What Is the RiftCompare Index?

What the RiftCompare Index Actually Measures

The RiftCompare Index is a single number meant to answer one question: is the Riftbound secondary market, taken as a whole, worth more or less than it used to be? It is not the price of any one card, and it isn't an average of "everything for sale." It's a tracked basket of specific cards whose combined value is rebased to a starting point, so the week-to-week movement of that basket tells you something about market direction rather than about one chase card getting hot.

Think of it the way a stock index works. The S&P 500 doesn't tell you what any single company is worth - it tells you whether large-cap US equities broadly went up or down. The Index on /market is built the same way for Riftbound singles: a defined group of cards, priced once a week, combined into one line you can watch over time.

This matters because individual card prices are noisy. A single copy listed low because a seller needed cash fast, or high because one shop is asking too much, can make a card's price chart look dramatic without meaning anything about the format or the game's overall health. An index smooths that out by design.

How the Basket of Cards Is Chosen

Not every card in Riftbound belongs in the Index, and that's intentional. The basket is the 200 cards with the highest recorded search count on RiftCompare that currently have a live price in the selected market: the cards players actually look up, which is the closest thing to demand the site can measure. Each card's weight follows its search count, capped so no single card can be more than a fifth of the basket.

The goal is representativeness, not completeness. Trying to include every printed card would let hard-to-find cards with one stray listing drag the number around. A basket of the cards people actually care about produces a steadier, more useful signal.

Why a Changing Basket Doesn't Move the Index

The basket is not frozen on a review date. It is worked out fresh from search data, so it turns over as the metagame shifts and as a new set's cards start getting searched. On its own, that turnover would make the Index jump whenever a differently priced card joined or left. The next section is how it doesn't.

How the Index Actually Combines Prices — Chain-Linking

Every card in the basket gets a price snapshot once a week - effectively a weekly "closing price": the cheapest in-stock listing we track for it across our markets, recorded in US dollars. It is a listing price, never a sale price; RiftCompare does not record sales.

The naive way to turn a basket of prices into one number is: average them each snapshot, then rebase the first snapshot to a round starting value like 100. That's how a lot of simple indices work, and it's how an earlier version of this one worked. It has a real flaw, though: the basket isn't fixed forever - it's the 200 most-searched cards right now, and that list moves as the metagame shifts and, especially, whenever a new set releases and a wave of newly-revealed cards suddenly gets searched heavily. A plain average jumps the instant a differently-priced card enters or leaves the basket, whether or not anything anyone actually owns changed price at all - which makes the naive version misleading at exactly the moment it matters most, a set launch.

The fix is called chain-linking, and real financial indices use the same idea to absorb constituent changes without a discontinuity. Instead of averaging price levels at each snapshot, the Index computes each snapshot's percentage move using only the cards that have a tracked price at both that snapshot and the previous one - then multiplies that move onto a running total that starts at 100.

The practical effect: a card that just entered the basket - a brand-new set's card climbing into the top 200 by search, say - has no earlier price to compare against, so it simply sits out the calculation for the snapshot it debuts on. It starts contributing from the next charted snapshot onward, once it has two consecutive tracked prices to form a ratio from. The basket can turn over completely over months and the Index still won't move on that account alone - only real price changes do.

This is also why the Index is most useful looked at over stretches of time rather than a single snapshot. The earliest stretch of the chart - back when only a handful of cards had any tracked price yet - is thinner than the rest for exactly that reason: fewer cards means more weight riding on each one. Chain-linking still keeps that stretch honest (a card joining never jumps the level, whether it's card three or card three hundred), it's just a noisier read of the market than a later week with the full basket reporting. The trend across weeks and months is where the signal lives.

When the way a price is sourced changes

Chain-linking has one weakness worth being open about: a change in how a price is measured looks exactly like the market moving, and because every step is multiplied onto the last, a one-off change would stay in the level forever. So when the data source itself changes, the Index is chained across it — the steps inside that window are charted flat and the prices re-base, which is what stock-index providers do over a methodology change.

That has happened once. On 23 September 2026 the US TCGplayer price switched from TCGplayer's market price to the cheapest English Near-Mint listing, which sits noticeably lower for most cards. Steps ending between 23 September and 1 October 2026 are charted flat for that reason, so the Index does not report the change as a market-wide fall. Card price charts are not re-based, so for about a week after the switch some US charts show a step down. The price movers and every 7-day change skip any comparison that would reach across the switch, so they never report it as a fall.

