# OP-Atlas Comparison — Supporting Data Companion to "BRING IT ON — OP-ATLAS DECLARES WAR ON THE WORLD GOVERNMENT". Everything below is dated and linked so it can be re-checked or disproved. Fieldwork 18–21 August 2026 (Asia/Bangkok). --- # APPENDIX A — How we scored ## The evidence layers We never collapse these three, for anyone, including ourselves. | Layer | Meaning | |---|---| | **Claimed** | The company's own marketing wording | | **Displayed** | What is actually visible in the product right now | | **Verified** | What we independently confirmed, with the method stated | ## Confidence — published next to every score | | | |---|---| | **High** | Verified: database query, reproducible test, or feature used first-hand | | **Medium** | Displayed: observed live in the product, not independently reproducible | | **Low** | Claimed only, or the evidence is behind a paywall, login or bot-block | | **Attested** | Our team's first-hand use of a paid product a reader cannot check without buying it | | **TBD** | We could not verify the category. We publish TBD rather than guess. | A High 7 outranks a Low 9. If you want to argue with a score, argue with the layer under it. ## The dimensions, and what a score means A 7 in Mobile is not a 7 in Data Coverage. Each dimension has its own scale. - **Data coverage** — cards represented, and whether prints, languages and variants are resolved as distinct. 10 = whole catalogue, every print language, variants distinct, counting method published · 7 = broad with real gaps · 4 = a useful subset · 2 = narrow slice. - **Market coverage** — independent markets, weighted by directness: direct feed > sold data > listing data > third-party dataset > link-out > claimed only. 10 = 6+ direct markets including genuine sold data, spanning Japan and the West · 7 = 3–5 direct · 4 = 2 markets, or many but all relayed second-hand · 2 = a single retailer. - **Price intelligence** — whether the number's construction is knowable. 10 = method published, per-card source count, freshness and confidence exposed, sold distinguished from asking · 7 = method public but incomplete · 4 = sources named, metric type undefined · 2 = a bare number. - **Market exploration** — whether you can interrogate the market rather than look up one card. 10 = segment and drill across sets, characters, rarities, graded and sealed from a question · 7 = several curated market views · 4 = movers and rankings only · 2 = search only. - **Collection intelligence** — six capabilities, scored separately, then averaged and penalised for absences: total value · cost-basis profit and loss · concentration · movement attribution · value drivers · anomaly detection. - **Sold data** — real completed transactions and how visible they are. 10 = deep, dated, visible · 7 = real sold data, partial depth or gated · 4 = present but shallow or indirect · 2 = asking prices shown near sold data without distinction. - **Grading** — graders covered, population data, grade-specific pricing, and stated provenance. - **Scanning** — 10 = AI bulk scanning verified · 7 = single-card scanning · 5 = graded-slab scanning only. - **Mobile** — 10 = mature native apps on both platforms at volume · 5 = installable web app · 3 = mobile web. - **Accessibility** — localisation reach: interface languages, currencies, card-text localisation, regional fit. *Deliberately excludes whether prices are free — that is a pricing fact, reported in prose, not a quality score.* - **Portability** — 10 = import and export, lossless, variant-resolving, ungated. A product you cannot leave scores at most 4. ## Rules we bound ourselves to 1. If OP-Atlas does not clearly win on evidence, OP-Atlas does not win. 2. No product is scored in a dimension where we lack meaningful evidence. TBD is a result. 3. If underlying data exists but is not exposed to users, it is not a product feature — for them or for us. 4. Roadmap does not count. Only what ships today. 