View Full Version : Design question about booster pack distribution model

05-15-2013, 09:23 AM
More of a question for anyone at Crypto, but some wild speculation or insight into other games would definitely be useful!

- Booster packs are generated using a random seed on the server and then passed to the game "client' upon triggering the opening

The question:
If the above assumption is true, are the seeds (boosters) pre-generated in lots using some algorithm that allows for set distribution of rarities or individual cards across X boosters, or are the seeds (boosters) generated individually and rely on pure probability for a specific cards (or types/rarities) to show up?

The former method is similar to what my understanding of how physical TCGs work, with there being a set number of "super-ultra-rare" per box of boosters and so many with a palette of boxes, etc. Of course, there is no concept of booster "boxes" (but there could be!)

The latter is how I've seen other collectible games handle it. This method lends itself to pure probability.

With the volume of digital boosters created, the whole thing might be moot, but I'm mainly curious to know if both approaches were considered and why one chosen over the other.

Thanks for any insight and design schooling you can give!

05-15-2013, 09:29 AM
From a coding perspective, I believe it would be easier to have the "pack" run a script at the time of "opening" that Random Number Generated the cards according to the alogorithm (15 cards, 1 rare/legendary, 4 uncommons, 10 commons).

The packs aren't "stored" anywhere per se. If the interface works like MTGO, then the packs are infinitely purchasable and are generated at time of checkout so I don't think the overhead would be worth it to generate each pack with "tags" for the cards contained in them prior to opening.

Just my 2 cents.

05-15-2013, 09:42 AM
@Deadpool Yeah, that's what we've done on other games I've worked on in the past. We decided to not pregenerate random seeds, but instead generate them at the point of purchase so that the code could pull the most recent (and tuned) version of the booster pack generating values.

I ask because I've seen the problem where, in a certain set of booster packs (more than what a typical run of a physical card game is), some cards just show up so infrequently (because it's "random") that is skews the overall rarity.

If all cards of the highest rarity showed up the same number of times per X cards (even distributed within their own rarity), that seems less likely to skew things.