Does our confidence label mean anything?
Every projected earnings date on this site carries a label: high, medium or low confidence. A label like that is worthless unless the higher tiers are actually more accurate, so this page checks. A confidence label that does not track accuracy is decoration, and decoration on a number people plan around is worse than no label at all.
| Label | Scored | Companies | Median miss | Within 3 days | More than 3 days off | Within 7 days | More than 7 days off |
|---|---|---|---|---|---|---|---|
| high | 4,668 | 135 | 1 d | 69.1%67.0 to 71.1% | 30.9%28.9 to 33.0% | 91.1%89.5 to 92.6% | 8.9%7.4 to 10.5% |
| medium | 113 | 113 | 1 d | 59.3%50.8 to 67.9% | 40.7%32.1 to 49.2% | 88.5%82.4 to 93.8% | 11.5%6.2 to 17.6% |
| low | none | none | none | none | none | none | none |
Small figures are 95% intervals: the companies, resampled with replacement 2,000 times from a fixed seed, so every build prints the same numbers, keeping the middle 95% of the results. Each draw keeps all of a company's quarters together, because a company that changes its schedule misses several quarters in a row. A tier's interval prints from 20 companies.
We cannot say the labels order correctly. On the point estimates the higher tier lands within seven days slightly more often (high 91.1% against medium 88.5%), but that held in only 80% of 2,000 company resamples, short of the 95% we require. The gap between the tiers is inside the noise.
Who is scored: every company in our earnings snapshot, 142 in all: the 81 most-watched large and mid caps on the earnings calendar and 61 small caps added for insider-buying coverage. The list was chosen for coverage, not for how well its dates project. 136 have five or more filed reports and enter the backtest; scoring starts at each one's fifth report, so 135 companies have scored points.
Read the sample sizes before the percentages. 97.6% of scored projections carry a single label, so the tiers are badly unbalanced. A label that reports the same value almost every time carries little information even when it is accurate, and that is a limitation of the engine, not a detail of this page. The low tier has zero scored points, and the reason is structural: scoring only begins at a company’s fifth report, which guarantees enough prior dates to reach a higher tier. This backtest therefore cannot test the bottom of the scale at all.
The live record: projections graded as published
The backtest above is a replay: its record runs from 2005-09-14 to 2026-11-17, over dates that were already filed. Live measurement began on 2026-08-19, when we started logging each projected window as we published it. Record from 2005-09-14, measured from 2026-08-19: a gap of 7,644 days. Everything before 2026-08-19 is simulation; this section is the part that was measured.
A rate prints from 20 graded reports on 20 companies. The interval: the companies, resampled with replacement 2,000 times from a fixed seed, so every build prints the same numbers, keeping the middle 95% of the results. Median days outside the window: 0.
- The claim graded for each report is the projection that stood at the end of the day before it was filed: the window a reader last saw. A row logged on the report day itself is never that claim, because ingest runs every six hours and logs the next quarter hours after the filing.
- A report counts once the company's 10-Q or 10-K confirms it, through the same classifier the backtest uses, so a recent release can wait days or weeks before it is graded.
- The log stores the window, not the label, so this record tests the dates and cannot test high against medium. 313 projections on 142 companies so far: 136 graded, 140 still standing, 35 replaced by a newer window before their report, 2 logged with a window already past, which we set aside. 1 settled report has no fresh projection standing before it and is not graded. 2 of the graded claims carried a confirmed date when logged.
