How PucaTrade Manages Point-Supply Data, Theory, and Observations
Over the last five years, I’ve been perpetually exploring the question: “How many PucaPoints should be in the PucaTrade economy to maximize the amount of exchange that can occur?” In pursuit of an answer, I’ve interviewed economics professors, commissioned consultants, scrutinized data, and read everything I could find about the theory of “money supply.”
To anyone familiar with macroeconomics, the answer will be unsurprising: “It depends.”
What I’ve come to learn is that there is no one “silver bullet” equation, metric, or variable that can answer this question in isolation. However, there are many key metrics that should be considered to keep us on the path of responsible stewardship of this economy.
Before we dive in, let’s take a step back and ask, “Why does it matter?”
PucaTrade’s goal is to put as many Magic cards into the hands of as many players as possible, and the number of PucaPoints in the system plays a role in the maximum number of cards that can be traded. Having too few points in the system would be undesirable, because it would be difficult or impossible for an individual to accumulate enough points for high-value trades to occur. Additionally, if supply of points is too low (and demand for points is too high), it would be more difficult to find a trading partner due to increased competition on the Send Page.
Having too many points in the system would be undesirable for the opposite reason; if supply of points is too high (and demand for points is too low), people will be less incentivized to send cards.
So, the goal is to find the “sweet spot” where the number of points in the system supports the maximum amount of trading in terms of quantity of cards traded, value of cards traded, and speed at which cards are traded (from the moment they become visible on the Send page).
Velocity of Trade
As a Magic player, once I’ve designed a new deck, I get pretty excited to build it and take it for a test-drive. Most of us love to get cards as quickly as possible. As a member of PucaTrade, the amount of time it takes for a trade to be confirmed from when I add it to my Wants List (the velocity of trade-initiation) is an important metric to me when determining how useful I feel PucaTrade is.
As the founder of PucaTrade, sometimes I get feedback that sounds like this, “Polluted Delta is taking forever for me to get! You must have printed too many points!” While this is important feedback, it’s also an oversimplification of something much more complex.
Many Forces Influence Trade Velocity
There are many factors that influence the speed of trade-initiation. On an individual level, having an incomplete profile (or even one that makes you look like you’d be a bad trading partner) can definitely have a negative impact on your incoming trade velocity. Other factors include the number of your Wants, your PucaPoint balance, your Pro status, the country you live in, and even luck (remember, PucaTrade isn’t intrinsically first-come-first-served).
Market dynamics in the greater sphere of Magic, such as upcoming set releases, out-of-print cycles, banned & restricted announcements, reprints, spoilers, and price fluctuations can all have a major impact on global exchange speeds as well.
Chart 1.1 illustrates the velocity of trade for two groups of five cards over time: Khans of Tarkir fetchlands (in red) and five other Standard staples (in green).
Chart 1.1: Velocity of trade-initiation in days over time on PucaTrade.com (higher on the y-axis is slower).
Since September of 2014 when KTK was released the velocity of trade for the Standard staples has consistently hovered at around 1-3 days. The fetchlands’ velocity was similar for the first few weeks of release, then basically held steady at a velocity of 6-10 days until the moment they went out of print in late July of 2015. At that time, their velocity slowed considerably over the course of the following year. This, of course, is not surprising at all. As Magic players stopped drafting and cracking packs of the recently out-of-print Khans, the KTK fetchlands already in the ecosystem found their homes and their velocity slowed down accordingly. Meanwhile, the other Standard staples continued to trade at their usual rate.
If you were to view the velocity of fetchlands in isolation (or if they were the only cards on your Wants List), it may seem reasonable to assume that all trade on PucaTrade is slowing down. For this reason, it's critical to compare the fetchlands velocity to the respective velocities of other cards as well, and to consider the larger context of market dynamics.
A quick note about these velocity metrics: this data is based on the date that a card was added to a Wants List compared to the date the trade was initiated. The metric does not take into account the number of points that the wanter has at the moment of clicking the want (thus the eligibility of the trade). A card that sits invisibly on someone’s list for 30 days before they earn enough points to afford the trade initiated on the 31st day (after becoming visible) would contribute to the average as 31 days (not 1). As such, this data shouldn’t be interpreted as an exact expectation for the amount of time it should take to get a given card; it’s more valuable as a tool to understand how velocity changes over time.
