Ophis takes dates on which similar events happened before, measures the distances between them, and projects those distances forward. Everything else it does is throwing away the results that bear no arithmetic relationship to what you gave it.
Ten minutes end to end. You need no maths beyond arithmetic.
Sixteen operations run against every pair of controls. Five controls make ten pairs, so five dates in gives 160 projections out — which collapse to fewer distinct days, because different operations often land on the same one. That collapsing is the first signal.
Every operation is an ordinary arithmetic expression starting X1+ or
X2+, which decides which of the two controls the result is added to. The rest is
a function of Y.
| O1 | X2+oph_round(Y) | 47 | → 10/06/2026 |
| O2 | X2+oph_flip(oph_round(Y)) | 74 | → 11/02/2026 |
| O3 | X2+Y/OPH_CRV | 9.25 | → 08/29/2026 |
oph_flip reverses the digits: 47 becomes 74. OPH_CRV is 5.08, one of
four constants the engine ships. Both are the author's, and neither is negotiable if you want
the numbers the original produced.
Alpha and Beta. Each operation carries a weight. At or above 1 it is an Alpha projection, below it a Beta. The author's own note: the Alpha projections are "future dates that are most likely to produce events that mirror the events of X-1 and X-2." The weight is not a label — it is added to the score directly.
Sixteen operations across ten pairs produce a lot of dates, and most of them mean nothing. The Multidimensional Spatial Recognition Filter is how the instrument decides which to keep: it takes the distance from each projection back to its control and looks that number up in three lists.
Those 390 numbers are the compressed output of six separate filters the author built — Annular Chronometric, Geodetic Chronometry, Hydrogen Spacing Angle, Fibonacci Sequence, IntraDecimal Matrix and Vortex Holography. Only the resulting numbers ever reached the software.
The broad set. A match is worth one point.
A smaller, stronger set — highly composite numbers, mostly. Worth two points.
Twelve numbers, and the only ones with a decimal. The author found them by subtracting every number from 1 to 15 000 from its own digit-reversal and keeping the ~5 % that never collapse to zero.
Order matters more than you would expect. Vortex is checked before anything else, so
43.5 matches even though half-day values are otherwise ignored — because 43.5
is a vortex number. And 76.3 matches Vortex 76.2, stealing what would
have been a plain Normal 76 and doubling its weight.
Here is the strongest row from the worked example the instrument opens with — five controls spanning a year, cast forward.
The rule worth internalising: the strongest resonance match multiplies rather than adds, and is withheld from the points sum so it is not paid for twice. The multiplier is a maximum, never a product — ten Important matches still multiply once.
Two operations agreeing matters more than one scoring well. A date several different operations independently land on is the thing the whole instrument exists to surface. Sort by Hits, not only by Score.
This is the part nothing in the software tells you, and it triples what you get out.
The author's flow chart takes three controls, not two: the two historical dates, and the date on which you are running the projection. That third date produces two more distances — one to each of the others — and every operation runs against all three.
Fourteen operations across three spans is the forty-two future dates the author's chart describes, collapsing to "38 or 39 actual dates" once duplicates merge. Enter only two controls and you see a third of the picture.
Add today's date as a third X-Date. That is the whole of it. The engine forms every pair automatically, so the two extra spans appear on their own — and the X-Dates panel has a one-click ☉ Today · Protocol Prime button for exactly this.
The author described Protocol Prime as a security measure — that an unauthorised user "can not use system to predict future events" without it. Read plainly, an operator who enters only two dates is simply not running the method.
Eight filters trim the output, and each can only ever remove rows. Three are on by default and do most of the work: hide anything before your last control, hide anything before today, and hide anything more than 2 559 days out.
Two more are worth reaching for once you have a feel for a data set. Minimum hits at 3 keeps only dates that several operations agreed on. Requires an MSRF match keeps only dates that resonate at all.
T-Dates work the other way round. Add the dates you already care about — when a team actually plays, when a market is actually open — and everything else disappears. It is the fastest way to ask "does anything land here?"
Ophis is arithmetic on the dates you supply. It projects intervals forward and ranks the results by how many independent operations and number-patterns agree. It does not know what your event was, and it has no access to anything beyond the dates in front of it.
The author was candid that the method eliminates rather than predicts: it "has ELIMINATED all future dates that bear no relativity to X-1 and X-2", leaving a shortlist on which "one, two or three of the projections will mark a future date when a similar event will occur again." A shortlist is what you get.
Treat it as a study instrument and a way of interrogating a calendar — which is what makes the Audit screen the most useful one in the application. Every number on every row can be traced back to the arithmetic that produced it. Click any result to see it.
Open the instrument