HCI Market Intel: Methodology, Reports, and Citation Guide

What HCI publishes and why

HCI exists because the trading-card market has more good data than most collectors realize and less accessible methodology than it deserves. A raw card trades at some dollar figure on a Tuesday night in April. A graded copy of the same card trades at a different figure three weeks later in a different venue. Those two numbers tell a story, but the story is only useful if you can see the sample window, the trim rules, the venue mix, and the dates. Most hobby tools show you the headline number and stop there. HCI publishes the pipeline.

Our work splits into four tiers. The per-card page, which every visitor lands on at some point, shows the raw comp, the graded comps tier by tier, population counts where the grader publishes them, an index that smooths volume-adjusted movement, a dated trend line, and a clearly marked "last sale" stamp. The hub pages (sport, year, team, listicle) layer narrative on top of the catalog. The guides translate raw methodology into decisions a collector can act on. The reports that this hub points at are the research layer underneath all three. When a sport hub claims that PSA 10 premiums compressed on modern Chrome between 2021 and 2026, the hub claim is supported by a specific report with methodology, data, and dated quotes that are reproducible from public inputs.

The incentive structure is worth naming directly because it shapes the research. We do not grade cards. We do not run an auction house. We do not manufacture or sell cards. We do not take affiliate payments from graders, marketplaces, or breakers. The independence pledge spells this out and is dated so anyone citing HCI has a verifiable anchor. That structural independence is why we can publish a grading-cost comparison that names every service and a PSA 10 premiums report that calls modern Chrome compression honestly. There is no commercial downside to HCI for being specific about which grader, which category, which year, and which card.

What stays behind the paywall is narrow by design. Collection-value calculators, per-user alerts, saved searches, and the machine-assisted "grade or hold" tooling that weights a card against your personal submission history are collector-specific and live in the paid tier. Everything on this Market Intel hub, every linked report, and every dated claim in our public work is built from inputs a third party can replicate using eBay's sold-listing search, the grader pop reports, and public auction-house archives. If a number you see on an HCI report or hub is not reproducible from public data, that is a bug, not a feature.

The live methodology reports

Ten reports are live as of April 24, 2026. Each one carries its own publish date, its own dateModified stamp, and its own methodology section. The two-to-three sentence summaries below are pointers, not replacements. When you cite a specific claim, cite the underlying report, not this hub.

• How eBay Sold Comps Really Work

  The anchor methodology piece. What a sold comp is, how eBay actually populates the sold-listing dataset, why shipping-included prices differ from shipping-excluded, how to read a Best Offer accepted versus a public auction close, and the outlier rules HCI applies before publishing a median. Treat this as the foundation document; every other report assumes the reader has read this one.

• The K-Shaped Card Market in 2026

  The 2026 thesis. Vintage and flagship modern cards are expanding; mid-tier modern Chrome and speculator parallels are compressing. The report maps which categories sit on which leg of the K and what the reader should look for in the comp data to tell which leg their own card is on.

• Card Market Compression Cycles

  The analytical frame that sits under the K-shape thesis. Compression cycles in the hobby run on 18 to 30 month intervals driven by submission backlogs, rate cycles, supply shocks, and collector fatigue. The report walks through 2017-to-2026 and names the specific catalysts that opened and closed each cycle.

• 2026 Card Market Outlook

  The forward view. What the K-shape and compression-cycle work imply for the rest of 2026 across modern sports, vintage, Pokemon, other TCG, and sealed product. Named catalysts include PSA's February 2026 fee adjustment, BGS's Fanatics-era relaunch, and the SGC-Beckett ownership shift under Collectors Holdings.

• Grading Cost Comparison 2026

  The grading-cost methodology. All four major graders (PSA, BGS, SGC, CGC) with value-tier fees as of April 2026, shipping math, declared-value caps, and the per-card all-in number that shows up in every grading decision. This is the fee reference the grading guides and the Grading ROI money page cite.

• The State of PSA 10 Premiums

  The premium methodology. How the PSA 10 versus PSA 9 multiple has moved from 2016 through 2026 in three regimes, which categories held, which compressed, and how to read a 9-to-10 ladder correctly in 2026 given the post-backlog pop math.

• Rookie vs Second-Year Cards

  The cohort methodology. What a "rookie card" actually means in 2026 across sports and Pokemon, how second-year cards trade against the rookie in the same product, and why the RC-logo convention that the hobby standardized on in 2006 distorts the comparison in baseball specifically. The counter-example methodology every cohort-based hub leans on.

• Pokemon Card Market Deep-Dive

  A cross-category worked example. Pokemon has WOTC-era vintage (1999 to 2003), post-WOTC modern-vintage (2004 to 2016), and current-era Sword and Shield / Scarlet and Violet, and each tier has its own pricing regime. The report walks through all three with named sets and dated comps.