To understand how point-creation is influencing the PucaTrade experience, it's more productive to consider general trends than to analyze individual card movement. Chart 1.2 below shows the monthly velocity of 10 Legacy staples. The period of erratic speeds from December 2013 to December 2014 illustrates both of the downsides to having too few points in the system (mentioned above).
Chart 1.2: Velocity of trade-initiation on PucaTrade.com. Notice the period of volatility followed by relative stability beginning in 2015.
The consistent baseline of zero days illustrates the high degree of competition on the Send page, which resulted in immediate trade-initiations (represented by an average value of 0 days). Conversely, the erratic peaks of 40+ days illustrate the long periods of time that a high-value card like Wasteland might have stayed on a Want List before a receiver earned enough points to afford it. This is consistent with most of the segmented velocity trajectories we’ve seen to-date, particularly for higher-value cards as seen in the parallel chart for Modern staples, shown below in Chart 1.3.
Chart 1.3 Velocity of trade-initiation on PucaTrade.com for 10 Modern staples.
This pattern is also visible when observing the average velocity of all cards traded on PucaTrade over time.
Chart 1.4 Average Velocity of Trade for all cards on PucaTrade.com (higher y-axis is slower). The period of time between October 2014 and August 2015 demonstrates an especially high degree of both velocity and stability, thus contains excellent targets for point-supply levels.
To fully understand the implications of these charts it’s important to consider the economic climate of PucaTrade in the year of 2014 and prior, when there were relatively few points in the system. Anyone who was a member of PucaTrade at that time will recall how competitive the Send page was, and how difficult it was to earn points as a result. During this time we received a fair amount of feedback from our members describing how challenging it was to find cards to send. The hashtag #SurvivalOfTheClickest was often used in reference to how quick you needed to be to initiate a trade before it got snagged by another trader.
During that time, the observation that trades were snatched up extremely quickly was frequently misinterpreted as a low-demand for Magic cards, especially by new PucaTraders who expected the Send page to feel similar to an online retailer’s shopping cart. Of course, there has never been a moment since 1993 when Magic players did not want infinite Magic cards (and certainly not in the history of PucaTrade.com). Demand for points was too high because supply of points was too low.
How PucaPoints Influence Velocity
In February of 2014, we launched the “Give and Let Give” bonuses that would allow a new member to earn 700 points for signing up and completing their profile. It’s important that point-creation events occur in parallel to the growth of our community and the total value of cards on Haves Lists because these three metrics are all correlated to the demand for points. If this sounds obvious, you may be surprised by how many people misunderstand this fundamental economic reality. In fact, misconceptions around this are so prevalent that it’s worthwhile to go a little deeper.
I often hear people say “Every time you print new points, you’re de-valuing my points! It’s obvious! If there exists 1,000 points and I have 100 of them, I have 10% of the points. If you create another 1,000 points, I’ll only have 5%. The value of my points will have decreased by half!” The crux of this misconception is that A) The value of a PucaPoint is relative to the total number of points in the system, and that B) The value of a PucaPoint is not influenced by the total number of people who want them. The implied hypothesis that “the number of points in the system should always remain fixed” does not account for increased pressures in demand due to new traders entering the system.
In the hypothetical situation above, what happens when the number of members double? The demand for the fixed quantity of 1,000 points would increase, and competition to earn them would increase in parallel. What if the number of members increases by 10x? It’s clear that the same 1,000 points would not be sufficient to service the needs of the new larger trading community.
It’s far more valuable to consider points per capita as a key metric for understanding point-supply. By pegging point-creation events to new member registrations, we’ve successfully ensured that the points per capita on PucaTrade remain relatively stable, as demonstrated by charts 2.1 and 2.2 below.
Chart 2.1 PucaPoints created each day are closely tied to new member registrations.
