Cheaper over there, after everything.
Every deals feed in this hobby compares a listing to a price in the same market — an eBay ask against an eBay sold price. This one does not. A row here is a live asking price on one market sitting below what the same card, in the same grade, actually sold for on another — after freight, the destination's import tariff, the currency conversion and the selling venue's fee.
We looked at thirty-five competing tools. Not one of them prices a card net of fees, FX and import tax; the closest tells its users to work those out themselves. That arithmetic is the whole page, and it is printed line by line on every row — because a spread with the costs folded away is not a deal, it is a number that looks like one.
Which way the money runs
Each row is priced on its own corridor, not on a global average: freight from Japan to the US is not freight from the UK to Germany, and the tariff into the United States is not the tariff into Japan. Every one is the arbitrage engine's published figure for that direction.
How big they are, after costs
Every qualifying row on the feed, bucketed by what is left after the crossing — not before it. This is the shape of the whole feed, including the rows locked below.
| Net spread | Rows | Median cost of crossing | Best net in band |
|---|---|---|---|
| under 10% | 32 | $24.99 | $93.23 |
| 10–25% | 31 | $24.25 | $578 |
| 25–50% | 26 | $32.94 | $569 |
| 50–100% | 19 | $36.02 | $1,031 |
| 100%+ | 2 | $220 | $1,009 |
The median row leaves $13.00 (20.8%) after paying $27.51 to move the card. 6,608 further ask-and-comp pairings were cheaper on one side and still did not make it — freight, tariff, conversion and the selling fee ate the whole difference. Those are exactly the rows a gross-spread feed would have published.
Worked end to end
Three of the feed, in full: the live ask with its venue and link, the median of the recorded sales in the other market with every one of those sales behind it, and the arithmetic from the asking price to the money that is actually left.
🇩🇪 eBay DE · seller martinbehrendt99 · Fixed price, offers accepted
🌐 PokéPriceTracker ×5
20 Jun 2026 – 16 Sept 2026
🇺🇸 eBay (US) · seller ks-tradewise · Fixed price
🇯🇵 Yahoo! Japan Auctions ×3
16 Aug 2026
🇩🇪 eBay DE · seller gianluccavall-2 · Fixed price
🌐 PokéPriceTracker ×9
5 Nov 2025 – 30 Aug 2026
Show all 9 recorded sales behind that median
| Date | Venue | Price | Proof |
|---|---|---|---|
| 30 Aug 2026 | 🌐 PokéPriceTracker | $4,370 | listing → |
| 10 Aug 2026 | 🌐 PokéPriceTracker | $4,000 | listing → |
| 8 Aug 2026 | 🌐 PokéPriceTracker | $2,500 | listing → |
| 30 Jul 2026 | 🌐 PokéPriceTracker | $4,999 | listing → |
| 10 Jul 2026 | 🌐 PokéPriceTracker | $520 | listing → |
| 29 Jun 2026 | 🌐 PokéPriceTracker | $1,948 | listing → |
| 19 Jan 2026 | 🌐 PokéPriceTracker | $3,500 | listing → |
| 17 Jan 2026 | 🌐 PokéPriceTracker | $3,000 | listing → |
| 5 Nov 2025 | 🌐 PokéPriceTracker | $4,177 | listing → |
The feed
The ranked list is the part that costs money: which listing, on which venue, at which price, with the link straight to the ask.
107 more live asks — ranked, with the listings linked
You have just seen the count, the corridors, the size distribution and three of them worked from the asking price to the money left over. The rest of the feed is the same thing for every remaining row.
- All 110 rows ranked by net spread, with the listing linked
- Every recorded sale behind each comp — dated, priced, linked
- Spread alerts: fire on the cross-market spread net of costs, not on an absolute price
- Real sold prices, Japan and the West, on every card page
Alerts that fire on the spread
Every price alert in this hobby watches one number in one market: tell me when this card drops below $X. That fires on a market-wide move, which is the moment everybody else gets the same notification. A Hoshino spread alert fires on the relationship between two markets, net of costs — when a live ask somewhere lands far enough below the sold comp somewhere else that the difference survives the crossing. It cannot fire on an absolute price move, and it will not fire against a price-guide point.
The two rules people actually want
- On one card
- “Tell me when this card's Japanese price falls more than 20% below its Western sold comp, net of costs.” Set the corridor and the threshold; it fires on the net spread, once per listing, ever.
- Across a watchlist
- “Tell me when anything I am watching becomes buyable in Asia below the Western floor.” One rule covers every card on the watchlist, and follows it as it changes.
Alerts land in the app, on your phone, and in the daily digest — the same delivery every other Hoshino alert uses. Set one from the Alert Center in the app.
How this is computed
The rules, and what each word means
- A live ask
- An active listing we hold, priced, seen inside the last 7 days, and — if it is an auction — one that has not already closed. Nobody has paid this price. It is what a seller wants, and it can vanish between our last look and yours.
- A real recorded sale
- The other side of every row is made of them: a transaction we recorded, at a price, on a venue, on a date. A price-guide point — TCGplayer market, Cardmarket trend — is not a sale and is excluded; we hold far more of those than real ones, which is exactly why. A listing-lifecycle inference is not a sale either. Sales our integrity filter flagged as shill, wash or outlier are dropped before anything is averaged.
- Two different markets
- The ask and the comp must be in different markets. An ask below a sold price in the same market is a good buy, and every other tool already shows you those.
- One grade tier
- The listing and every sale behind the comp share one grader, one grade, one print run and one raw condition. A PSA 10 is never priced against a raw copy; a 1st Edition never against an Unlimited. Where we cannot confirm the tier, the row is dropped rather than assumed.
- Net of costs
- Tracked insured shipping for that corridor, the destination's import tax — 15% into the United States since the $800 de-minimis exemption ended in August 2025, 10% into Japan, 20% into the EU — the currency conversion a real buyer pays off mid-market, and the selling venue's own fee on the sale. Every one is itemised on every row. A row whose difference does not survive all four does not appear.
- 3 sales behind the comp, minimum
- Two prices are not a market. Below 3 recorded sales the row is absent — not shown with a warning, absent. The comp must also be recent: its median sale date within 90 days of now, because the ask is live and a comp from last season measures time rather than distance.
- Not advice
- A row is an observation about one asking price and a set of prices people already paid. Hoshino does not forecast, does not tell anyone what to buy, and does not promise any spread here can be captured. Shipping a card across a border carries real risk of loss, damage, customs delay, misdescription and grading disputes that no price difference accounts for.
Of 207,813 live asks on cards that hold a usable sold side, 77,104 had no comp in their own grade tier at all, 32,981 had a comp under the sales bar, 62,199 had one only in their own market, 4,884 had a comp too widely spread to be a single price, 2,902 were further apart than one printing can be, 1,708 rested on a comp too old to describe today, and 6,608 simply did not survive the costs. We publish those numbers because a count you cannot interrogate is a marketing figure.
Read the full standard on what each price on Hoshino means, and see the Japan–West sold-against-sold comparison on the Edge board.
Recomputed from the live database. Last run 19 Sept 2026.
The only price that matters is the one after costs.
Watch a card and we will tell you when it becomes cheaper somewhere else — net of everything.