You add the same cards to two collection apps and get two different totals. Not a little different. Hundreds or thousands of dollars different.
Neither app is wrong. They are measuring different prices, and almost none of them tell you which price they measured.
We valued the same 25-card collection four legitimate ways and got totals from $17,148 to $27,516. The difference came from the pricing method, not an error in the apps. In our ten-card check, Collectr and Pokedata often showed the same price on liquid cards; the bigger gaps come from thinly traded cards and from method choices. The data behind this is TCGinvest's own: we track every card's European and US price along with the number of completed sales behind each figure, which is what makes the comparison possible.
The experiment
Collectr and Pokedata are the two apps we get asked about most, but the same issue affects almost every collection tracker, so we started by taking the apps out of it entirely.
We took 25 well-known Pokémon cards (Charizards, Umbreon VMAX, Lugia, Shining Mewtwo, Pikachus), the kind of collection a serious hobbyist actually owns. One identical basket. Then we valued it four different ways, using four price types that real tools really use.
Every method below is legitimate. Every one is used in production by somebody.
The same 25 cards. The same day. A 60% difference between the highest and lowest total.
Widen it out and the pattern holds: across all 3,049 cards we can price by all four methods simultaneously, the same basket comes to $136,241 or $213,270 depending on method, a 57% spread. This is not a quirk of the cards we picked.
Two forks sit between your cards and the number on the screen: which market, and which price. Change either one and the total changes, without a single price being wrong.
Why the gap exists
An asking price is not a sale. A trend line is not today. A modelled estimate is not a receipt.
An app showing you $27,516 and an app showing you $17,148 are not contradicting each other. One is totalling what sellers want. The other is totalling a lagging average.
The part that decides how big your gap is
Here is what most comparisons miss: the disagreement is not evenly spread across your collection. It concentrates almost entirely in cards that rarely sell.
Two cards from the same basket:
Umbreon VMAX, Evolving Skies #215. 60 completed sales behind it.
All four land within about 13% of each other. When a card trades constantly, every method converges, because they are all describing the same busy market.
Lugia ex, Unseen Forces #105. 30 completed sales, a card printed in 2005.
Same four methods. 3.5x apart. Nothing is broken here either. There is simply no crowd enforcing a price, so one optimistic listing and one lagging average pull in opposite directions.
We measured this across the whole dataset in our study on why price sites disagree: Pokémon cards with fewer than 31 completed sales show a 2.46x median gap between the highest and lowest pricing method. Cards with 31 or more show 1.82x.
So the practical rule: the more vintage and thinly-traded your collection is, the further apart different pricing methods become. Apps that rely on different methods will naturally produce different totals. A modern collection of heavily-traded chase cards will look roughly the same everywhere. A vintage collection will not.
The second reason: two continents
Three of the four numbers above started in euros and one in dollars, and converting between them does not make them comparable, because Europe and the US are genuinely different markets.
Across the 3,370 Pokémon cards we can price in both, the median card sells for about 31% less in Europe than in the United States. An app pricing your collection off European data and an app pricing it off US data will disagree by roughly a third before any methodology difference even enters.
There is a limit to that measurement worth stating: the European figure comes from Cardmarket's 30-day average, which includes every language and condition sold, so it sits below a Near Mint English copy. Treat 31% as the widest reasonable reading of the gap, not a like-for-like figure.
How to read your own collection total
- Ask, or completed sale? An ask is what someone wants. A sale is what someone paid.
- How many sales is it built on? A median from 60 sales is dependable. A median from 3 sales is an anecdote wearing a fact's clothing.
- Which market? A euro price and a dollar price on the same card are two different markets, not one number in two currencies.
If an app cannot tell you which of those three it used, its total is a number without a question attached.
What we do about it
At TCGinvest, we show Cardmarket data for Europe alongside PriceCharting and TCGplayer data for the US, with each figure labelled by source and method, and the sales count shown next to it. Collapsing them into one number would throw away the most useful thing about having both. Where a card's data is too thin to trust, we show a dash rather than invent a number.
That is the only real defence: say which question you are answering, and show your sample size.
So we checked the apps themselves
Everything above uses our own data. But the question people actually type is "why do Collectr and Pokedata show different values." So we looked up ten cards in both, logged in, and put their displayed prices next to what those cards actually sold for. We deliberately picked ten cards spanning modern and vintage, liquid and thinly traded. The goal was not to audit every price in the apps, but to see whether the same cards produce materially different answers.
Here are seven of the ten, with each app's displayed price next to the median of completed US sales over the last 180 days:
The result was not what "Collectr vs Pokedata" implies. On three of the cards, the two apps showed the identical price to the cent (Umbreon VMAX at $2,242.68, Rayquaza VMAX at $1,249.89, Mewtwo-GX at $751.62). For those cards, choosing between the two apps made no practical difference. That result is consistent with both apps displaying the same underlying TCGplayer value for those cards.
The bigger gap was not between the apps. It was between both apps and completed sales. Where the two agreed, they still sat 7 to 25% above the 180-day completed-sale median we measure. Whatever the displayed figure is built from, on these cards it consistently sat above the middle of recent completed sales.
They only genuinely disagreed on the vintage card. Lugia ex from Unseen Forces: Collectr $902, Pokedata $600, a 50% gap on the same card. Exactly where the sales dry up, exactly as the rest of this article predicts.
So the honest answer to "which app is more accurate" is: on cards that trade constantly, our sample suggests they often show the same number anyway. On thin cards, no shared figure can help, and the most informative number is the one shown with its sales count, because it answers the question "what have buyers recently paid?"
Frequently asked
Why do Collectr and Pokedata show different values for the same collection?
Often they do not. In our ten-card check, three cards showed exactly the same price to the cent, consistent with both apps displaying the same TCGplayer value. When totals do differ, it usually comes down to the price method (asking price, trend line, modelled estimate, or median of completed sales) and the market (Europe or the US). Those choices alone move an identical 25-card collection between $17,148 and $27,516. The larger gap is usually not app versus app, it is app versus what the cards actually sold for.
Which value should I trust for insurance or selling?
For selling, the median of completed sales is the closest thing to what a buyer will actually pay. For insurance, replacement cost is closer to the asking price, since that is what you would pay to buy the card back. They are different questions with different correct answers.
Why is the gap bigger on my vintage cards?
Thin markets. Cards with fewer than 31 completed sales in our dataset show a 2.46x median gap between pricing methods, against 1.82x for cards with more. Fewer sales means each individual listing moves the number more.
See it yourself: every card page on TCGinvest.io shows the European (Cardmarket) price and the US (TCGplayer and PriceCharting) price side by side, each labelled by source and shown with the number of completed sales behind it.
TCGinvest measures prices daily from Cardmarket (EU) and PriceCharting/TCGplayer (US). The 25-card basket and the 3,049-card universe both require all four price types present, at least 8 completed sales, and exclude 1st Edition printings. EUR converted at 1.1573. Nothing here is financial advice, and every number moves.
