, small window — so it reacts violently by design. That is its job: surface who's hot now.
- Real tilt. Some cappers genuinely spiral — the post-loss behavior that a 30-day window catches but an all-time grade buries.
How far do grades actually travel? Here is the full distribution of letter-grade moves across the 239 cappers who carried both grades:
| Change in grade | Cappers | Share |
|---|---|---|
| Stayed put | 68 | 28% |
| 1 letter | 34 | 14% |
| 2 letters | 42 | 18% |
| 3 letters | 42 | 18% |
| 4 letters | 30 | 13% |
| 5+ letters | 23 | 10% |
Read that again: 28% stability is the best case, the "nothing happened" outcome. The remaining third of the field is in motion every single month — which is why a grade from 90 days ago is mostly history.
How to use this: treat the all-time grade as identity and the L30 grade as momentum. A capper with a stable all-time A+ and a live L30 trend is the rarest combination — that is your follow list. A capper whose grade just collapsed is either regressing to the mean or tilting; both are signals. This is exactly why every CappersTracked profile shows both numbers side by side.
5. The Bottom Line
Grades are samples of recent behavior, not permanent verdicts. 71.5% of cappers changed grade within 30 days — plan for it. Follow the ones whose grade survives contact with the calendar.
All-time grade for identity, L30 for momentum — both sit side by side on every capper profile.