Sun & Moon
As mentioned in the introductory post to the Sun & Moon series, in my model I am assuming 5.6 holos, 4.0 “regular” Pokémon-GX, 1.0 full art Pokémon and 1.4 full art trainers/secret rares per box on average. The latter sheets are labelled “GX”, “FP” and “FT” in the tables. From this set onwards, one Basic Energy Card is included in each pack, bringing the total number of cards to 11. I am including it in the table (the sheet is labelled “E”), but not in the sum of cards (“set”) on the right; the master set for Sun & Moon including reverses thus includes 289 cards and not 298. Here is the rarity table:
It is a very large set, and the rarest cards (10 of the secret rares) have the second-lowest pull rate (1 in 642.9 packs) of any set up to this point (Plasma Freeze still holds the record according to my model, with 1 in 720). As mentioned in my introductory post, I am assuming each full art Trainer to be printed 10 times on the FT sheet, and 10 of the secret rares to be printed 4 times and 4 of the 5 times on the same sheet.
The raw data for the various sheets is shown below, with the dashed lines indicating where the transition from H8 to H7, etc., would occur:
Two observations:
As we will see in later sets, the practice of making the non-holo rares which were theme deck exclusive holos slightly rarer than the others was very much still a thing during the Sun & Moon era, but for this set I see no conclusive evidence for it. There are three theme decks featuring exclusive holo versions of Decidueye, Incineroar and Primarina. Decidueye and Primarina did appear the fewest times in the sample, but not by much, and Incineroar appeared very frequently, so I am guessing there are no artificial rarity differences.
For the secret rares, the ones with the higher observed pull rate are all rainbow rares and not Items/Energies. If this is replicated in the other two sets for which I am assuming different pull rates for different secret rares (Celestial Storm and Unbroken Bonds - see introductory post for the Sun & Moon series), it may be estimated to be significant.
Edit 2024/03/24: Changed sequence length from 180 to 360; this has no impact on pull rates. Made minor edits to text.