What the Atlas Estimate Is, and How It's Calculated
The Atlas Estimate is our best answer to one question: what level is this card actually trading at right now?
It is not the highest price we can find. It is not what someone is asking. And it is not an average of every website glued together.
Atlas collects readings from several markets every night, then checks them one stage at a time — is this trustworthy, and does it really belong to this card — before anything reaches the final number.
TL;DR
Atlas pulls from seven marketplaces every night. We do not average everything we find. First we ask:
- Is this price really for the same card?
- Is it a real sale, or just someone's asking price?
- How much should this marketplace count for this language?
- How fresh is it?
- Does it make sense against the card's own price history?
Whatever fails is dropped. What survives is weighted, trimmed of outliers, and combined into the Atlas Estimate.
And sometimes the answer is no price — on purpose. A wrong price is worse than no price.
Here's what happens behind the scenes.
Step 1 — Seven marketplaces
Atlas reads:
- 🇺🇸 TCGplayer
- 🇺🇸 eBay
- 📊 PriceCharting
- 🇯🇵 SNKRDunk
- 🇪🇺 Cardmarket
- 🇪🇺 CardTrader
- 🇯🇵 yuyu-tei
But they do not all count the same.
What actually sold on TCGplayer is a very different fact from someone listing a card at five times the market and waiting. So Atlas labels every reading as either
💰 SOLD or 🏷️ ASKING
because those two things are not the same kind of evidence.
Step 2 — Is this even the same card?
This may be the most important stage of all.
Marketplace search will happily mix together:
- the base card
- a parallel
- an alt art
- a manga rare
- a promo
- another language entirely
and some of those share the same card number.
So Atlas does not ask "what price did we find?" It asks:
"Can we prove this price belongs to exactly this card?"
- An exact match scores full marks.
- A price blended across several variants gets its weight cut.
- A price that belongs to the wrong card is thrown away.
Otherwise a $600 card quietly inherits the $4 price of its base version. 💀
Step 3 — Not every market gets an equal vote
For English cards, TCGplayer and Cardmarket carry the most weight.
For Japanese cards, yuyu-tei and SNKRDunk do.
The reasoning is simple: we lean on the market that actually trades that language.
Freshness matters too. A sale from yesterday tells you far more than a sale from eleven months ago, so some sources decay in weight over time while the ones we read daily stay current.
Step 4 — Throw out what doesn't belong
Now we look for readings that don't fit the rest.
If the data says:
$4
$5
$6
$5
$80
…that $80 deserves another look.
Atlas uses several rules to catch bad readings:
- outliers far above or below the rest
- data that has gone stale
- suspicious asking prices
- prices leaking in from a different print
- eBay with too few sales to be meaningful
And one detail that matters: we don't let the biggest marketplace decide on its own. We compare against the shape of everything else that survived. A single strange price should never get to set the market.
Step 5 — But what if the market really moved?
This is the interesting part.
Atlas keeps an established level for each card, based on where it has been trading recently. It works as a reference point, and it stops noisy data from throwing the price around.
But we don't want that reference to become a cage.
So when several marketplaces move to a new level and agree with each other, Atlas can accept the change.
Last week:
$0.24
Then several markets start showing:
$4.20
$4.50
$4.80
Atlas can conclude: okay — this looks like a real move, instead of stubbornly holding the card at $0.24 forever.
Step 6 — Finally, the estimate
After all the filtering, weighting and checking, whatever survived is combined.
The result is a weighted mean. Not "take the highest". Not "take the lowest". Not simply "take the middle". The final number comes from everything that made it through.
And we tell you how confident we are
Every estimate arrives with context.
- 🟢 Several sources that agree → higher confidence
- 🟡 Limited data → medium
- 🔴 A single source → low
We also show the range the surviving readings covered.
And if everything left is an asking price with no real sale behind it, the card is marked 🏷️ Listings Only, so you know the number reflects what sellers want, not what anyone has actually paid.
Sometimes Atlas shows no price at all
This might be my favourite rule in the whole system.
If every reading fails its checks, we do not invent a number.
- No dressing up a 2022 sale as today's price.
- No borrowing the neighbouring card's price.
- No giving an expensive parallel the base print's number.
- No treating someone's $2,000 wishful listing as market value.
Just: "No price."
Because the goal was never to put a number on every card. The goal is for a card to carry a good, evidenced price when the data genuinely supports one.
So the Atlas Estimate isn't just "a number"
It's a conclusion drawn from a lot of data that had to survive a lot of checking.
And you can open the whole thing up on any card page:
- Open a card
- Press "How we got this"
- See what was kept
- See what was rejected
- See why
The estimate is the last line. Everything before it is the evidence.
Disclaimer
OP-Atlas is an independent project and is not affiliated with, endorsed by, sponsored by, or associated with Bandai, Shueisha, Toei Animation, or the One Piece intellectual property holders. Market data may contain errors, delays, incorrect matches, incomplete data, or other discrepancies, and the Atlas Estimate is an analytical estimate, not a guaranteed market price or a value at which a card can actually be bought or sold. Prices can change rapidly. This post is for informational and entertainment purposes only and does not constitute financial, investment, trading, or legal advice.
Content note: AI was used as a tool for quality assurance (QA) and translation support. Final content was reviewed and edited by the OP-Atlas team.