How Long Do You Have to Hold a Pokemon Card?

Buying a card is easy. Getting your money back out costs about 13% between selling fees and shipping, which means a card has to rise 14.94% before a round trip breaks even. We measured how long that takes across every hold window we hold, to 2026-09-22, on a population that has to be named before its numbers are read: the 2,766 cards that stayed continuously quoted for twenty months, 13.92% of the tape, through a window that was a sustained rising market. On those survivors, each start week against each horizon, 254,082 windows at 30 days and 91,986 at a year. The median 30-day hold loses 11.27% after costs. The median 90-day hold loses 6.26%. The median hold first clears the cost floor somewhere between 150 and 155 days, and at a year it returns 22.93% net. Everything shorter than five months is, at the median, a donation to eBay.

Selling a card into a retail marketplace costs roughly 13% of the sale price once the final value fee and shipping are counted. Because the fee comes off the sale and not off the gain, the gross move required to break even is 1 divided by 0.87 minus 1, which is 14.94%. That is the fee-only floor and it is the optimistic one. It assumes you sell at the quoted market price, and the quoted market price is an ask. Allow a 10% gap between the ask you bought at and what a seller actually realises and the required gross move rises to 27.71%. We report both settings throughout and never average them.

Those are gross numbers and they look generous. Now apply the floor. A 30-day hold needs 14.94% and delivers 1.99% at the median, so the median 30-day holder is down 11.27% after costs and only 12.15% of all 30-day windows came out ahead. That share is the number to carry around. Six in ten 30-day windows ended higher than they started, and only one in eight ended high enough to pay for the round trip. The gap between those two facts is the entire difference between watching a price and owning a card.

Walking the horizon week by week pins the crossing precisely. At 150 days the median hold is still down 0.25% after fees. At 155 days it is up 0.42%. The median holder of a Pokemon card in this population begins to make money somewhere in that five-day band, and not before.

One thing this study can rule out cleanly. Entry price does not meaningfully change the odds on a year-long hold. Across the four bands holding more than a thousand windows each, the share clearing the fee-only floor sits between 74.14% and 76.71%, and only the $100-and-up band falls away, to 69.81%. The median gross return is flattest of all, between 35.43% and 45.36% across every band. Whatever separates a good hold from a bad one in this population, it is not the price tag.

That result sits comfortably beside the wider factor work in What Predicts a Pokemon Card's Return, which found a rank correlation of 0.0180 between price level and forward return across 41,359 cards, and it points the same way: price level is not a signal. It also explains why so much retail advice about holding periods is untestable. The advice is usually about which card to buy, and the measurable part is how long you hold it.

As of 2026-09-22. Population: the per-card weekly `ppthist:<tcgId>` series in `api_cache` joined to `cards` on `tcg_id`, restricted to cards whose series reaches 2025-01-27 with at least 40 recorded points, giving 2,766 cards and a median recorded span of 603 days, the modal last recorded date being 2026-09-22. For each card and each recorded week, a hold window opens at that week's price and closes at the price recorded on or before the date exactly H days later, accepted only when that closing point falls within 10 days of the target; windows whose closing date exceeds the card's last recorded week are not opened. That produces 254,082 windows at 30 days, 200,688 at 90, 175,130 at 150, 163,466 at 180 and 91,986 at 365. Every panel is keyed on the TCGplayer product id and carries exactly one row per product. 51,719 catalogue rows carry a TCGplayer product id, and they resolve to 50,507 distinct ids: 1,212 ids are claimed by two rows each, 2,424 rows in all. Those pairs are not a Japanese print against its English twin, which is what we assumed before counting them. Every one of the 1,212 is a same-language pair, 972 Japanese and 240 English, and they are overwhelmingly one product indexed under two catalogue spellings of the same set, as the XY trainer kits are. Each id collapses to a single row, preferring an English row and then the lexically first catalogue id, so no product enters a cross-section, a factor sort or a census twice. The tie-break is stated rather than assumed away, because it decides which of the two spellings a shared product is filed under. Windows overlap by construction and are not independent observations; no significance test is run and none should be inferred. Both legs are TCGplayer market prices, which are ASKS, so every gross figure is an ask-to-ask change and no graded sold median enters any return. The cost model is applied as a single deduction on the sale side: net equals one plus the gross return, multiplied by 0.87 for fees and shipping, and by a further 0.90 in the second setting for the realisation gap, minus one. The 13% figure is the ordinary eBay final value fee plus shipping; the 10% realisation gap is our chosen severity rather than a measured clearing discount, and it is reported as a separate column rather than blended into a single headline. Breakeven gross is the reciprocal of the surviving fraction minus one: 14.94% at fees only and 27.71% with the gap. Grading costs, sales tax, insurance and storage are not modelled. The median convention is the upper-middle element on even samples. The broad-tape contrast uses `price_history` days 20695 and 20718 across the 43,800 products priced on both.