The reverse holo is the printing people throw in the bulk box. On 4,528 products where the same dump prices both a reverse holofoil and an ordinary printing on the same day, the reverse is worth more 95.54% of the time, and worth more than double on 50.13%. The median ratio is 2.03. This is not a rare card story. It is a story about the most common cards in every set.
The premium by era
| Era | Pairs | Median reverse / base |
|---|---|---|
| WotC | 110 | 11.15x |
| EX | 112 | 8.35x |
| XY | 281 | 2.45x |
| SM | 1,873 | 2.00x |
| Sword & Shield | 1,925 | 1.89x |
| Scarlet & Violet / Mega | 128 | 1.79x |
The era gradient is the mechanism. A modern set prints reverse holos of everything and they are easy to find, so the premium settles near 2x. The older the set, the fewer survived in condition, and the ratio climbs. EX Team Rocket Returns runs 12.68x across 95 pairs, with a median reverse of $65.00 against a median base of $5.82. EX Delta Species runs 8.49x across 107 pairs. At the other end, Champion's Path is 1.48x and only 75.9% of its pairs are above parity.
The cards that look like typographical errors
Rhyhorn from Legendary Collection is $0.48 as a normal printing and $92.46 as a reverse holo, a ratio of 192.6. Spearow in the same set is $0.56 against $100.00. Slugma is $0.53 against $65.00. Magby $3.30 against $124.99. Bagon $1.11 against $57.50. Every one of these is a card most people would not look at twice in the ordinary printing.
The pairs worth looking at
Population: per printing rows inside the daily TCGplayer dump, restricted to products carrying a reverse holofoil printing and either a holofoil or a normal printing, both with a market price above zero. That is 4,528 pairs, 4,491 of which match a catalog card. Both legs come from the same dump row on the same day at the same venue, so there is no time or venue misalignment. The codebase's own printing incoherence filter, which rejects a printing whose market price is below a third of its own listing floor, was applied and removed zero rows. Figures are medians and share above parity rather than means, because thin printings are noisy. Bucketing is by era and set only: we deliberately did not bucket by price band, because the base price is the ratio's denominator and banding on it would create a mechanical relationship. No regression is run.