Chart 2.2: Fixing point-creations events to new member registrations has a stabilizing effect on our points per capita metrics. Recall that the average velocity of trade was at its peak between October 2014 and August 2015. As such, 700-900 points per capita is a great target range for us to preserve.
After witnessing positive results from the launch of the new-member bonuses in February of 2014, we launched the referral bonuses in October of 2014. Note how the increase in point-creation late in 2014 aligns with the velocity trajectory in the following months (as illustrated in Charts 1.2 - 1.4). Figure 1.4 shows how the speed of trade immediately stabilized at around 5 days on average, and began gently trending upwards in the following months. Concurrently, the number of cards traded per day and their corresponding values were increasing in tandem, demonstrating that the increased levels of point-creation succeeded in the goal of maximizing trade as illustrated in chart 2.3 below.
Chart 2.3 Trades initiated each day increased from 2,500 in early 2014 to 15,000 per day in early 2016.
Is the average value of traded cards increasing or decreasing?
Quality is as important as quantity. To ensure that an even distribution of cards of many values are being traded on the site, it’s important to consider average values of cards traded each month as well. Chart 3.1 illustrates that the average value of cards traded on PucaTrade has consistently hovered around 375 points throughout PucaTrade’s history.
Chart 3.1 The average value of cards traded on PucaTrade has remained constant, reflecting an even distribution of cards traded.
Gauging Demand: How many PucaPoints are in escrow?
Every time a trade is initiated, points that were previously held by a receiver are moved into escrow until the card arrives and the trade is complete. In that way, points in escrow can be considered points “in use.” It’s important that points we print are being used as opposed to surplus points that could potentially contribute to a disproportionate decrease in demand. 33% is our traditional average and a great benchmark for the points in escrow.
Chart 4.1: As new points enter the system, consistent numbers of points in escrow demonstrate consisent demand for them.
How does the site feel?
Earlier we touched on the limitations of anecdotes, understanding that there are many factors that contribute to velocity metrics. Even so, the general feel of the site plays an important role as one data point among many, because ultimately that’s the primary metric available to our site-members. We listen to feedback on this topic, and consider how it matches up with our own experiences trading on the site. This is one of the many reasons it’s helpful that everyone on the team is an avid Puca trader as well.
Silver Bullets and Sweet Spots
Basing our point creation policies around data and solid macroeconomic principles is crucial for the longevity of this community. Observing and reacting to the results of these policies is vital as well. Understanding how points are being used by members of our community is a critical component of our economic strategy; it serves as a litmus test to ensure we’re on the right track.
I started this article by claiming that there was no “silver bullet” equation or formula that could accurately predict optimal point supply. However, analysis of the extensive data we have access to indicates that the best target to enable the maximum number of trades and the highest velocity of those trades is 800 points per capita.
In all of the velocity charts, the period of time between October 2014 and April 2015 represents the first period of time in PucaTrade’s history when trade-speeds didn’t fluctuate with a high degree of volatility. During that time period, the average velocity held steady at 4.48-4.82 days. Anecdotally, October 2014 was the month that we stopped receiving feedback from traders about how excessively competitive the Send page was, and was also the last month the hashtag #SurvivalOfTheClickest was used.
As such, our current goal is to gradually reduce the number of points per capita from 950 to 800 over the course of the next 6 months. To achieve this, our we'll be reducing the new-member bonus from 700 points to 350 points per new member beginning one week from today on 5/11/2016. Given that roughly 200 people register for accounts every day on average, we expect to reach our target by October of this year (discussion of this policy update is happening here).
One of the strengths of the PucaTrade economy is that it’s a flexible system. As the caretakers of this incredible network of Magic players, every decision we make on Team Puca is executed with the longevity of the community in mind. In any new community the early days are the most volatile and tumultuous, but we've perservered through those days and are now settling in to a period of sustained growth. With almost 5 million trades and counting, we’ve proved that this ecosystem for exchange is resilient, vibrant and burgeoning. With the upcoming release of Magic Online and non-English cards, we’re poised to welcome Magic players from all over the world as we secure the title of most popular network of Magic traders on earth. It truly is a joy to serve you in the process.
Eric Freytag is founder and CEO of PucaTrade.