How the Rift Signal Works: Riftbound Zone's Price Outlook
The Rift Signal is Riftbound Zone's indicator that estimates whether a card's price is more likely to rise or fall over the next 2–4 weeks. You'll find it as a ▲/▼ arrow on the prices grid, on Price Movers and as the "Outlook" block on every card page. This guide explains exactly what it measures, which data it uses and how to read it.
Which price the prediction is based on
Key point: the Rift Signal predicts the Cardmarket Price Trend, the reference price Cardmarket computes for the whole product page, aggregating all languages and conditions (English, Chinese, French…). It does not predict the "EN minimum" you see as the headline price on our pages: that's the lowest price among EU-27 sellers for English cards only — a different, often more volatile series.
In practice: a ▲ arrow means "this card's overall reference value tends to rise", not "the EN minimum will surely go up". The two series usually move together, but they are not the same thing.
Which data it uses
Every night the algorithm recomputes the signal for every card with at least 14 days of history and a trend above €0.30, combining four families of data:
- Competitive meta — how often the card appears in recent tournament top decks, and whether its presence is growing or declining;
- Market supply — available copies and seller counts on Cardmarket and CardTrader: shrinking supply pushes prices;
- Price momentum — the trend direction over the last 7 and 30 days, with a penalty when the move looks already priced in;
- Cross-market confirmation — when Cardmarket, CardTrader and TCGplayer move in the same direction, the signal is more reliable.
The infrastructure. All of this — multi-market price collection, history, the predictive signal — runs on a high-end dedicated server we manage ourselves, not on shared hosting: that's what lets us refresh thousands of cards every night and answer in moments.
The two lanes: play cards and collector cards
Not all cards move for the same reasons, so the algorithm uses two distinct models:
- Play lane — regular deck cards: competitive meta and supply weigh the most;
- Collector lane — Showcase, Signature, alternate variants and cards above ~€100: supply, graded copy values (PSA/BGS/CGC) and how rare the graded population is weigh more.
How to read the signal
- ▲ Likely to rise / ▼ Likely to weaken — the estimated 2–4 week direction. We only show the signal when confidence is at least medium: no arrow means the model expects a stable price or lacks data.
- Score (−100 to +100) — the signal's intensity: the further from zero, the more the factors push in the same direction.
- Confidence — how complete and concordant the data is (history, multiple sources, confirmations).
- Reasons — the 2-3 main drivers behind the verdict, always consistent with the indicated direction.
How reliable it is
We validated the model with a walk-forward backtest over three months of history (May–August 2026): the model is "sent back in time" and its predictions are compared with what actually happened. Result: rise signals proved correct in 78% of cases at 14 days (versus a 48% random baseline), fall signals in 58%. The benchmark here is the Cardmarket Price Trend too — the same series the signal predicts.
The limits (really, read them)
The Rift Signal is an experimental statistical indicator. It doesn't know about Riot announcements, tomorrow's tournament results, reprints or bans: events like these can flip any prediction overnight. Use it as one more tool to orient yourself, not as a promise. It is not financial or purchase advice.
Comments
Log in to comment.