What is DataNeverLies?

The flagship. Every night the engine scores every actively traded card and grade from 0 to 100 with a proprietary multi-factor model. Buys flag cards set up to rebound, meaning a durable turn rather than a one-day blip. It then publishes plain-language BUY calls where the backtest supports them, currently sports cards only, with the strongest calls labelled STRONG BUY. Other categories publish when they pass the same test, and not before. The exact signals and their weights are ours.

Testing that gets stricter as our data does. Before launch we replayed the engine point-in-time across ~2 years of real daily market prices (100 weekly windows), using only the data that existed at each point. Since then we have graded outcomes against real completed sales, and we re-test after every data-cleaning pass. One of those tests withdrew our highest-scoring band for a month when it measured negative; a later cleaning pass found the bad sale data behind that reading, and with it removed the band re-measured as our strongest, so it publishes again. The latest full test puts published buys about 5 points ahead of the market over the following days, with solid statistical confidence. After typical selling fees the absolute return was not distinguishable from zero, so treat signals as a tool for buying at the right price, not a promised return. Every call is logged before its outcome is known.

And the part most sites won't tell you: where the test doesn't prove an edge, we don't publish. The sell side is the working example. It never produced enough graded results to show whether sell calls worked, so we stopped publishing them instead of leaving something unproven running. That's what "DataNeverLies" means.

A signal tells you when to act. These tools tell you at what price. For any card and grade, get a ready-made sell ladder (95 / 90 / 85 / 80% of market value) and buy ladder (80 down to 60%). On eBay, the ~13% seller fee is netted out of your sell targets so you see what you'd actually pocket (TCGplayer gets the same treatment on raw TCG cards, with its ~12.75% in fees); dealing cash at a show, the numbers work differently.

The buy advisor goes one step further: type in an asking price, pick the venue, and it tells you steal / buy / fair / rich / pass, factoring in the card's liquidity (thinly traded cards get an automatic discount) and any active signal on the card. It's guidance, not financial advice, and every buying decision, along with its outcome, is yours alone.

Thinking about grading a raw card? The pre-grade tool examines your photos (front and back), measures centering to the pixel, and returns a decimal-precision grade estimate, like PSA 9.54, inside a likely range, with per-category subgrades and plain-language reasons for the number: what supports it and which flaw is holding it back. It adapts its inspection to the card in the photo, whether that's a holofoil Pokémon card, a modern chrome parallel, or vintage paper. All before you spend grading fees and months of turnaround on a card that's unlikely to earn a top grade. It's an estimate to inform the decision, not a grade. A free account includes 2 pre-grades to try it; Pro removes the limit.

The same AI scanner identifies any card from a photo (set, number, parallel) and pulls its live prices. Slabbed card? It reads the label and cert number in seconds so you can verify it on the grader's site before money changes hands.

Prices for every tracked card come from actual market activity: primarily JustTCG's per-condition market prices (TCGplayer market price fills any gap) across all 13 tracked trading card games (Magic, Yu-Gi-Oh!, Pokémon, Pokémon Japan, Dragon Ball Super, Flesh and Blood, Star Wars Unlimited, One Piece, Union Arena, Digimon, Lorcana, Gundam, and Riftbound), and real recorded sold prices (from eBay and the major auction houses) for sports cards. No asking prices, no listings, no made-up numbers.

Raw and ungraded prices come straight from live market data (per-condition market prices). Graded values (PSA, BGS, CGC, SGC) show real graded-sale prices where they exist, and estimates everywhere else. Every estimate is labeled, never mixed in with real prices, and excluded from the signals and indexes. Raw TCG cards also get the standard condition ladder: Near Mint is the live market price, and Lightly Played through Damaged use standard market discounts, labeled as estimates.

Add your collection and it becomes a live portfolio: current value by real market prices, gains and losses against what you paid, and a performance chart over time. It's like a brokerage account for cardboard.

Watch any card and set price alerts in either direction; alert digests land in your inbox, so the market comes to you.

The built-in AI analyst answers with the actual guide in front of it: real comps, your real portfolio, live category indexes. Ask it what to sell first, whether a price is fair, or how a set has moved since January, and it cites the data it used.