How RFID playing cards work

An RFID playing card is a normal card with a thin inlay laminated into its core: an antenna loop of printed or wound conductor, and a chip a few millimetres across. There is no battery. The reader’s field powers the chip, which answers with a unique identifier.

These are high-frequency tags operating at 13.56 MHz, not the UHF tags used for pallets and inventory. HF is the right choice here for an unglamorous reason: its read range is short and its field falls away sharply. A card is legible when it is on the table and unreadable from across the room, which is a security property before it is a technical one.

The chip stores an identifier, not a card face. Nothing on the tag says “ace of spades” — that mapping lives in the system, which is why a deck has to be registered before it can be played.

Antennas sit beneath the felt — typically one per seat for hole cards, plus the board, the muck and the dealer position. A reader energises them in sequence and listens for whatever answers.

Two cards in the same field at the same moment would normally collide and garble each other. The protocol handles this with anti-collision: an arbitration scheme that lets the reader enumerate several tags individually rather than hearing mush. It is the reason a seat can be read as a pair rather than one card at a time, and it is the first thing to ask a card supplier about.

A deck that reads perfectly on a bench can misbehave on a table. The usual causes, roughly in order of how often they bite:

Detuning. An antenna is tuned to its environment. Metal in the table, a rail, a laptop, even a change of felt shifts that tuning and shortens range. A system commissioned in an empty room is not commissioned.

Stacked and overlapping cards. Cards shield each other. Two cards squared perfectly on top of one another are a harder read than two spread apart, and a player who protects their hand with a chip stack has added metal to the problem.

Adjacent antennas. Neighbouring positions bleeding into each other causes a card to be reported at the wrong seat — which is worse than not reading at all, because it looks like data.

Chip and reader mismatch. The single most expensive mistake. Several HF tag families exist and they are not interchangeable; a deck built around one is not readable by a reader stack built for another. The cards arrive, and they are decoration.

Wear. Inlays fatigue. A deck is a consumable and should be treated as one, with spares sealed and rotation planned rather than improvised on show day.

Specify the chip against your reader, not against a price list. Ask a card vendor which tag family they are supplying and confirm your readers speak it, before quantity or unit price enters the conversation.

Read-test before you commit. On your antennas, on your table, with cards stacked and offset the way players actually leave them.

Plan the consumable. Decks wear, get bent, and go missing. The order quantity is a show-operations question, not a procurement one.

Ask who owns the mapping. Registering a deck binds identifiers to faces. If that process is opaque or manual, it is a failure point on the day.

The moment a table can identify a face-down card, the production is carrying the most sensitive data in the building. Anyone with access to that feed knows what the players do not.

That is why a real system is built around a broadcast delay enforced in the infrastructure rather than the schedule, tiered access so that most operators never see live card data at all, and an audit trail that can answer afterwards who saw what and when. A vendor who talks only about read rates and not about the delay chain has solved the easy half.

We have been building this end of it since 2004. The record is public.