Built in the dark. Graded in public.
The Terminal is an NBA-only pricing operation. This is the working method, including the rules that stop us selling you our own mistakes.
Updated 5 October 2026 · the blend
Every line, priced before tip
Every night the Terminal builds its own projection for every player market on the slate: minutes and role first, then volume and efficiency separately, then matchup, pace and rest — simulated thousands of times into a full distribution, not a single number. That gives our own chance on every line, with the range around it.
What a row shows you: our chance the line lands, the range around it (how sure we are), the best price and which book has it, break-even (the chance that price needs), and a verdict. The lines on the night’s list carry a BLEND tag.
The blend, and the night’s list
Our model is one opinion. The market — every book’s price with its margin taken out, taken together — is another, and over a season it is a hard one to beat. The blend weighs the two into one chance, by how well each predicted results last season. The market carries most of the weight; our model moves the number where it knows something the prices don’t.
A line makes the night’s list — a blend edge — only when all of this holds: a price is more than four per cent above the fair price the blend puts on the line; our model leans the same way; at least three books price the line; and it is one of the three strongest in its game. Each one comes with a floor: the lowest price still worth taking. Below it, pass.
The weights are fitted on last season and frozen for this one. We tested the whole method on last season first, night by night, fitting only on the months before each night. About half the picks win, and that is the design, not a flaw: a blend edge is a price worth taking, not a lock. On a night with nothing worth taking, the list is empty and says so.
Three reads, one number
Three reads sit under our own number. The data read is what he has actually done. The model read is the simulation. The market read is what the books, between them, are saying. Our number lands between them, and every line shows all three, so you can see when they disagree — that disagreement is usually the story.
The verdict
Our own number carries a verdict of its own. Robust means the bet still pays at the low end of our range against the price on offer. Fragile means it only pays at our exact number — the range touches break-even, so it is a lean, not a play. Pass means no edge to claim. Withheld means we held the claim back: an edge past our sanity caps is treated as a pricing or data error until it is explained, never as an opportunity.
The verdict is the evidence under the blend: press any pick and you see why it is there. Most nights, most lines get a pass, and the Terminal says which ones.
News changes the minutes
On game days the injury watcher reads the official injury report every fifteen minutes. When a player is ruled out, it reprices his teammates from with-and-without-him minutes patterns — measured, tested out of sample, and applied only when the signal actually fires. Each repriced line lands in Alerts with a before and an after. Prices refresh every twenty minutes, and a line that moved is marked. On a quiet night the honest move is no move.
Last season’s test, and What hit
Every night of 2025–26 was priced as it would have been before tip-off and graded on the official box scores, losses included. It is a labelled backtest — it shows what the Terminal does, not how often it will land this season — and the night of blend edges on the front page comes from it. Inside the Terminal, last season is gone: from opening night it keeps its own record, What hit — every signal it published, graded off the box score, won and lost, by night, by week and by season. Never only the winners.
The community, and your record
The community chat has a thread for tonight and one per game. Tag a line on anything you say — type the player, tap his line — and it shows on your message as a card anyone can open. “I’m on this” counts people, never money. “Your number” is your own chance on a line; after the game the Terminal grades it off the box score, and the ladder ranks who reads the night best.
My Edge is your own bets, graded off the box score. It is encrypted on your device with a passphrase only you hold; the copy we keep for you is ciphertext we cannot read.
The rules we can’t break
- No naked percentages. Every probability ships with its range. A number without its uncertainty is marketing, not analysis.
- Too good is suspect. Claimed edges above our sanity caps are withheld, not sold. The best-looking number on a night is the one most likely to be wrong.
- No invented data. A failed feed is reported as failed. An empty state is a valid output. Nothing is backfilled to look complete.
- The record is append-only. Nights are committed before tip-off, no-play nights included, so history cannot be quietly rewritten. Misses stay published.
- Small samples say so. 22–11 is promising and not proof. When a range includes break-even we print that sentence ourselves, before anyone asks.
On multis
Books love multis because stacking legs compounds their margin. Most same-game prices pay far less than the legs are jointly worth — correlations are haircut crudely, and always in the house’s favour. The Terminal prices a multi with actual correlation maths and, most nights, the honest output is: this multi is worse than its legs. When a combination genuinely clears the bar, it says so, with the fair price next to the book’s.
What we are, and aren’t
Ritz Ariona is an analysis service. We publish probabilities, prices and a graded record. We are not a bookmaker, we take no bets, we hold no money, and we never place bets for anyone. Whether to bet, and how much, is yours — our job is to make sure that if you do, you are looking at an honest number first.
Not affiliated with or endorsed by the NBA or any team. All statistics from public sources. 18+, or 21+ where your jurisdiction requires it.