| Ticker | Logged | Published window | Report filed | Result |
|---|---|---|---|---|
| DIS | 2026-08-19 | 2026-11-07 to 2026-11-17 | 2026-11-17 | inside |
| HD | 2026-08-19 | 2026-11-12 to 2026-11-22 | 2026-11-17 | inside |
| DGII | 2026-08-19 | 2026-11-06 to 2026-11-16 | 2026-11-12 | inside |
| CEVA | 2026-08-19 | 2026-11-04 to 2026-11-14 | 2026-11-09 | inside |
| MARA | 2026-08-19 | 2026-10-29 to 2026-11-08 | 2026-11-09 | 1 day outside |
| KOP | 2026-08-19 | 2026-11-01 to 2026-11-11 | 2026-11-06 | inside |
| NHC | 2026-08-19 | 2026-11-01 to 2026-11-11 | 2026-11-06 | inside |
| REI | 2026-08-19 | 2026-11-01 to 2026-11-11 | 2026-11-06 | inside |
| USPH | 2026-08-19 | 2026-10-31 to 2026-11-10 | 2026-11-06 | inside |
| ABNB | 2026-08-19 | 2026-10-31 to 2026-11-10 | 2026-11-05 | inside |
| COP | 2026-08-19 | 2026-10-31 to 2026-11-10 | 2026-11-05 | inside |
| CRK | 2026-08-19 | 2026-10-28 to 2026-11-07 | 2026-11-05 | inside |
| NOG | 2026-08-19 | 2026-10-31 to 2026-11-10 | 2026-11-05 | inside |
| LMAT | 2026-08-19 | 2026-10-31 to 2026-11-10 | 2026-11-04 | inside |
| MCD | 2026-08-19 | 2026-10-30 to 2026-11-09 | 2026-11-04 | inside |
| MGY | 2026-10-12 | 2026-10-23 to 2026-11-02 | 2026-11-04 | 2 days outside |
| MTRN | 2026-08-19 | 2026-10-23 to 2026-11-02 | 2026-11-04 | 2 days outside |
| MWA | 2026-08-19 | 2026-10-31 to 2026-11-10 | 2026-11-04 | inside |
| PARR | 2026-08-19 | 2026-10-29 to 2026-11-08 | 2026-11-04 | inside |
| QCOM | 2026-08-19 | 2026-10-30 to 2026-11-09 | 2026-11-04 | inside |
| UBER | 2026-08-19 | 2026-10-29 to 2026-11-08 | 2026-11-04 | inside |
| ADUS | 2026-08-19 | 2026-10-28 to 2026-11-07 | 2026-11-03 | inside |
| AMD | 2026-08-19 | 2026-10-29 to 2026-11-08 | 2026-11-03 | inside |
| ATRC | 2026-08-19 | 2026-10-23 to 2026-11-02 | 2026-11-03 | 1 day outside |
| CRVL | 2026-08-19 | 2026-10-29 to 2026-11-08 | 2026-11-03 | inside |
| PFE | 2026-08-19 | 2026-10-29 to 2026-11-08 | 2026-11-03 | inside |
| SHOP | 2026-08-19 | 2026-10-29 to 2026-11-08 | 2026-11-03 | inside |
| SMCI | 2026-08-19 | 2026-10-29 to 2026-11-08 | 2026-11-03 | inside |
| TALO | 2026-09-22 | 2026-10-31 to 2026-11-10 | 2026-11-03 | inside |
| XYZ | 2026-08-19 | 2026-10-31 to 2026-11-10 | 2026-11-03 | inside |
| AAPL | 2026-08-19 | 2026-10-24 to 2026-11-03 | 2026-11-02 | inside |
| ALG | 2026-08-19 | 2026-10-31 to 2026-11-10 | 2026-11-02 | inside |
| NSSC | 2026-08-24 | 2026-10-28 to 2026-11-07 | 2026-11-02 | inside |
| PLTR | 2026-08-19 | 2026-10-28 to 2026-11-07 | 2026-11-02 | inside |
| TNC | 2026-08-19 | 2026-10-28 to 2026-11-07 | 2026-11-02 | inside |
| ABBV | 2026-10-12 | 2026-10-25 to 2026-11-04 | 2026-10-30 | inside |
| CVX | 2026-08-19 | 2026-10-25 to 2026-11-04 | 2026-10-30 | inside |
| AMZN | 2026-08-19 | 2026-10-24 to 2026-11-03 | 2026-10-29 | inside |
| CAT | 2026-08-19 | 2026-10-23 to 2026-11-02 | 2026-10-29 | inside |
| COIN | 2026-08-19 | 2026-10-24 to 2026-11-03 | 2026-10-29 | inside |
How this is measured
- Walk-forward, no lookahead. To score the report on a given date, the engine sees only the dates filed before it — both to project the date and to decide which confidence label it would have assigned at that moment.
- Real filed dates only. Every point is an 8-K earnings date the company actually filed. Nothing is simulated except the ordering.
- A pre-announcement is not the report. Companies file Item 2.02 for more than results: date notices, preliminary revenue, deliveries, guidance updates. An Item 2.02 counts as the quarter’s report only when the company’s 10-Q or 10-K follows it before its next one. The projection still sees every earlier filing, pre-announcements included, because that is what the engine had at the time. 352 filings in this sample are set aside that way and never scored as reports.
- Scoring starts at the fifth report per company, because the projection needs at least three prior dates and two usable quarter gaps before it can say anything.
- The label rule, as the engine stores it. High: five or more past reports, with quarter-to-quarter gaps (those of 75 to 110 days) whose standard deviation is at most 15 days when last year's same-quarter report anchors the projection, or at most 6 days with one weekday in at least 60% of reports when it does not. Medium: four or more past reports otherwise. Low: fewer than four. A cadence that has broken carries no label. The backtest assigns each point the label this rule would have given on the day before the report.
The backtest is not a track record. It replays the engine over dates that were already filed. It is not the same claim as the max pain accuracy ledger, which records what we published as we published it and refuses backfill. The live record above is our track record for earnings dates, and it starts the day logging started.