Home › Accuracy
How accurate are our predictions?
We grade ourselves against reality. Every figure on this page measures one thing: how close our predicted dividend announcement dates came to the day the board actually declared.
Predicted announcement (declaration) dates
This is what the site predicts: the day a board is expected to declare its next dividend, weeks before that dividend even has an ex-date. That estimate carries its own High or MediumMed quality badge, graded separately because the two are not equally precise.
Backtest — measured on historical declarations
This is a backtest: for every past declaration, the estimate is reconstructed using only data available before it, then checked against what the board actually did (2023-onwards holdout). It is not a live record graded against real declarations as they land — that's tracked separately, below, once there's enough of it to publish. Right week is the calendar week we filed the prediction under, not a seven-day window either side of it — which is why it can read slightly lower than within 3 days: an estimate three days early can still fall the wrong side of a Monday.
| Quality | Exact to the day | Within 3 days | Right week | Beyond 30 days |
|---|---|---|---|---|
| High | 62% | 86% | 84% | 0.7% |
| MediumMed | 41% | 83% | 78% | 0.2% |
Confirmed in real time
Separately, we keep a running record of predictions we published before the company announced anything, then grade each one the moment the real declaration lands — never a figure recomputed afterwards with the benefit of hindsight. So far 122 predictions have been graded this way:
| Quality | Predictions | Exact to the day | Within 3 days | Right week | Beyond 30 days |
|---|---|---|---|---|---|
| High | 106 | 54.7% | 89.6% | 89.6% | 0.0% |
| MediumMed | 16 | 37.5% | 87.5% | 62.5% | 0.0% |
Recent predictions vs. actual declarations
Every dividend declaration in our data over the past 60 days that we predict, with the date we gave and the date the board chose. 315 declarations; at High quality 63% landed on the exact day and 93% within three, at MediumMed quality 57% exact and 93% within three.
Each prediction is computed from declarations dated before the one it is graded against, so it never sees its own answer. We show a per-company list because it is checkable in a way an aggregate is not.
| Stock | Quality | Predicted | Board declared | Off by |
|---|---|---|---|---|
| TOL | High | Wed, Sep 16 | Wed, Sep 16 | on target |
| OLP | High | Wed, Sep 9 | Wed, Sep 16 | 7d early |
| NWFL | High | Tue, Sep 15 | Wed, Sep 16 | 1d early |
| ISTR | High | Wed, Sep 16 | Wed, Sep 16 | on target |
| GSBC | High | Wed, Sep 16 | Wed, Sep 16 | on target |
| GHI | High | Tue, Sep 15 | Wed, Sep 16 | 1d early |
| PEB | MediumMed | Mon, Sep 14 | Tue, Sep 15 | 1d early |
| LADR | MediumMed | Mon, Sep 14 | Tue, Sep 15 | 1d early |
| IIPR | MediumMed | Mon, Sep 14 | Tue, Sep 15 | 1d early |
| HBNC | High | Tue, Sep 15 | Tue, Sep 15 | on target |
| FULTP | High | Tue, Sep 15 | Tue, Sep 15 | on target |
| CTRE | MediumMed | Mon, Sep 14 | Tue, Sep 15 | 1d early |
| PEGA | MediumMed | Mon, Sep 14 | Mon, Sep 14 | on target |
| RWT | High | Thu, Sep 10 | Thu, Sep 10 | on target |
| INVH | High | Fri, Sep 11 | Thu, Sep 10 | 1d late |
| AMAT | High | Thu, Sep 10 | Thu, Sep 10 | on target |
| VZ | High | Fri, Sep 4 | Wed, Sep 9 | 5d early |
| ORCL | High | Tue, Sep 8 | Wed, Sep 9 | 1d early |
| NLY | High | Wed, Sep 9 | Wed, Sep 9 | on target |
| AGNCP | High | Thu, Sep 10 | Wed, Sep 9 | 1d late |
| AGNCN | High | Thu, Sep 10 | Wed, Sep 9 | 1d late |
| AGNCM | High | Thu, Sep 10 | Wed, Sep 9 | 1d late |
| AGNCL | High | Thu, Sep 10 | Wed, Sep 9 | 1d late |
| UHT | High | Wed, Sep 9 | Tue, Sep 8 | 1d late |
| SUI | High | Wed, Sep 2 | Tue, Sep 8 | 6d early |
| ALCO | High | Fri, Sep 11 | Tue, Sep 8 | 3d late |
| PLOW | High | Fri, Sep 4 | Fri, Sep 4 | on target |
| VIRC | High | Tue, Sep 1 | Thu, Sep 3 | 2d early |
| VICI | High | Thu, Sep 3 | Thu, Sep 3 | on target |
| HOFT | High | Tue, Sep 8 | Thu, Sep 3 | 5d late |
| CASY | High | Thu, Sep 3 | Thu, Sep 3 | on target |
| NTAP | High | Wed, Aug 26 | Wed, Sep 2 | 7d early |
| HPE | High | Wed, Sep 2 | Wed, Sep 2 | on target |
| FUND | High | Wed, Sep 2 | Wed, Sep 2 | on target |
| SELF | High | Tue, Sep 1 | Tue, Sep 1 | on target |
| RVT | High | Wed, Sep 2 | Tue, Sep 1 | 1d late |
| RMT | High | Wed, Sep 2 | Tue, Sep 1 | 1d late |
| PMT | High | Wed, Sep 16 | Tue, Sep 1 | 15d late |
| JHS | High | Tue, Sep 1 | Tue, Sep 1 | on target |
| JHI | High | Tue, Sep 1 | Tue, Sep 1 | on target |
How we measure this — honestly
It would be easy to cheat at this: re-run today's model over past dates and "predict" dates it has already seen. We don't. For every past declaration, we rebuild the prediction using only the data available before that date — the model never sees the answer it's being graded on. That's the same point-in-time discipline as a proper backtest, and it's why these numbers reflect what you'd actually have got at the time, not hindsight.
Predictions are still statistical estimates — a company can change or cancel a dividend without warning, which no method can foresee. This is educational information, not financial advice. See how the predictions work or browse the upcoming calendar.