Where does Prospect Card Radar get its pricing data?

Pricing data comes from eBay's official Browse API for live active listings and the Finding API for recent sold listings. PCR queries each card type for each tracked prospect on a recurring schedule, parses the listing titles to classify parallels and grades, and stores the results.

How often is pricing data refreshed?

The top 100 prospects refresh four times per day (roughly every six hours). Ranks 101 to 300 refresh once per day. Stats from the minor leagues snapshot once per day, in the morning. Rankings update daily from a composite of MLB Pipeline, Fangraphs, and Baseball America.

What prospects are tracked?

The top 300 MLB prospects, with the catalog sourced from MLB Pipeline, Fangraphs Top 300, and Baseball America consensus. The list refreshes daily and is fluid as players debut, get traded, or drop in rank.

What card types are priced?

Bowman Chrome 1st Bowman autos, Bowman's Best autos, Bowman Pro Debut autos, plus the major refractor parallels (refractor, atomic, gold, orange, red) for each. The exact set varies by prospect based on what's been released.

What do the underpriced, fair, and overpriced verdict labels mean?

Verdict labels compare a prospect's current market price percentile against their stats and ranking percentile. Underpriced means price is significantly below where stats and rank would suggest. Fair means price is roughly in line. Overpriced means price has run ahead of the underlying performance. These are observations about the data, not investment advice.

How accurate are the prices?

Prices reflect a specific, narrow sample of eBay listings that match PCR's query filters at refresh time. They are most reliable for the top 100 prospects with high listing volume. For lower-ranked prospects with thin sample sizes, the displayed average can move significantly based on a single listing. PCR shows sample size with each price to make this visible.

Methodology — How Prospect Card Radar Works

On this page

Data sources
The prospect catalog
Refresh cadence
Card parser & parallels
Verdict labels
Known limitations

Data sources

Pricing data comes from eBay's official APIs. Active listings and asking prices come from the Browse API. Recent sold prices come from the Finding API. Both are queried per prospect, per card type.

Stats come from Fangraphs minor league pages and from MLB.com's prospect stat feeds. These are scraped or fetched on a daily schedule and stored in PCR's database.

The prospect ranking is a composite of three sources: MLB Pipeline (top 100, scraped twice for stability), Fangraphs Top 300, and Baseball America Top 100 when available. Each prospect's composite rank is the average of their rank in whichever sources cover them. The catalog rebuilds daily.

The prospect catalog

The site tracks 300 prospects at any given time. The catalog is dynamic, not a fixed list. When a prospect debuts in the majors, gets traded, or drops out of the consensus top 300, they roll off. When a new player enters the top 300 (a new draft class, an international signing, a breakout), they get added.

For each prospect, PCR maintains:

Name, current team, position (P, C, IF, OF)
Current rank (1 to 300) from the source above
List of tracked card products (varies by player based on what's been released)
Headshot image (sourced from MLB media when available)

Refresh cadence

The exact schedule of automated refreshes:

What	How often	Notes
Top 100 prices (active listings)	Every 6 hours	Four snapshots per day, split into batches to stay under API rate limits.
Ranks 101–300 prices	Once per day	Runs in the morning, batched in 9 parts.
Minor league stats	Once per day	Pulled around 11:30am Eastern.
Prospect rankings	Daily	Composite rebuilds daily at 11:10am Eastern to catch source updates.
Sold listings (historical sales)	Once per day	Last 90 days of sold listings for each tracked card.

Every page has a "last updated" timestamp visible at the bottom so you can confirm freshness.

Card parser and parallel classification

eBay listing titles are messy. The same Bowman Chrome 1st Bowman auto can be listed as "2024 Bowman Chrome Prospect Auto," "2024 BCP Chrome Auto," "Bowman Chrome 1st Auto," and a hundred other variations. PCR's parser does its best to normalize these into a consistent set of card types and parallels.

The classification hierarchy:

Product: Bowman Chrome, Bowman's Best, Pro Debut, Topps Chrome, etc.
Card type: 1st Bowman auto, regular auto, base card, insert.
Parallel: base refractor, atomic, gold (/50), orange (/25), red (/5), superfractor (1/1).
Grade: raw vs PSA 10, PSA 9, BGS 9.5, SGC 10.

Each card is given a single canonical bucket. Listings that don't match a recognized pattern are dropped from the price calculation so they don't pollute the average.

Verdict labels: underpriced / fair / overpriced

The verdict is a single label per prospect that compares two percentile ranks:

Price percentile: where this prospect's price sits within the top 300.
Performance percentile: where their current stats sit relative to peers, blended with their consensus ranking.

If the gap between the two is meaningfully positive (price is well below performance), the label is underpriced. If it's meaningfully negative (price is well above performance), the label is overpriced. The middle band is fair. The thresholds are intentionally conservative so the label doesn't change every refresh.

To be clear: this is a math observation about the data PCR has. It is not investment advice. Cards can be cheap because the player is injured, has off-field issues, or is fading without it showing up in stats yet. The label is a starting point for your own research, not the answer.

Known limitations

The things the site can't currently do well, kept here so you know what to discount.

Small sample on low-ranked prospects. A prospect ranked 250 might only have two or three active listings of a specific card type at any moment. The displayed average is mathematically real but can swing a lot based on a single new listing. PCR shows sample size next to every price so you can judge confidence yourself.

Parser edge cases. The classifier sometimes mis-tags cards. The known active case is some IP-signed base cards getting bucketed as 1st autos when the seller titles them ambiguously. The team is filtering these out where possible, but the fix is ongoing.

Listings are not sales. Asking prices on active listings are what sellers want, not what buyers will pay. PCR cross-checks against the recent sold prices when available, but for cards that haven't sold recently, the active listing average is what's displayed.

No condition grading for raw cards. A raw card can be NM/Mint or scuffed to hell, and PCR can't tell from a listing photo. Raw prices reflect a blend.

International signings and new draftees lag. If a player just got drafted or signed and hasn't been added to the consensus rankings yet, they won't appear in the catalog. PCR refreshes the list daily to catch these as fast as possible.

If you've spotted something wrong, please tell me. The site only stays honest if collectors flag the mistakes.

Last updated: 2026-05-16

How PCR works