5. Scores were assigned centrally from collected evidence, so one scale governs every product. --- # APPENDIX B — The scorecard Scores as of 21 August 2026. Format: score · confidence (H/M/L/A). | | OP-Atlas | Collectr | Card Ladder | OP.TCG | OP Collector | Card2Price | OnePieceCardValue | Data Card Game | |---|---|---|---|---|---|---|---|---| | Data coverage | **9.0 H** | TBD | TBD | 6.5 L | 5.5 M | 5.0 M | 3.0 M | TBD | | Market coverage | **8.5 H** | 4.5 M | 6.0 M | 6.0 L | 6.5 H | 3.5 M | 4.0 M | 4.0 M | | Price intelligence | **8.5 H** | 4.0 L | 8.0 L | 4.0 L | 4.5 H | 3.5 M | 6.0 M | 5.5 M | | Market exploration | **9.0 H** | TBD | 7.0 M | TBD | 6.0 H | 4.0 M | 2.0 M | TBD | | Collection intelligence | 7.5 H | 7.0 H | TBD | 6.0 A | 6.5 H | 2.0 M | 0 H | TBD | | Sold data | 6.0 H | 5.0 M | **9.0 M** | 6.0 A | 2.5 H | 7.0 H | 3.0 M | 3.0 M | | Grading | 7.0 H | 5.5 M | **9.0 M** | 4.5 L | 0 H | 5.0 M | 6.0 M | 3.0 M | | Scanning | **0 H** | 8.5 M | 5.5 M | **9.0 M** | 0 H | 0 H | 0 H | TBD | | Mobile | 5.5 H | **9.0 M** | 8.5 M | 8.5 M | 4.0 H | 3.5 H | 3.0 M | 3.0 M | | Accessibility | **9.5 H** | 4.0 L | 2.0 M | 6.0 M | 3.5 H | 4.0 H | 2.5 M | 4.5 M | | Portability | **8.5 H** | TBD | 7.0 M | 1.0 M | 1.0 H | TBD | 0 M | TBD | ## Notes that travel with the scores **Where OP-Atlas loses outright.** Scanning: we score 0 — we do not have one. Mobile: we are an installable web app, not a native application. Sold data and grading: Card Ladder is ahead. **Card Ladder.** Their One Piece **index** publishes "Total Cards Tracked: 124" — an index basket of graded cards from 2022–2025 sets, not a catalogue. Their wider One Piece ladder renders 20 rows behind "Upgrade to Card Ladder PRO to gain access to all cards and sales data", and no public total exists, so we score their One Piece data coverage TBD rather than guess. Free chart ranges are 2 weeks, 1 month and 3 months; 6 months, 1 year, 2 years and all-time are Pro features. We could independently verify One Piece sale records back to 24 January 2026. Their "every public sale going back to 2000" is a claim about sports cards; the One Piece Card Game launched in 2022, so it cannot apply here. Their scanner is real and current (v4.2.1, 20 August 2026) but scans **graded slabs**, not raw cards. We did not buy Pro, so several of their cells stay TBD. **OP.TCG.** Prices sit behind their paid tier, so most of what we could examine is marketing copy and store listings. Their collection intelligence and sold-data scores are **attested by our own team's use of the app**, not independently reproducible — labelled accordingly. Two figures we can only report as published: their site and Google Play state "20,000+ references" while their App Store listing states "over 10,000 references" (both observed 19 August 2026). We are not going to decide for them what either number counts. **OP Collector.** The quiet surprise of this comparison, and the product that broke a claim we had intended to make. Free, no account, it ships both concentration and movement attribution: "Top Characters — By total portfolio value" and "Top Gainers — Biggest 7-day price rises in your portfolio". It carries no sold data of its own, but relays TCGplayer Market Price and Cardmarket Trend, both of which are sold-derived — which is why it scores 2.5 rather than 0 on sold data. **Collectr.** Ships movement attribution on its free tier: "Trending Today — The cards in your collection moving most, up or down, today." **What nobody has — including us.** Across every product we examined, we found no evidence of **anomaly detection** at collection level. The words "anomaly", "driver", "attribution", "concentration" and "exposure" appear nowhere in the marketing or product copy of the four largest products in this comparison. **OP-Atlas does not claim anomaly detection as a shipped collection feature either** — it is one of the six capabilities in this rubric that we score ourselves as absent. Five of the six are verified in our production code; that sixth one is not. --- # APPENDIX C — OP-Atlas's own numbers Same treatment we applied to everyone else. Every figure below comes from our production database, queried 20–21 August 2026. ## Catalogue | Measure | Count | |---|---| | Cards carrying a live Atlas Estimate (bake of 20 Aug 2026) | **11,247** | | — Japanese prints | 4,528 | | — English prints | 6,719 | | — variant subprints (parallels, alternate arts, reprints) | 2,998 | | — promotional cards | 409 | | Total price observations recorded | **2,309,294** | **How we count.