Claim levelRung 1. No validated predictive claim.
The questionHow much of whitetail movement timing comes from public data, and how much has to come from your own ground?
The ruleThe test is set and frozen before we see the answer. We publish the result either way.
ScorecardPublished and frozen before the season
Season testPre-registration open
RackIQ™ is being built to read your ground, make a call, and keep score of whether it was right. We have not proven it predicts anything on your property, and we will not tell you we have.
This page is where we commit to how we will be graded, before the season that grades us. The scorecard below is frozen. It says exactly what counts as a hit, what counts as a miss, and what we are not allowed to count at all, and it was written before we had a single result to be tempted by.
As far as we can find, no company in this category publishes its grading rules at all, let alone before the season. That is the part we think is worth publishing. The engine itself, the datasets we test against and our internal validation write-ups are not published here, and that is deliberate: they are of no use to a hunter and of real use to a competitor. They are available to researchers and reviewers on written request.
The scorecard · Frozen August 17, 2026, before the season
What counts as a hit, and what we are not allowed to count
This is the whole of it. It was written before we had a single season result to be tempted by, and it does not change now that the season is starting.
The governing principle: a call is never scored differently because the hunter followed it or ignored it. Compliance changes which bucket a call lands in, never how generously it is judged inside that bucket. Anything else lets us quietly reward obedience and call it accuracy.
| Situation | Class | Why |
|---|
| Called good, hunter sat the called stand in the called window, saw target movement | hit | the call was tested and held |
| Called good, hunter sat the called stand in the called window, no movement | miss | tested and failed, and this is the row we must never quietly drop |
| Called poor, hunter sat anyway, no movement | hit | a correct negative is a correct call |
| Called poor, hunter sat anyway, saw movement | miss | tested and failed |
| Hunter did not hunt at all | untested | never enters the numerator or the denominator |
| Hunter hunted a different stand than the one called | untested for the stand claim, gradable for the day claim | the timing call was still exposed; the location call was not |
| Hunter killed at a different stand | untested for the stand claim | a kill is not evidence about a stand nobody sat |
| Sit cut short, bumped a deer walking in, weather, anything | ungradable if the window was not covered | exposure was lost, and a lost sit is not a failed call |
| Called good, hunter sat, saw nothing, and no camera on the property was operational | ungradable for the movement claim | no exposure means no observation |
Two arms, never pooled. The stand claim (where to sit) and the movement claim (whether deer moved in that window) are scored separately. Collapsing them lets a right-for-the-wrong-reason result read as a clean win.
The messy cases resolve toward ungradable, not toward a hit. Where a situation could plausibly be argued into either column it goes to ungradable, and the count of ungradable results is published alongside the rest. A high ungradable rate is itself a finding about the method, and hiding it would be the easiest way to flatter the number.
No zero without exposure. A non-observation only counts as evidence when we can show the woods were actually watched. A call that was never tested is not a miss.
Pre-registration 001 · Frozen before the answer
The season test
Our pre-season work measured the public-data layer with our own ground read switched off. The season test switches it on, and it is the one that matters.
The question. Does a property's own history improve the next call on that property, above a locked public-data-only comparator?
How it is scored. Each call is recorded forward, dated, before the day happens. It cannot be back-fitted. At season's end the calls are reconciled against what was actually logged and the number is published here.
What it has to beat. A separately locked comparator built from public data only, fixed before the season and not adjustable afterward. An earlier version of this page named a single number carried over from our collar work. That pass line was retired on August 20, 2026 by a published amendment to the protocol, before the season opened and before any gradeable outcome existed, because a collar-derived number and a hunter-outcome number do not measure the same thing and cannot be compared honestly.
What we do if it fails. We publish it, we say what we think went wrong, and we name the next test. A result that only counts when it is good is not a result.
Note · August 28, 2026
What changed on this page
This page previously carried three detailed write-ups of our own pre-season validation work, with the datasets, the figures and the engineering behind them. They have been moved off the public page.
The season test below is unchanged. Its grading rules are the ones frozen on August 17, and on the date of this edit no gradeable season outcome existed, which is a fact in our own data rather than an assurance. Nothing here has been revised to look better, and the commitment is the same one: we publish the season result, pass or fail.
The reason for the change is narrower than it looks. That material described our test apparatus rather than our claims, it is of no use to a hunter, and its actual readership was people building competing products. Researchers, reviewers and anyone with a serious interest can have the full pre-season write-ups, methods and figures on written request.