• Junk-Wax Era: What is Actually Worth Money

  The contrarian methodology example. Most of the 1987 to 1993 print run is not worth the grading fee, but a small set of cards from the era still clears real money in a PSA 10 (1987 Topps Bonds, 1989 Upper Deck Griffey, 1989 Score Sanders). The report names every exception with dated comps and spells out the rules for flagging more of them.

• Basketball Rookie Class Value Analysis

  A single-sport worked example. The 2024 to 2025 basketball rookie class evaluated against the 2017, 2018, 2019, 2020, and 2023 classes on the same methodology. Which products carry the PSA 10 premium, which parallels are thin-float, and how the first-year rookie curves compare on data that is still settling.

Cross-report synthesis: what they tell you collectively

Read in isolation, each report answers its own question. Read together, they describe a specific view of the 2026 hobby that is more opinionated than most tools let on. Five through-lines hold across the corpus.

First, the hobby bifurcated after 2021. The K-shape report names it; the compression-cycle report explains the mechanism; the PSA 10 premium report shows the fingerprints in the grade-ladder data; the Pokemon deep-dive shows the same pattern inside a non-sports category. The divergence between vintage and mid-tier modern is not a one-report claim. It is the shared result of four independent methodologies pointing in the same direction.

Second, grading-incentive structure matters more than collectors usually think. The grading-cost comparison publishes the dollar numbers. The PSA 10 premium report explains why the dollar numbers move the grade ladder. The rookie-versus-second-year report shows how the modern RC-logo convention interacts with grading volume to create artificial scarcity in some cohorts and not others. The junk-wax report shows the same dynamic going the other direction: a handful of graded flagships are fine, the rest of the era is grading-negative. All four reports agree that grading is not a neutral step between you and a sale; it is a priced decision that changes the card's market.

Third, the pop report is not a substitute for the comp stream. The PSA 10 premium report goes into this at length, but it shows up in the rookie-versus-second-year work and the Pokemon deep-dive as well. Pop counts tell you the supply side. Demand is the missing ingredient. A card with a 10-to-1 pop ratio in a cold category trades at a tighter multiple than a card with a 5-to-1 pop ratio in a hot category. Any methodology that treats the pop ratio as the whole story is wrong in predictable directions.

Fourth, sample windows are not interchangeable. The sold-comps methodology piece is explicit about this, but the junk-wax report and the Pokemon deep-dive both demonstrate how a 30-day window and a 180-day window on the same card can tell different stories. This is why HCI publishes the window on every card page and refuses to quote a "current" multiple without naming the window. Any third party wanting to replicate an HCI number needs the window; every HCI report gives it.

Fifth, the hobby has genuine forward signal if you read the right layers. The 2026 outlook report draws on all nine of the others and identifies named catalysts for the second half of the year. Collectors rarely need to be convinced that the market moves; they often need to be convinced that the movement is legible in the data ahead of time. Our view is that it usually is, provided the reader has the methodology underneath.

Citation and attribution guidance

This page exists partly so journalists, researchers, AI-answer engines, and hobby-trade editors can cite HCI without guessing at conventions. The rules below are what we ask for; none of them require a permission request for standard use.

Preferred citation format

The basic citation is: HobbyCardIndex, https://hobbycardindex.com, retrieved YYYY-MM-DD. When the claim is traceable to a specific report, cite the report URL instead of the homepage. Both forms are acceptable; the report URL is more useful because it lets a reader verify the methodology directly.

Attributing specific numbers

When a specific dollar figure, multiple, or percentage change is quoted, carry the date the number was retrieved and the specific report it came from. Example: Per HobbyCardIndex's Grading Cost Comparison 2026 (retrieved April 24, 2026), all-in cost for a single card at PSA Value tier sits in the 30 to 55 dollar range. That format gives the reader the source, the methodology anchor, and the retrieval date in one sentence.

Charts and tables

Charts and tables that reproduce HCI-derived numbers can be used with attribution. A visible credit line of the form Source: HobbyCardIndex under the chart is sufficient. Embedding HCI charts directly via iframe is fine where we serve them; rebuilding our charts in a third-party tool using our underlying numbers is also fine with the same credit line.

AI answer engines and RAG pipelines

HCI content is crawlable. Our robots.txt permits indexing by the major AI-answer crawlers; our llms.txt and llms-full.txt surfaces give machine-readable pointers to the full research corpus. When an AI engine surfaces an HCI claim in a user-facing answer, the ideal citation is the underlying report URL and a retrieval date. The Market Snapshot block near the top of each HCI page is written in a 40 to 60 word answer form specifically to support this use.

