The Secret, Real-Money Economy for Digital Trading Cards
A data-driven analysis of collecting cards in Topps digital apps
September 2, 2015
This Topps Bunt 2015 Mike Trout Variant Signature card sold for $440.00 on eBay on July 14, 2015 — the highest recorded real-money transaction for a single digital Topps Bunt card. Yes, this is a screen grab of a digital baseball card in a smartphone app.
Executive summary
- Digital cards in Topps apps like Bunt, Huddle, Kick, and Star Wars Trader have real-money value.
- I outline the value indicators, the value factors, and the value phases for these digital cards using Topps Bunt as the data source.
- I present data that validate the card odds that Topps gives us.
- Data and analyses are presented on the value of cards, and show that secondary market (eBay) costs for Topps Bunt signature cards are ~20% of their in-game costs.
This is a long read. If you are an experienced Topps app user, skip right to the good stuff in section 4: The Data.
1. Background
Prior to downloading the Topps Bunt digital baseball card app it in the middle of the 2014 baseball season I had never made an in-app purchase in my life. I had never spent more than a couple bucks on any app in the iTunes app store. I’ve got to hand it to Topps.
Over a year later I am now several thousand dollars deep into Bunt purchases: some on the app, and some outside of it. It may be an addiction; it may be a healthy form of entertainment. Either way, I take it seriously, and have attempted to become “good at” collecting digital baseball cards (whatever that means).
Some background. Topps Bunt is a mobile app for your smartphone. It is fundamentally a fantasy baseball game where the players come in the form of digital baseball cards (yes, images on your phone screen you can flip over to see stats on the other side like real baseball cards). You can buy better odds at winning in the fantasy games by spending more money on packs that contain better-scoring cards. Topps also makes apps for football (Huddle), soccer (Kick), and Star Wars (Star Wars Card Trader).
But wait, you say — I’ve never “played” my physical cardboard baseball cards before. So on the other hand, Topps Bunt is fundamentally a medium for collecting digital cards. “Utterly insane,” you may retort; “Why would anyone collect digital images in a closed app run by a company with an uncertain future?” Yes, the logic is faulty in the long run. What if Topps fails to support the app- a very valid concern for anyone who collects these digital cards. No app, no cards.
Despite the risks, the fact is that a very healthy economy for these digital cards exists- not only in the game, but on eBay as well as on at least one separate, stand-alone digital card marketplace. As someone who falls squarely into the “collector” category rather than the “player” category, I’ve attempted to gain a deep understanding of this new digital card economy, and I attempt to share some insight here.
2. Collecting Digital Cards
Within the Topps app, there are two ways to acquire cards: one, via the purchase of packs with coins (900,000 coins can be bought within the app for $100). The most consistently popular cards are “signatures,” which are indeed cards that have a digital signature imposed upon an image of a ballplayer. One of the most popular signatures is that of Mike Trout. On opening day of the 2015 baseball season, Topps released 150 Mike Trout signature cards (somewhat predictably, since they did the same on opening day of 2014, so users were standing by for the release), at odds of roughly 1:150 (meaning one in 150 packs would contain the card). The pack sold out within an hour, meaning all 150 of the cards were pulled in packs in that hour. At 1:150 odds, 150 inserts being pulled over the course of one hour meant that the velocity of packs being bought was 22,500 per hour. At 5,000 coins per pack and $100 for 900,000 coins, Topps made over $12,000 in that hour. The value of these cards is real.
The pack purchase velocity of the 2015 Trout Signature may very well have been the highest of any card ever, but there is a wide spectrum of pack purchase velocities depending on the card; some won’t sell out for days. Either way, if a user missed the opportunity to log on in this hour of madness to try to pull the card in a pack, they could potentially acquire it via the second means: trading with other users.
Which brings us to the third way cards can be acquired, and the dirty little secret among most of Topps Bunt’s collectors: buying cards on eBay. Yes, cards can be bought and sold on eBay. A search for “Topps Bunt” on eBay at the time of writing this yielded 3,714 active listings for cards. In an eBay transaction, cards are transferred from owner to buyer via a trade in the app after the purchase is made and the buyer’s Bunt username is communicated to the seller. These is very little risk as any scam artist will get poorly reviewed on eBay. I’ve made well over a hundred Bunt eBay transactions on both sides of the table without a single problem.
