Scorecard
The record, counted in the open.
Every story names the dated tests that would prove it wrong. This page counts them: how many are on the clock, how many land this month, and how the resolved ones came in. Reference date: .
The Docket
Tests on the clock.
A dated test has a day or a month on the calendar; a signal-based test resolves on a print, a filing or a price and joins the calendar when its date is known. Of the 11 due in October 2026, 10 are still pending. Pre-specified means the words were approved on or before the test’s date. Specified after the date means they were written afterwards: such a test is counted and shown, and never rated. An open question has no words yet.
- Tests
- 75
- Dated
- 49
- Due this month
- 11
- Resolved
- 2
- Pre-specified
- 37
- Specified after the date
- 1
- Open questions
- 37
Results
By verdict.
No hit rate is shown until ten results exist. 2 of ten so far. Until then the counts stand on their own; the rate per verdict appears at the tenth result, because a smaller sample flatters or punishes by chance.
The hit rate counts only tests whose words were approved before the date and whose result held or missed the Call; a Call held means the result the Call expected, whatever the test’s question asked.
Inconclusive share: 0 of 2 results inconclusive (0%). An inconclusive result is allowed only when the named series was late, revised or silent, and it says what was read, what was missing and the next step with a date.
- Results
- 2
- Call held
- 2
- Call missed
- 0
- Inconclusive
- 0
- Call informed
- 0
Open the Docket Read the results Download the tests and results
The numbers
Every number the Calls gave, on one line from no to yes.
36 tests carry a number recorded before the date; a resolved one shows the answer at the end of its line. The Call’s number is the chance, from 0 to 100, that the Call gives the yes answer to a test’s question, recorded before the test’s date; it is DeepStack’s own reading, not a market price and not advice. The readers’ mark on the same line is the readers’ median, the middle of the numbers signed-in readers gave the test, shown from 25 readers.
- p07-t02 · no
- p13-t01 · yes
- p09-t01 · yes
- p07-t01 · either
- p10-t01 · yes
- p11-t01 · yes
- p12-t01 · no
- p14-t01 · yes
- p08-t01 · either
- p12-t02 · no
- p09-t02 · no
- p03-t01 · yes
- p14-t02 · no
- p09-t03 · either
- p11-t02 · yes
- p12-t04 · no
- p16-t02 · either
- p16-t03 · either
- p16-t04 · yes
- p08-t02 · yes
- p13-t02 · no
- p17-t03 · yes
- p15-t01 · no
- p04-t01 · no
- p15-t02 · yes
- p14-t03 · yes
- p14-t04 · no
- p17-t01 · yes
- p17-t02 · yes
- p13-t03 · yes
- p17-t04 · no
- p15-t04 · no
- p15-t03 · no
- p10-t06 · either
- p12-t05 · When it filesyes
- p12-t06 · When it filesyes
0 · no50100 · yes
Brier score: not shown until ten results carry a probability recorded before the date (1 so far). From the tenth result whose specification carried a probability before the date, the Scorecard prints the Brier score: the mean of the squared gap between the probability the Call gave the yes answer and what happened, where 0 is perfect and 0.25 is a coin flip; a public consensus is printed beside it where one exists.
The Call, the readers, the machines
One rule for the house, the readers and the machines.
Every number recorded before a test’s date, the Call’s, each signed-in reader’s and each model’s, gets a Brier score when the test resolves: the squared gap between the number, read as a fraction, and the answer, one for yes and zero for no. Lower is better; 0.25 is what a coin flip earns. The readers’ median on each test is scored as one forecaster. A mean prints from the tenth resolved test, with the count beside it.
| Forecaster | Resolved tests | Mean Brier score |
|---|---|---|
| The Call | 0 | Not shown until ten resolved tests (0 so far) |
| The readers’ median | 0 | Not shown until ten resolved tests (0 so far) |
The readers’ median and the Call have no resolved test in common yet. As of October 21, 2026.
The readers’ table Put your own number on a test Forecasting bots on the Docket
The Calls
Where the house stood, counted.
Each Call takes a side and a conviction from 1 to 5, the whole scale: a 1 or a 5 is allowed. The tones and the convictions of every Part are counted here at each build, so a lean in the house’s Calls shows as a number.
- Constructive
- 1
- Split
- 9
- Skeptical
- 7
- Conviction 1
- 0
- Conviction 2
- 0
- Conviction 3
- 11
- Conviction 4
- 6
- Conviction 5
- 0
The scorecard counts tests and results; the readers appear only as their median, one forecaster. No person named in a story is given a score, a ranking or an accuracy. Status moves only when a result is published after a human review. Analysis and opinion, not investment advice.