Methodology

Sources

Decks and events come from tcdecks.net (the community's reference for Premodern results: events, players, placings) and MTGTop8 (placings and card lists). We use the card lists only for statistics (card trends, archetype evolution, prices); every deck and event opens on its source site. Cards and images come from Scryfall, prices from the Cardmarket price guide (EUR) and TCGplayer market prices via Scryfall (USD), and ban-list dates from public announcements (0 changes recorded). The latest event in the data is from 25 Oct 2026; recent months keep filling in as results are posted, so the last month is usually incomplete.

The field and shares

The field is every tcdecks.net deck plus every MTGTop8 deck that is not the same deck as a tcdecks one (we link decks reported on both sites, so each deck counts once). Tcdecks.net supplies 100% of the field.

Share is an archetype's decks ÷ all decks in the field for the chosen range and filters. We call it "share of reported decks" because it is not a share of everyone who played: see the top-8 bias below.

Card lists. 0% of decks have a published card list (from MTGTop8; a tcdecks deck uses its linked MTGTop8 list). Card statistics, signature cards and prices come from these lists only, and every such number shows how many lists it is based on.

Top-8 bias

Most events only publish their top 8, so 94% of reported decks are top-8 finishes. Shares therefore describe what does well at events, not everything that is played. Sources publish placings, not match results, so this site has no win rates; be wary of sites that show win rates without saying where they come from.

How decks are classified

Each deck in the field got its archetype from, in order of preference:

• MTGTop8 label, where it reliably maps to one archetype: 3,144 decks (100.0%)

When a deck has no usable label, a classifier compares its card list with lists of known archetypes. Tested on 12,450 decks that appear on both sites (the tcdecks label is the answer), the card list alone gets 94.1% right and the combined rule we use 94.3%. The most common mistakes are between close cousins (UW Control vs Landstill, Mono Black vs Pit Rack, Oath vs Oath Ponza). Unclassified decks are listed on their own archetype page.

Rogue decks

An archetype counts as rogue in a range when it has under 1% of the field there, or it is the catch-all “Rogue” archetype, or the deck has no archetype. The Rogue page lists their top-8 finishes at events with 32 or more players (16+ and 64+ as options), and the archetypes that were under 1% in the previous range of the same length and are growing (10+ decks now).

Changes and significance

Changes compare the selected range with the range of the same length just before it (or two ranges you pick on Trends). A change is shown in ▲ green or ▼ red only when a two-proportion z-test says it is unlikely to be chance (95% level); otherwise it is ■ grey. With about 6,500 decks in three months, a 5% archetype moving by 0.4 points is normal noise. Risers and fallers only include archetypes with at least 40 decks across both ranges.

Sample sizes are shown next to shares (decks) and next to card statistics (lists). Treat anything based on fewer than about 50 decks or 10 lists as anecdotal.

Top-8 conversion

Conversion = an archetype's share of top-8 slots ÷ its share of the field (1.0 = average). Because 94% of reported decks are already top-8 finishes, it stays close to 1.0 for almost every archetype and is not a measure of strength. We show it only in the full archetype table, with this caveat. It says a little more with an event-size filter, where full standings are more common.

Prices

Prices are shown in EUR (Cardmarket; latest: –); the Prices menu in the header switches between EUR and USD. They are two markets, not a conversion: reserved-list and old staples cost far more in the US while commons cost about the same, so each currency picks its own cheapest printing. Both are refreshed weekly with the rest of the data (the price cadence is configurable, separately from the deck refresh).

• EUR: the Cardmarket price guide. A printing's price is its Cardmarket trend (what copies actually sell for); a printing with no trend only counts when no printing has one, and then its lowest listing is used. We use the trend rather than the lowest listing because the lowest listing is often a damaged or foreign copy and made some competitive lists look like they cost under €10. The guide's own 30-day averages give the "prices to watch" from the first day.
• USD: TCGplayer market prices (sales-based, non-foil) as published daily by Scryfall. Scryfall describes its bulk prices as estimates, meant to track trends or give a general idea of a card's value, with no guarantee; that is how this site uses them. Scryfall carries no 30-day average, so the USD "prices to watch" compare the site's own weekly snapshots a month apart and appear once a month of them exists.
• Cheapest printing: the tournament-legal paper printing with the lowest price in the chosen currency. Every "price to build" on this site uses it: median list prices, cost drivers, cheapest top-8 lists.
• Original printing: the price of the card's first printing (the one whose image we show). Useful for collectors; never used for deck totals.

A list's price is Σ copies × price over main deck and sideboard. An archetype's price is the median of its lists in the range. The cost driver is the card with the largest part of the archetype's total cost (copies × price over all its lists). Basic lands and a few unmatched cards have no price; list prices that miss a card are shown with "≈". Trend views (the evolution table, What changed) leave basic lands out.

Filters

• Range: whole months, by event date. The default is the last 3 months with data.
• Platform: MTGTop8 tags MTGO events; its untagged events count as paper (webcam events count as paper).
• Event size: keeps events with at least N players. MTGTop8-only events have no player count, so they drop out when a size is set.
• Top 8: only decks that finished in the top 8. Since most decks already are, it changes less than you might expect.

All filters live in the URL, so any view can be shared or cited as a link.

Old School

The format switch in the header (Premodern | Old School) shows the same pages for Old School 93/94. Results come from tcdecks.net only (MTGTop8 has no Old School format), mostly each event's top 8. Archetypes are tcdecks' own labels. The default range is the last 12 months, because there are far fewer events.

Rule sets come from the event name: X-point (a 10-point budget on Atlantic rules), Eternal Central, Atlantic, Swedish, and “other” for local rule sets (LCOS, Molins, Madrid, Paladin…) or names that do not say. 7-Points Singleton events are left out, and so are events tcdecks lists twice (same date, same players).

No card lists. There are no Old School deck lists here. Card numbers come from its public “most played cards” page limited to the event's date: copies of each card across that event's listed decks (top 100 cards per board). Events that share their date with another Old School event have no card numbers. So there is no list over time, typical list, cost or budget for Old School, and card numbers are per deck on average, not per archetype.

X-points use the current X-point list (xpointoldschool.com) and only X-point events since January 2026. Points spent at an event = copies × points across its listed decks; divided by the decks listed, that is the average a deck spent of its 10 points.