The promise
Every dealer level Runir publishes gets graded against what price actually does the next session. Four discrete outcomes, four published states. The grade is banked alongside the underlying measurements so it's reproducible, not retrofitted.
Those grades compound. Per-ticker accuracy appears on each name's page once enough has been scored. Universe-wide accuracy ships to Discord every Friday close. The numbers are observed, not predicted — and the sample size always ships with the rate.
What gets published
Four times every market day, for the most options-active names in the universe, Runir publishes the levels active options traders care about:
- The call wall — where dealer hedging tends to defend against price moving up.
- The put wall — where dealer hedging tends to defend against price moving down.
- The gamma flip — where dealer hedging switches from pinning a range to chasing the trend.
- Heat — a one-number summary of how active a name is right now.
- Plays at the walls — the covered-call and cash-secured-put strikes worth selling, sized to the dealer levels.
How accuracy gets measured
The morning after every close, an automated process reads yesterday's published walls and asks the same question for each one: did price stay near it, move away from it, blow through it, or never test it?
The four answers — pinned, rejected, broke, untested — are simple, mathematical, and determined the same way every time. There's no human judgment involved in the classification. The thresholds are set in advance and applied uniformly across every name in the universe.
Every scored wall is banked with the raw measurements alongside the verdict, so the rate we publish is reproducible from the data itself — never an after-the-fact re-telling.
What we publish about ourselves
Per-ticker reliability appears on every name's page once enough walls have been scored to be statistically meaningful. The ribbon shows: total called, held (with percentage), broke (with percentage), best call, worst miss. Below that threshold, the ribbon stays hidden — at small sample sizes there's nothing honest to claim.
The Friday Accountability Card posts to Discord at every Friday close. It aggregates the week's grades across the whole universe. While the cumulative sample is small, the card says so explicitly in its subtitle.
Horizons
Every options trade has a time window, and different traders care about different windows. A weekly trader cares about gamma scalping today's flow; a wheel seller wants the canonical 30–45 day premium plays; a positional investor sells quarterly+. The same row of data — same strike, same delta, same vol/OI — means very different things at different horizons.
On the Unusual + Premium boards, every row carries a colored badge disclosing its trade horizon. Classification is by days-to-expiry against the trading day:
- WEEKLY — 1 to 7 days. Short-term flow, gamma scalpers, weekly sellers. The dominant horizon on most "unusual" rows because vol/OI ratios spike fastest near expiry.
- MONTHLY — 8 to 45 days. The canonical wheel range. Most retail premium-selling lives here.
- LEAPS — 46+ days. Quarterly to annual. Positional commitment, not short-term tactics.
0DTE contracts (expiring same day) are filtered out of both boards before ranking. End-of-day churn produces huge vol/OI ratios that don't represent positioning, and by the time a Runir post lands those contracts are already worthless. The boards only show contracts that have at least a session of life left.
Thresholds are stable across every ticker and every session. The badge is a colored credibility signal, not a filter — every row is still ranked by the same vol/OI or yield math; the badge just helps the reader self-filter at a glance for the horizon they actually trade.
Wall enrichments
Sometimes a published wall coincides with a level from a different framework. When that happens, the wall row says so. Today the page checks for Fibonacci confluence: if a wall sits within 0.5% of a standard Fib retracement (0.382, 0.5, 0.618, or 0.786) of the trailing-90-calendar-day swing range (~60 sessions), the annotation fires.
Framing matters here: this is a credibility multiplier on the wall, not a separate level. The wall is still the primary claim. When two independent frameworks point at the same price, the wall is more believable — that's what the annotation is reflecting. Runir is not a charting tool.
The swing window is fixed at the trailing ~60 sessions across every ticker. Tolerance is fixed at 0.5%. Both are stable across days — no cherry-picking the window that makes confluence look strongest. The architecture supports additional enrichers (range percentile, prior reaction zones, VWAP proximity) which will appear on the same row pattern when they prove out.
Why this matters
Most finance-Twitter accounts and most paid services publish predictions without ever publishing their hit rate. A 64% claim at n=8 is noise. The same claim at n=200 is a real signal. Without the sample size, you can't tell which one you're reading.
Runir publishes the rate, the count, and the window. Always. This is the brand promise made structural — and it's the part that competitors can't easily clone, because the rate is a time-banked artifact of months of scored predictions, not a number anyone generates on demand.