** One card = one print in one language. The Japanese and English printings of the same card number are two cards. Each parallel, alternate art and reprint subprint is its own card. We publish this because a card count means nothing without it. **The comparison that proves the point.** An independent One Piece database, optcg.one, publishes "3,047 cards across all sets" (observed 21 August 2026). We publish 11,247. **Neither of us is lying.** They appear to be counting released card numbers; we count every print, language and variant separately. The same catalogue is honestly 3,047 or 11,247 depending only on the counting rule — which is exactly why "20,000+ cards" or "9,000+ cards" or "10,000+ cards", published without a counting rule, tells you almost nothing. That includes ours. ## History — and its limits | Series | Earliest record | Recorded days | |---|---|---| | Raw source prices (Cardmarket, TCGplayer, PriceCharting, yuyu-tei) | 30 April 2026 | — | | The Atlas Estimate series itself | **9 May 2026** | **56 days recorded** | | eBay sold | **16 July 2026** | 18 days, 3,110 cards | That is months, not years, and the recorded days are not continuous. We are not going to call it anything else. Publicly, every card page shows 30 days of history today. **From 1 September, free accounts see 7 days and Pro sees all-time.** We are telling you before it changes, not after. ## Market sources Atlas currently draws from seven live market sources: TCGplayer (daily) · Cardmarket (weekly) · CardTrader (daily) · PriceCharting (twice weekly) · yuyu-tei (weekly) · SNKRDunk (every 5 days) · eBay Sold (weekly). Yahoo! Auctions data is collected but not yet surfaced, so we do not count it. **A disclosure we owe you, by our own standard.** PriceCharting is itself an aggregate built on eBay sold data. So two of our seven sources are not fully independent of each other, and our estimate engine de-weights that pair deliberately. If we are going to ask whether someone else's "market" is direct data or a third party's dataset, we have to answer it about ourselves first. Likewise: our PSA, BGS, CGC and SGC graded prices are sourced **via PriceCharting**, not from the graders. Our first-party graded data comes from SNKRDunk and from eBay sold records. The freshness stamp on a card page describes when the estimate was last drawn, not when every underlying market last moved — several of those sources refresh weekly. ## Data decisions | Measure | Count | |---|---| | Data decisions logged | **31,251** | | Applied | 25,649 | | **Rejected** | **5,602** | | Distinct cards touched | 13,695 | **We are not calling this community-verified.** Eight accounts have ever filed a decision, and one of them accounts for 99.87% of them — our own, including a substantial amount of manual review alongside automated matching. The honest description is that these are our decisions, they are auditable, and roughly one in six gets rejected. ## Accessibility Interface in 8 languages (English, ไทย, Deutsch, 日本語, 简体中文, Bahasa Melayu, Español, Tiếng Việt) and 8 currencies (USD, EUR, GBP, JPY, CNY, THB, MYR, VND). **Card ability text is translated into Thai and English only** — a reader using the Spanish or Vietnamese interface still sees English card text. ## Portability Spreadsheet import (.xlsx, .csv, .tsv or pasted rows) with per-row print resolution: rows with one candidate resolve automatically, rows with several let you pick the exact print, and unmatched rows are reported rather than silently dropped. Export is one click. Neither is gated. **Portability caveat.