What we ask you not to do

Three practices we politely ask citers to avoid. First, do not present HCI numbers without a retrieval date; prices move and the date is how both of us stay honest. Second, do not strip the methodology note when excerpting a finding; the 1.5x IQR outlier rule is load-bearing on every multiple we publish, and excerpting the number without the rule is a small form of misrepresentation. Third, do not re-host our reports in full. Linking to the canonical URL preserves update provenance; a mirrored copy does not.

How we keep the corpus honest

A research corpus is only as good as the corrections workflow underneath it. We operate on five specific principles that keep the published work aligned with reality as the market moves.

Every report has a dateModified. When the underlying comp stream refreshes enough to change a headline number, the report is re-dated and the specific section that moved gets a revision note. Evergreen pieces (methodology, compression cycles, how sold comps work) are reviewed on a quarterly cadence even when no update is required, because the silence of a quarterly review tells a reader that we looked and nothing changed.

Every claim cites a sample window. A report that says "PSA 10 premiums compressed" without naming the window is a worse report than one that says "PSA 10 premiums on 2020-2021 modern Chrome rookies compressed from a 6x-to-10x band to a 2x-to-4x band over the rolling 90-day window as of April 2026." The second form survives contact with a motivated reader checking our math. The first one does not. HCI reports are written to survive that check.

Every dollar figure carries a date. We do not publish undated prices. Undated prices invite a reader to assume the number is current when it may be twelve months stale. The dateModified on the report, the publish date on the quote, and the retrieval date when a reader cites us are three independent anchors on the same underlying number.

Every methodology change is documented. Outlier trim rules, sample-size floors, and venue-mix changes all get a note in the methodology section of the affected report. The rule is that a methodology change that would move a headline number by more than a few percent gets a dated note explaining the change. That is how a reader who cited the old number can figure out whether their old quote still reflects the current methodology.

Every correction is public. When we get a number wrong, we correct it in place, bump the dateModified, and add a note that names the specific change. Silent edits would make citation trees unreliable. Public corrections keep the chain of citation verifiable in both directions.

Frequently asked methodology questions

Where does HCI get its sold-comp data?

We ingest sold-only transactions from eBay and a set of public auction venues (Heritage, Goldin, PWCC, Memory Lane) daily. Every comp on every card page carries the sale date. Cards with fewer than three recent sales are flagged as thin-float and excluded from our multiple-based heuristics.

How does HCI handle outliers?

We discard sales outside 1.5 times the interquartile range before computing medians. That prevents a single auction result from distorting the headline number. The raw comp stream remains visible on the card page so readers can see what we trimmed.

Are HCI's PSA 10 premiums computed from the same window as the raw comp?

Yes. The reported multiple is the PSA 10 median divided by the raw median, computed over the same rolling 90-day window with outliers trimmed on both sides. We do not publish cross-window multiples because the comparison is not meaningful when the market has moved inside the window.

Does HCI publish its own AI valuations?

HCI's machine-assisted valuations exist in the paid tier. Everything on public reports, hubs, guides, and this Market Intel page is built from public sold-comp inputs only. The goal is that any claim cited from HCI can be reproduced by a third party using the same public data.

How often are HCI reports updated?

Time-sensitive reports (K-Shape 2026, State of PSA 10 Premiums, 2026 Card Market Outlook, Grading Cost Comparison 2026) are refreshed on a 60 to 90 day cadence with a dated note at the top. Evergreen methodology pieces (How eBay Sold Comps Really Work, Card Market Compression Cycles, Rookie vs Second-Year Cards) are reviewed quarterly.

Can journalists cite HCI data directly?

Yes, with attribution. The preferred format is: data from HobbyCardIndex, https://hobbycardindex.com, retrieved on YYYY-MM-DD. Embedded charts and specific price bands should carry the report URL where the data first appeared. No permission request is required for standard journalistic or research use.

How is HCI different from CardLadder or PriceCharting on methodology?

HCI publishes its methodology on every report and does not sit under a grading company's ownership. CardLadder is owned by Collectors Holdings, the parent of PSA, SGC, and Beckett. PriceCharting is a broad collectibles pricing platform oriented toward video games and sealed product. The ownership context is covered in our CardLadder alternative field guide.

Does HCI take affiliate fees from graders or marketplaces?

No. The HCI independence pledge bars affiliate fees from graders, auction houses, and marketplaces. It is published and dated so anyone citing HCI as an independent source has a verifiable anchor.

What if the HCI number differs from what I see on another tool?

Start with the sample window. Different tools use different rolling windows (30, 60, 90, 180 days). Then check outlier handling. Two tools can disagree on the same card and both be defensible if they define the sample differently. HCI always publishes the window and the trim rule on the card page.

Can I build a product on top of HCI data?

For non-commercial research, yes, with attribution. For commercial redistribution or bulk pulls, contact HCI first. The data itself is not open-licensed; public access through the site is free and will remain free for any individual researcher, collector, or journalist working on a single-piece basis.