3. Card Value and the eBay-Bunt Arbitrage
There are several ways of quantifying a card’s popularity, using indicators to predict a card’s value. Some value indicators I look for:
1. How quickly the pack containing that card sells out.
2. The ratio of the velocity of posts on the Fan Feed message board within the Bunt app that request a certain card relative to how many posts are offering that card. Popular cards will have everyone begging for one while none are being offered; for an unpopular card, there will be many trade offers that stay active because nobody wants the card.
3. How much people are willing to trade to get the card, which is related to point 2 above and, while subjective, a strong indicator if you’re an experienced trader with accumulated knowledge of card values.
4. The price on eBay.
When collecting, I now exclusively look at eBay price as a value indicator. I’ve found that it correlates very accurately to any of the first three value indicators.
I’ve also been able to roughly determine what makes a card valuable, but with much higher variability. Roughly speaking, value is some convolution of the following value factors:
1. The popularity of the player. Mike Trout, Derek Jeter, Klayton Kershaw, Giancarlo Stanton, etc., all command the highest values and often by a wide margin. While normally signatures are the highest-value cards in the game, some inserts for elite players can out-value signatures, especially if they meet some of the other following criteria.
2. The design of the insert. Signatures command a lot of value because autographs are unique. Sets with artwork can sometimes command a higher value for the same reason. This is highly subjective, though, and iffy at best; the Frozen set (in my opinion one of the ugliest sets in the game) is also perhaps the highest-value set in the game because of criteria 3–5 below.
3. The quality of players in an insert’s set. Most inserts are part of a small set, say 10–20 cards, released over time. Topps often indicates who the players in the set will be upon the release of the initial set. A set of entirely elite players will usually be of higher value and so each card within it will usually be of higher value.
4. The rarity of the set. This is the only quantifiable part of the card’s underlying value, but may not a good indicator of the card’s actual value due to the strong contributions from these other value factors. Each card has different odds, a different pack price, and a different card “count” (the total number in circulation). A card from the Limited set, at 1:500 odds with a total count of 50, in a pack that costs 5,000 coins, has a good chance of being much more valuable than a card at 1:30 odds with a total count of 1000 in a 2,500 coin pack. There are exceptions; a bad player fitting the former criteria may end up being cheaper than Mike Trout fitting the latter.
5. The award given for collecting the whole set. Often times, for collecting a set (like signature series cards 1 through 10), an award card will be given out. This award card will naturally be of significantly lower count than any of the cards in the set since not everyone who collected the lowest-count card in the set will have collected the entire set. These award cards themselves can be very valuable, and so the cards that are needed to attain them are also very valuable before the award is given.
So, does ranking each card on these five value factors allow one to accurately predict the trade or eBay purchase value of a card? Not so fast. Because of both the way Topps has structured the release of cards as well as the natural cycles of popularity that players see from season to season, we also have to take into account the structural changes in value over time. I’ve also been able to identify fairly clear phases that card values cycle through. A card goes through four main value phases:
1. Initial frenzy on the release day;
2. “Active chase,” ie, the period during which the card is part of a set whose reward is not yet given, or still “active;”
3. Rest-of-season, post-award;
4. Long tail, post-season.
One can only compare card values based for cards in the same value phase. Why? A card that has low value factors might command a higher price or trade value in the same-day frenzy or active chase phases than a card with very high value indicators from a set two years old.
Now for the fun part. Let’s look at the data that support the above statements.
4. The Data
Signature insert cards are, even according to Topps, the most popular cards in the game, and the ones that have the most consistent value indicators and value factors. For the past few months I’ve been taking data on my in-app purchases of Bunt signature packs and the sale values of signature cards on eBay. The data here are fairly representative of cards in other sets from my observations. And some of the data are surprising.
First, let’s check Topps’s math. I kept track of the number of packs I purchased for the three-month period starting May 27, 2015, along with the number of signature inserts I pulled. A summary of the data is in this table:
Then I calculated whether the actual odds I was observing were consistent with Topps’s stated odds. By tracking Topps’s stated odds in their in-app articles, I calculated the binomial distribution of the probabilities of pulling x number of inserts (the top gray curve, below). By tracking how many signature inserts I pulled (15) over the course of buying 711 packs in that timeframe, I calculated the binomial probability of that outcome (x = 15, the red dot below). In a statistically perfect world, I would have pulled 16 signatures in the 711 packs I bought given the cumulative odds of each pack (which vary from signature to signature).
