Bubble Pack

Armchair analysis · 3 of 3

8 min read

Napkin math: the life of a pack

A pack sells 69% of its six years in year 1, and a 10k reprint at month 24 takes 40 months to sell: why packs rotate out

In this article · 6 sections
  1. What’s built
  2. What’s played
  3. What sells
  4. Why packs rotate out after 2 to 3 years
  5. Check it yourself
  6. Sources

Every day in 2025, the community published 30 decks and logged 400 games, and each product sold 5 copies on Amazon US. marvelcdb.com tracks the decks, marvelchampionstracker.com the games, and Amazon US the monthly sales estimates, which I captured by hand in the second half of the year. I kept the quality submissions only: no decks without a description, with a default title, or under 40 cards.

This is back-of-the-napkin math. Every number below is an estimate built on public data and a few of my own records, and I will refine them as better data comes in. Beast checks the arithmetic.

Decks, sales and plays are correlated. What people build is what they play, and sales follow. Doctor Strange is the exception that proves it: no sales (the pack was out of stock), few new decks (28th), and still the 11th most played hero. Three movements follow: what’s built, what’s played, what sells. The last one explains why a pack rotates out after 2 to 3 years.

What’s built

The average deck holds 42 cards. I was curious about the power of double resources: to price a draw in ER, I needed to know how often decks run extra ones, the “Power of” family, which speeds up the cards you can play. On the way I found how popular they are, and it was too good a curiosity not to share: 44% of decks run them, and the share changes with the aspect.

Aspect”Power of” usage
Aggression48.6%
Leadership45.3%
Justice40.5%

What’s played

Standard I is not the default

Only 65% of games are played in Standard I: 81,350 of the 124,350 plays across the three difficulties. Standard III is a resounding success: 12.8% of plays, 59% as many as Expert, and the best win rate of the three.

DifficultyPlaysShareWin rate
Standard I~81,35065.4%69.7%
Expert~27,05021.8%66.0%
Standard III~15,95012.8%72.5%

Players pick the powerful heroes

Spider-Man has the most plays, 7,450, nearly 5% of all games and 26% more than Captain America. Captain America wins more (79.3%, against 75.3% for Spider-Man) and sits in both top 5s. The five best win rates sit between 79.1% and 81.3%, 10 to 12 points above the 69.3% average, and Doctor Strange leads.

Most playedGamesWin rate
Spider-Man7,45075.3%
Captain America5,90079.3%
Iron Man5,55072.3%
Captain Marvel5,00075.4%
Black Panther4,45068.7%
Best win rateWin rateGames
Doctor Strange81.3%3,150
Storm79.6%2,000
Captain America79.3%5,900
Maria Hill79.3%1,750
Magik79.1%2,350

Maria Hill has more plays (1,750) than the other new heroes, which says powerful heroes are really appreciated.

New players keep coming in

Rhino appears in 15,400 logged encounters, about 10% of all games, more than any other villain. He is the first scenario of the core set, and expert players rarely come back to him. His lead measures the core set’s weight in the experience: new players are still coming in, and they start with Rhino.

What sells

From Amazon US to global sales

Multiply Amazon US sales by 10 to estimate global sales. Amazon US is one storefront in one country and one channel, so two ratios scale it up:

  • Amazon US is 25% of US sales. Online is 20% to 25% of US toy sales and Amazon is about 60% of that online revenue, so 12% to 15% for toys as a whole. Hobby games sell more online than toys, and board game roundups go up to 35%, none of them audited. ICv2 sizes the US and Canada hobby store channel at $2.84 billion in 2024, and leaves Amazon out of that number.
  • The US is 40% of global sales. Market research roundups put North America at 42% of the board game market. For this license, the Marvel movies give a gauge: the US and Canada took 33% of Infinity War’s box office and 31% of Endgame’s. The license is not more US-centered than games in general, so I keep 40% and test 30% and 50%. Asmodee’s own accounts show 15.7% of net sales in the US (28% for the Americas), but that is the whole catalog of a European group.

