About the tracking dots

Most home and office printers add metadata to printed sheets in the form of very tiny yellow dots (sometimes called a machine identification code) that can’t be seen with the naked eye. The layout of the dots are different between printer brands and some don’t leave any at all. Information like serial number and sometime the print time is encoded in these dots. They can act as a “signature” for the printer that law enforcement uses as document forensic evidence (like in cases of forged currency).

To view the dots you need to a magnifying glass or high resolution image and you need to adjust the colour channels to emphasize the yellow. More info: How to Find Yellow Dots in Prototypes – DIY Guide

Important note:

The dots are printed on by the kind of printers you and I would buy. Things like magazines, posters, and real Pokemon cards will not have dots. They are printed by large industrial offset printers.

How to read the dots

Different brands use different dot encoding patterns, and not all of these can be decoded. The companies don’t reveal this information so any known pattern has been cracked by someone from the general public.

The most well known of these is was broken in 2005 and often referred to as the “Xerox DocuColor” code – named for the printer models used to discover it. Some brands other than Xerox also use this same pattern.

In this case, a 15×8 grid repeats like a checkerboard across the entire printed sheet. Each repeated grid has the same dots. The columns represent binary numbers For example, the eighth column encodes the year in binary. “2024” would appear as “011000” which is 24 in binary.

If you have a “Xerox DocuColor” style code, you can use @mika’s decoder implementation here to decode it: Yellow Dot Decoder – Fake Prototype Playtest Cards

Tracking dots on the “prototype” Pokemon cards

A slew of cards that were originally believed to be early prototype and cards used for playtesting have been appearing in public auctions starting in 2024. Most are believed to trace back to Takumi Akabane – someone who was involved in the early development of the TCG. The combined sales across all auction websites likely exceeds $10M. Individual cards were selling between four to six figures, based on the variant and the popularity of the Pokemon.

Because these cards are printed from a home/office printer, many of them have tracking dots. And many have a dot patter that suggest a print date in 2024.

Note the decoding result. We see this serial number reoccur across many cards.

Checking the dots like the above example can be done using public images if they are sufficiently high resolution. Such as this signed one where the autograph was witnessed by CGC:

The dots reveal the 2024 year and the same printer serial number.

Variant breakdown

The development of the Pokemon TCG was an iterative process, and many prototype and playtest variants have surfaced. Based on evidence inherent in the cards, this is what I can put together to the best of my ability

Note: only Alpha playtest and Beta presentation have any documented evidence of their existence that predates 2024

Alpha prototypes

• there are no yellow tracking dots since they were printed in b&w
• 2 variants, we can call them HQ and LQ: Many of the Pokemon playtest cards were likely printed in 2024 – #8 by pfm
• easily distinguished by the background
• there is one set of HQ, all other copies appear to be LQ
• LQ appear to be copies of HQ, as they pick up the same printing artifacts seen on the HQ version.

HQ: inconclusive.
LQ: likely a copy of HQ, when the copy was made is inconclusive.

Alpha prototypes

• Most carry a unique dot pattern (different from beta) Many of the Pokemon playtest cards were likely printed in 2024 – #18 by pfm
• Likely from a Konica Minolta (Japanese brand) or Epson printer Many of the Pokemon playtest cards were likely printed in 2024 – #421 by pfm
• Some copies have BOTH alpha dots and “xerox” dots that can be dated. The hypothesis is that some alpha cards were scanned/copied (picking up alpha dots) and reprinted in 2024: Many of the Pokemon playtest cards were likely printed in 2024 – #143 by pfm
• the alpha with inflated pop reports are more likely to have the double pattern (many trainers): https://www.cgccards.com/population-report/tcg/pokémon/2/test-prints-oddities-other/692/alpha-playtest-thick-lines/13952?populationID=1288703
• a small handful have noticeably higher quality backs such as the three starters and this gyarados: Many of the Pokemon playtest cards were likely printed in 2024 – #59 by pfm

High quality back: dated to 1996 for the three starters
Alpha pattern only: inconclusive.
Double dot pattern: all observed copies so far have 2024 dots

Beta playtest

• two variants, HQ and LQ: Many of the Pokemon playtest cards were likely printed in 2024 – #299 by pfm
• HQ has yet to be observed with any dots
• LQ has always been observed with 2024 dots
• there only only appears to be 1 HQ beta per Pokemon, implying the vast majority of betas have 2024 dots
• HQ betas have subtle traces of the alpha pattern on the back, suggesting that the HQ beta back was scanned/copied from a card that has the alpha print pattern – this implies the HQ betas were printed after the alphas and the authenticity is contingent on the authenticity of the alphas: Many of the Pokemon playtest cards were likely printed in 2024 – #745 by pfm
• it is unlikely that the LQ are simply scans/copies of HQ given that not a single print artifact or dirt spec has yet been observed to be transferred from HQ to LQ

HQ beta: inconclusive
LQ beta: all observed copies so far have 2024 dots

Delta playtest, Alpha presentation and Gamma playtest

• There are few of these graded but they all look similar to LQ beta and all observed and tested copies of all three variants have 2024 dots
• No copies of Gamma have been graded. The art does not match any existing card.

Beta presentation

• No hi-res scan available yet, but I see no evidence of dots so far
• They match well with the corocoro images and the print quality suggests a different printer was used for these than anything mentioned so far
• inconclusive so far

Acknowledgements

Thanks to @tediorso for suggesting this avenue of investigation.
Thanks to @mika @linkdu83 and @HumanForScale for helping with data collection and analysis
Thanks to @BANKS for working on information summary