A Concrete Example: A New Set's Launch

Say Origins, Spirit Forged and Unleashed cards make up most of the Index the week before a new set releases, each trading in an ordinary $5-$40 range. The set launches, hype cards get searched heavily, and within two or three weeks several of its cards - some already trading at $80 and up - climb into the top 200 by search and enter the basket.

Under a plain average, that alone would drag the Index's level up sharply the moment those cards entered, even though none of the older cards had moved at all - a false signal blamed on "the market," when it was really just new, expensive cards joining the average. Under chain-linking, those new cards have no price before their own release, so they aren't part of any snapshot-to-snapshot return until they've been tracked for two consecutive snapshots. The Index keeps telling the same story it was already telling about the older cards, and the new set's cards only start moving the number once there's an actual prior price to compare against.

The Exact Formula

The plain-English version above is what you need to interpret the number day to day. This is the literal arithmetic, published because a health check you can't check isn't much of one - and it's exactly what the code computes, not a simplified stand-in for it.

1. Constituents. The 200 cards on RiftCompare with the highest recorded search count that currently have a live price in the selected market.

2. Weight. Each constituent's raw weight is 1 + searchCount (the +1 means a market with zero recorded searches still produces a valid weight instead of dividing by zero). Every raw weight is then capped at 20% of the basket's total raw weight - a card whose raw weight would be worth more than a fifth of the whole basket is clipped down to exactly that fifth. Call constituent i's resulting weight w[i], and the sum of every constituent's weight W.

3. Per-snapshot price. p[i, t] is constituent i's lowest tracked live price on snapshot t, carried forward from its last known price when this exact snapshot has no fresh reading for it (so a card briefly out of stock doesn't vanish from the calculation).

4. Starting point. The series starts at the first snapshot t0 where any constituent has a tracked price at all - as far back as the data goes, full stop. Index(t0) = 100.

5. Chain-linked step. For each snapshot t after t0, let C be the set of constituents priced at both t and the previous charted snapshot t':

return = ( Σ w[i]·p[i,t] for i in C ) ÷ ( Σ w[i]·p[i,t'] for i in C )

Index(t) = Index(t') × return

That's the whole thing - five steps, no persisted state, recomputed from scratch every time from the raw tracked prices, nothing set-specific hard-coded anywhere in it. Step 5 is the literal answer to "what happens when a new set releases": a card with no price at t' is outside C for that step by construction, so it cannot affect return no matter how differently it's priced from the rest of the basket - it only starts actually moving the number from its first two consecutive tracked prices onward. Notice C doesn't need to be most of the basket - it just needs to be non-empty. An earlier version of this formula also required C to clear a coverage threshold before trusting a step, on the theory that a thin C meant an untrustworthy reading. It didn't actually add any protection beyond what the exclusion rule above already gives for free, and it meant the chart started later than the real data did - so it's gone.

Regions. The Index defaults to the US market. Every other tracked region (AU, UK, SG, CA, EU) runs the same five-step computation with its own constituents - the most-searched cards with a live price in that market - on the same weekly price series, converted into the region's own currency at our reference rates. Pick one from the Market selector on the page.

Index vs. Movers: Two Different Questions

It's worth being explicit about what the Index is not for, because /movers exists to answer a genuinely different question. Movers is about which individual cards changed price the most recently - the specific singles that jumped or dropped week over week, useful if you're trying to time a buy or sell on a particular card.

The Index doesn't try to do that job. It won't tell you that one card spiked because of a tournament result or a reprint rumor. What it tells you is whether the format as a whole is trending up or down. A card can be a huge mover in either direction while the Index barely budges, because it's one card out of a basket. Conversely, the Index can drift steadily even when no single card is making headlines that week - that's often the more meaningful signal, since it reflects broad, sustained demand rather than one card's news cycle.

If you're deciding whether to buy a specific card right now, check Movers. If you're trying to understand whether Riftbound singles in general have gotten more or less expensive since you started collecting, the Index is the number to watch.

Using the Index as a New Collector

If you're still getting oriented in the game itself, it's worth pairing this with our beginner's guide to Riftbound before you lean too heavily on market data - understanding what makes a card mechanically strong or scarce will help you interpret why the Index moves the way it does, not just that it moved.

Used honestly, the Index is a health check, not a trading signal. It won't tell you when to buy a specific card. It will tell you, over time, whether the market you're buying into is expanding or contracting.