One line for the whole market: the share of actively traded raw, near-mint English cards whose price today is above their own 30-day average. Above 50 means more of the market is rising than falling. The question this page answers is not which card won — it is how many of them are winning, and whether that number is confirming the market composite or quietly disagreeing with it.
-2.0 pts in a week · 35,633 actively traded
Between 40 and 60: no side is holding.
Market making new highs on fewer cards: composite +6.5%, score -5.3 pts since 2026-08-30.
Each reading is that day's own score, never an average of the days between.
Counted over the cards whose level actually changed between 2026-11-15 and the day before it. The cards that did not move are excluded from the median rather than dragging it to zero — which is what a median over the whole priced cohort reads, every day, on a guide that re-prices a card every few days.
Updated weekly, Sunday night by a pipeline step (last run 2026-11-16), never read live: the score's per-day, per-card 30-day lookback touches two years of candle history for the whole catalogue, and every aggregation runs in Postgres so the request-time read stays one small select. Every figure on the page is as of the build's evidence day, printed at the bottom — not as of the moment you loaded it.
A card counts only when it was priced on at least 20 of the last 30 days. The published line is filtered with a 3-day exponential average, because the raw line carries a measurable Monday step from the stretch where prices were written weekly. The full method below.
The share of actively traded raw, near-mint English cards whose price today is above their own 30-day average — 54.2% of 35,633 cards today. Above 50 means more of the market is rising than falling; the further from 50, the broader the move.
Because a guide price only moves every few days. Comparing today with one exact past day flips whole groups of cards in and out from one day to the next, and the line zig-zags. Comparing today with the average of a card’s own last month cannot: one re-price moves that average by a thirtieth.
A card priced on at least 20 of the last 30 days, with a TCGplayer listing or sale recorded somewhere in that window wherever those counts exist. A card priced twice a month has a "30-day average" that is really a two-point average, and its side of the line says nothing about the market.
$10+, $50+, $100+, $200+: the same score, recomputed over the cards at or above each level on the day. A market where the cheap half is rising and the $200+ half is not is a different market from one where both are, and only splitting it shows you which you are in.
Risk-on is the score holding above 60 for at least 5 days running; risk-off is below 40 for the same. Fixed thresholds, not percentiles of the chart's own window, so "risk-on" means the same thing in a quiet year as in a violent one. A single day over the line is not a regime.
Every priced RAW, near-mint, English card this platform tracks. For each of the last 730 calendar days, each card's own moving average is the mean of every level it carries in the trailing 30 calendar days, and the score is the share of the actively traded cohort whose level that day is strictly above it. The per-card level is the platform's standard precedence: TCGplayer's market guide, then the guide relay, then the midpoint of the sold low and high, then the sold median.
A card counts on a day only when it carries a level on at least 20 of that day's trailing 30, and when a TCGplayer listing count or sales count above zero appears somewhere in that window — for the stretch of history where those columns carry anything at all; where a card's candles never carry either column, it is judged on its price history alone rather than silently excluded. A day under 25 qualifying cards is marked thin, not hidden.
$10+, $50+, $100+, $200+: the identical statistic, restricted to cards whose level ON THAT DAY is at or above the threshold. The threshold is never applied to some historic price, so a card that crosses $50 joins that band from the day it crosses and leaves it if it falls back. All five lines are published; the chart draws the unrestricted one and offers the rest as legend pills.
Nothing before 2024-02-08 is averaged with anything after it. Before that day the platform's history is a monthly guide's ungraded summary and after it TCGplayer's daily near-mint market — two different sticks — so the moving average's lookback is clamped at the seam rather than quietly chaining across it.
Every build folds the raw line onto its weekdays and measures how far the best and worst weekday sit apart. This build's line is filtered with a 3-day exponential average: it carried a Monday step of about seven tenths of a point against every other day, from the stretch of history where prices were written weekly rather than daily, and that step is large enough to read as a market turn when it is not one. The stored payload always holds the RAW line and the verdict beside it, so the filter is applied once at the point of reading and never compounds across builds.
The shaded stretches are runs of at least 5 consecutive days with the score above 60 (risk-on) or below 40 (risk-off). The thresholds are absolute, so the shading is a statement about the market rather than about the window being plotted, and the five-day floor is why the chart no longer carries a stripe per day. A day with no score breaks a run rather than continuing it.
As of 2026-11-15 · built 2026-11-16 · Updated weekly, Sunday night
Measured over Nov 16, 2025 to Nov 15, 2026.
Change is against 30 days ago. Actively traded means a card priced on at least 20 of the last 30 days, with a TCGplayer listing or sale recorded in that window wherever those counts exist.
Market making new highs on fewer cards: composite +6.5%, score -5.3 pts since 2026-08-30. A divergence is the composite making a new 90-day high while the score is lower than at the last high — the advance being carried by fewer cards. Updated weekly, Sunday night; the last build ran 2026-11-16. 37,603 priced cards, 2026-05-20 through 2026-11-15.
The grey line is the platform's market composite — an unweighted, chain-linked average of whatever was priced on both of two consecutive published days, not an index of a selected basket. A bearish divergence is the composite setting a new 90-day high on a day when the score is LOWER than it was at the previous high in that window; a bullish one is the mirror at a new low. Only a signal within the last 30 days is shown, and an opposite-kind signal ends the run — that is the market answering, and the card must stop claiming the old direction.
Advancing and declining counts, and the median percent move, over the cards whose level actually CHANGED between the build's evidence day and the day before it. The median over the whole priced cohort is also stored, and is almost always exactly 0.0% — a fact about how often the guide re-prices a card, not a reading of the market, which is why it is no longer what this block prints.
The 7 days, 30 days, 90 days lines behind the legend's advanced pills are the page's earlier headline: the share of cards priced on BOTH a day and the day that many days earlier whose later price is strictly higher, an exact calendar-day match with no tolerance. They are kept because they are a real statistic and some readers want them — but they are jagged by construction, which is why they are no longer the picture the page leads with.
Up to 10 sets, ranked by the size of their own 7-day percent change — read straight off this platform's own Hot & Cold board rather than computed a second time, so the same set-level move never quietly disagrees between the two pages.
37603 priced cards, 2026-05-20 through 2026-11-15. Updated weekly, Sunday night, from the last COMPLETE candle day — never the day in progress, which carries a stub of prices until the writers have been round. A moving-average comparison is a lagging one: a market that turns hard turns this line a few days later, and a card that stops being priced leaves the cohort rather than registering a fall. This page never distinguishes condition beyond near-mint, and never splits by grading company, printing or language beyond English.
Research classification generated from observable market data with stated confidence and assumptions. Not personalized financial advice.