** Import, print resolution, export and the absence of tier gating were all verified. A full export → reimport round-trip was **not** independently tested in this pass, so we do not claim lossless round-trip behaviour. --- # APPENDIX D — Sources All captured 18–21 August 2026. Links are live; check them yourself. **Products compared** — [OP.TCG](https://optcg.app) · [Collectr](https://getcollectr.com) · [Card Ladder](https://cardladder.com) ([One Piece index](https://cardladder.com/indexes/onepiece), [One Piece guide](https://www.cardladder.com/ladder?category=OnePiece)) · [OP Collector](https://opcollector.com) ([sets](https://opcollector.com/sets/)) · [Card2Price](https://card2price.com) · [OnePieceCardValue](https://onepiececardvalue.com) · [Data Card Game](https://data-cardgame.com) · [RakaCard](https://rakacard.com) · [SlabSiam](https://slabsiam.com) · [TCGCardPrice](https://tcgcardprice.com) · [TCG.IN.TH](https://tcg.in.th) · [Card Radar](https://cardradar.gg) · [Mugi TCG](https://mugitcg.com) · [Haki TCG](https://hakitcg.com) · [Kaizoku](https://kaizoku.app) · [Logia](https://app.logiatcg.com) · [CardExplorer](https://card-explorer.com) · [OP.LOG](https://op-log.gg) · [OneJum](https://onejum.com) (pack-opening, not a price tracker) · [optcg.one](https://www.optcg.one/) · [Limitless](https://onepiece.limitlesstcg.com/cards) **Market sources** — [TCGplayer price guides](https://www.tcgplayer.com/categories/trading-and-collectible-card-games/one-piece-card-game/price-guides) and their [Market Price definition](https://help.tcgplayer.com/hc/en-us/articles/213588017-TCGplayer-Market-Price) · [Cardmarket](https://www.cardmarket.com/en/OnePiece) · [CardTrader](https://www.cardtrader.com/en/one-piece) and their [price definition](https://static.cardtrader.com/en/pages/autopricer) · [PriceCharting](https://www.pricecharting.com/category/one-piece-cards) and their [methodology](https://www.pricecharting.com/page/methodology) · [yuyu-tei](https://yuyu-tei.jp/top/opc) · [CardRush](https://www.cardrush-op.jp/) · [SNKRDUNK](https://snkrdunk.com/en/brands/onepiece/trading-cards) · [eBay Marketplace Insights](https://edp.ebay.com/api-docs/buy/marketplace-insights/static/overview.html) **Our own surfaces** — [op-atlas.app](https://op-atlas.app) · [data freshness](https://op-atlas.app/data-freshness) · [pricing](https://op-atlas.app/pro) · [patch notes](https://op-atlas.app/patch-notes) · [leaderboard](https://op-atlas.app/leaderboard) · [terms](https://op-atlas.app/terms) --- # APPENDIX E — What we could not verify Published because a comparison that hides its gaps is advertising. 1. **Anything behind a paid account we chose not to buy.** Card Ladder Pro ($19.99/month) gates their collection tools, sales dates and all chart ranges beyond 3 months. OP.TCG gates prices. Those cells are TBD or attested, never guessed. 2. **Sample size.** Our hands-on price and interface testing used a 10-card benchmark fixed in advance, across four products, tested logged out on 18–19 August 2026. Indicative, not statistically conclusive. 3. **Logged-out testing under-represents every product with value behind sign-in — including ours.** Our own eBay sold values, per-source labels and longer history are all behind a free account and were not counted in that pass. 4. **One reference marketplace.** Cross-checking prices against source marketplaces got as far as TCGplayer before we ran out of time. Cardmarket, PriceCharting, yuyu-tei and eBay were not captured. 5. **Update speed.** Measuring how quickly each product reprices needs repeated observation over days. We did not do it, so we make no claim about anyone's freshness — including our own beyond what our stamps say. 6. **Sealed products.** Not tested, for anyone. 7. **No screenshots.** Every competitor quote here is text captured from a live page with its URL and date. Pages change; that is why the dates are on everything. **If we got your product wrong, tell us.** Send the card, the source, the method and the number to hello@op-atlas.app and we will re-check and correct it publicly. A comparison that cannot change its mind is advertising.