Multiplier: 1 / (25% × 40%) = 10. A product selling 5 copies a day on Amazon US sells 50 worldwide. Here is how it moves, with the months a 10k reprint needs to sell at 25 copies a month on Amazon US:

US share of globalAmazon US 12% of USAmazon US 25%Amazon US 35%
30%27.8× · 14 months13.3× · 309.5× · 42
40%20.8× · 19 months10.0× · 407.1× · 56
50%16.7× · 24 months8.0× · 505.7× · 70

Each cell is 1 / (US share × Amazon US share), then 10,000 / (25 × multiplier). At 50 copies a month on Amazon US, every duration halves: from 7 to 35 months. A reprint that sells through in 24 months or less fits the 2 to 3 year window: 3 of 9 cells at 25 copies a month, 6 of 9 at 50.

The shelf life of a pack

A pack sells 69% of its six years of sales in its first year. Silk sold 1,500, 1,000, 500, 300 and 100 copies on Amazon US in her first five months. Assume the drip then falls from 100 to 25 copies a month by month 24, and stays at 25. Silk is one pack, a hit. I lost the spreadsheet with the other packs’ monthly numbers, so I will add more curves later, built with the script at the end of this article.

PeriodAmazon USWorldwide (×10)Share
Year 14.0k40k69%
Years 2 to 61.8k18k31%
0 5k 10k 15k 15k M1 10k 5k 3k 1k M5 M12 250 M24 M36 Floor: 250 a month

Silk, worldwide copies per month (Amazon US × 10). The yellow bars are the floor after month 24.

Silk’s first five months are 3,400 copies on Amazon US and 34k worldwide, which is 59% of her six-year total. She never went out of stock, so her first run was bigger than 10k: 15k worldwide sold in the first month alone.

The core set never stops: 400 copies a month on Amazon US in 2025, which is 4k a month worldwide and 48k over a year, more than the 40k of a hero pack’s best first year.

Why packs rotate out after 2 to 3 years

Since February 2025, Fantasy Flight’s current environment holds roughly 2 to 3 years of expansions. Older ones are retired and not reprinted. The napkin math shows why: a reprint has to sell through in 24 months or less to fit that window.

At month 24, a pack sells 25 copies a month on Amazon US, which is 250 worldwide. Printers have been indiscreet with me, and a minimum order quantity of 10k is right in the range of what they quote. A reprint then takes 40 months to sell (10k / 250). At 50 copies a month on Amazon US, it takes 20.

Storage is not the problem. 3PL warehouses charge about $20 per pallet per month, from $14 at enterprise volumes to $40. A hero pack box measures 15 × 9 × 2.5 cm, which is 0.012 ft³, so 10k packs are 120 ft³, a few pallets. A hero pack retails around $13.5 to $20 in the US, and a publisher keeps about 40% of the retail price: $6 on a $15 pack, so $60k for a 10k reprint.

Take 5 pallets at $20: $100 a month, so $4k over 40 months, which is 7% of the $60k. Revenue at the floor is 250 × $6 = $1,500 a month worldwide. The reprint beats its printing bill if printing costs less than $5.6 a pack ($56k / 10k), and it takes 10k × the printing cost / $1,500 months to earn it back: 17 months at $2.5 a pack, 37 months at $5.6.

The cost of a reprint is time and cash, not shelf space. At the floor, a pack earns $1,500 a month worldwide, locked behind a printing bill paid on day 1, and Asmodee releases over 1,000 new SKUs a year. At 25 copies a month on Amazon US, 3 scenarios in 9 sell a reprint inside 24 months, and a pack is slow money. At 50, 6 in 9 do, and you reprint it.

Check it yourself

Amazon US prints a rounded “bought in past month” figure under each product. Open an Amazon US search for Marvel Champions, paste this in the browser console, and read the floor of each pack’s monthly sales:

[...document.querySelectorAll('[data-component-type="s-search-result"]')]
  .map((r) => [
    r.innerText.split("\n").find((l) => /Marvel Champions/.test(l)),
    r.innerText.match(/([\d.]+K?\+) bought in past month/)?.[1],
  ])
  .filter(([title, bought]) => title && bought)
  .map(([title, bought]) => `${bought}  ${title}`);

The figure is rounded down to 50, 100, 200, 300, 400. Run it once a month and you have a sales curve, and an answer to the question the conclusion hangs on: does the floor sit at 25 copies a month or at 50?

Sources