Now, something for my fellow math nerds. I then normalized the x-axis to standard deviations from the mean, and calculated the z-value of my mean relative to Topps’s stated mean. Low and behold, the z-value of 0.4 means my observed mean of 15 signatures is only 0.4 standard deviations away from Topps’s mean of 16 signatures. This is pretty darn close, statistically speaking. Topps’s stated odds are believable. Feel free to check my math here (Excel download).
But whoa: something bothered me. The actual cost per card is a whopping $85.17. Meaning if I buy coins, then buy packs, and try to pull signature insert cards, over time the average cost will be eighty-five bucks per card! I may get lucky sometimes and pull in the first pack, but I may get unlucky and pull in the 1,000th pack. Thank you, binomial statistics.
So next, I took a more granular look at the data on individual cards- both my data and some data I pulled from eBay.
Let’s say you want to collect signatures and are willing to wait until a late-stage, lower value phase to purchase them. Is it worth it?
Abso freaking lutely! Wow, the difference is huge. The data above show very clearly that prices of Topps Bunt cards on eBay are significantly lower than if one were to try to pull them via packs in the app with purchased coins. And this is why I stopped buying packs and went from being a Black VIP (~$500 spent per month) to spending $0 in the app. Cards on eBay consistently cost ~20% of the cost of cards acquired in Topps Bunt through coin purchases and packs. Yup, 80% off on eBay.
The previous data used the cheapest seven transactions on eBay for each card, which usually fall into the low-value, later phases of a card’s life. Is it always the case that cards can be had so cheaply, even in the initial high-value phases? No, but as the following data show, they can still always be had on eBay for 50–90% cheaper than what you’d get them for in the app. Let’s plot the value of a few cards over the 4 value phases I outlined earlier in the article:
This is a real economy with big money being spent both in the app as well as in the dark secondary market for digital Bunt cards on eBay. Only time will tell how much these secondary market values equilibrate towards the full in-game cost of the cards; the better question is if they ever will fully.
As I pointed out earlier, eBay values seem to correlate very strongly to the in-game trade value of cards. I.e., if someone offers me two cards that sell for ~$10 each on eBay for a single card that sells for $20 on eBay, it’s likely the trade complete successfully from my experience. This begs the broader question: why do people buy so many cards in the app and then sell them at a value lower than their cost? This is one for psychologists and economists to answer, but I’ll venture a guess that there are three reasons. First, that the majority of cards for trade on eBay were pulled at low cost, as in, the user got lucky and pulled a card without having to spend a lot of coins. Second, that users don’t yet appreciate how much cheaper cards are on eBay relative to pulling them in the app, i.e., the eBay market does not yet have the liquidity of the in-app trading market. Third, and I think most obviously, that there is indeed some value that users derive from “playing the game” in the app. There is value in the experience of buying packs (which is a form of gambling), combined with some perceived value of the high-circulation and low dollar-value “base” cards that one gets in a pack and can play in fantasy games.
5. What’s Next?
Bunt has come a long way and needs to improve further. And Topps will need to figure out how to profit from the secondary market if values there begin to appreciate.
There are a few things I’d like to see changed in Bunt. The fantasy game structure is broken, in my humble opinion; 2014’s system was great, but in 2015 the introduction of “decks” that needed to be loaded in advance mean another barrier to entry (re: time), and so I haven’t played at all. This is why I was driven to collecting rather than playing.
The app is not all about fantasy sports. Topps makes a lot of money off the collectable cards like signatures and other inserts that offer no advantages in the fantasy games (the very cards that drive the secondary markets). If users lose trust in this aspect of the game, Topps loses. Could Topps take this to the next level by adding blockchain technology or something like it to create copies of cards that stand alone without an app, and could even be traded as raw data? Or, at least, create desktop support. A collector can dream…
Overall, Topps has created a new economy for a new good — digital trading cards — and, given the momentum in users and revenue that Bunt and Topps’s other apps are seeing in 2015, it seems like this economy may be here to stay.
