Testing Hypotheses Around Token Economies
In a previous deconstruction, we looked at the token and economic systems of a variety of games, applying a “smoke test” to look for potential sustainability problems. That test involved making an assumption that the number of players remains constant (with new and churning players in equal amounts) and seeing if the economy can sustain.
Part of that analysis involved creating flow charts to roughly model how tokens and value moved through the system. In this process, we came across a tool called Machinations that is designed to visually model game economies through functional flow charts. This seems like a perfect opportunity to try the tool out and challenge our theories with a dynamic model.
We’re not going to go through the derivation and explanation of each of our own flow charts here – please check out the previous deconstruction for details there. Instead, our focus is going to be on testing hypotheses and seeing how shifts in the assumptions impact the system.
As we will see, Illuvium’s ILV token has some advantages over Splinterland’s SPS token, largely thanks to its separation from gameplay and its redistribution of revenue cycle. Splinterlands can still find success, but it will be more challenging and, at the moment, it’s not going well.
As before, this is the key for the elements in our flow charts.
Disclaimer: Naavik did not receive any compensation for this review and this is not an endorsement. Machinations did however give us free access to the tool for the purposes of writing this paper and provided technical assistance with setting up the diagrams.
Second Disclaimer: The values in the diagrams are for illustration only. We are not projecting the amount of spending by players or representing the actual staked values or prices.
Teaching the use of Machinations is outside the scope of this essay, but a very quick introduction to see what the different parts are will be helpful to understand the charts we produce. At its core, Machinations converts flow charts into dynamic systems. There are a few basic types of nodes and connections:
- Pools – Represented as circles, these can collect and hold resources.
- Sources – The upward triangles are sources that produce (mint) resources.
- Sinks – Conversely, downward triangles destroy (burn) resources.
- Resource – Solid lines can carry resources from one node to another
- State – Dashed lines will influence other values. So instead of transferring resources, a state connection could set the rate of a resource connection to a different value.
In the above animation, a source produces resources and sends them to the two pools. You can see the value on the top connection starts at 2, so two resources are sent across. The state connection has a +1 in it, which means it adds one to the value it’s pointing to every time it triggers (in this case, every time the pool it is coming from receives a resource). This makes the number of resources created increase over time, and the top right pool collects them, even as a slow drip heads to the sink.
Machinations has much more functionality than this, but understanding these elements will be enough to look at the diagrams in this essay. In many cases there will be background calculations in the diagram that we will hide for the sake of readability. If you wish to see all of the details you can make a copy of the live diagram and unhide the calculation layers.
Illuvium’s ILV Token
Illuvium is one of the most anticipated AAA-tier blockchain games. It combines land-based farming, which produces Fuel token, with a 3D overworld to explore and strategic autochess-style battles. The primary NFTs are land and Illuvials – the creatures used in battle.
Illuvium’s governance token, ILV, is separated entirely from gameplay. Instead, it functions as a way of both establishing voting power and redistributing earnings to Stakers. Specifically, Players will be spending ETH for a variety of in-game items and features (Fuel, Land, IlluviDEX transactions) and all of this ETH is collected by the Illuvium Vault. The only other fungible tokens are the Fuels, which are uncapped and minted by both players and the IlluviumDAO.
The vault then periodically spends that ETH to purchase ILV from the Sushi Liquidity Pool, and distributes that ILV to Stakers. Stakers can then choose to sell ILV back to the Liquidity Pool for a profit. In the diagram below we removed the Fuel Balancer Pool and just represent player spending as a single connection to the Vault.
Colors in Machinations have a specific function to track the type of resource being used. So instead of green for sources and red for sinks, we’ll be using green for ETH and blue for ILV. Sources and sinks are identified as the triangle nodes.
ILV Diagram 1
In this model, the players are spending 200 – 300 ETH per step. The vault sells the ETH to the Liquidity Pool at the current ETH/ILV exchange rate, which is updated each step based on the supply of each. The Vault then sends its ILV to the Stakers, who in turn sell it to the pool. We’re assuming Stakers are not selling any staked ILV (30k worth), but are selling 100% of the distributed ILV.
The ETH/ILV price (seen in the black square) remains reasonably steady throughout this process, and does so even if we adjust the player spending up or down. Of course if spending goes down the Stakers will earn less, but the currency value would not inherently collapse.
But what if the Stakers get scared and start to unstake and dump their ILV? In this version of the model, instead of keeping an initial 30k staked, they start selling off 10% of the total amount of ILV every step.
ILV Diagram 2
As one would expect, the exchange rate in the pool quickly drops from the initial 0.3 ETH / ILV due to the sudden influx of ILV. However, it starts to stabilize around 0.16 ETH / ILV, even with the Stakers continuing to sell off 10% of their holdings every step.
You can see this stabilization point by opening the chart at the bottom. The equilibrium happens when the amount of ILV being sold by Stakers roughly equals the amount purchased by the Vault.
So while a drop necessarily occurred due to an influx of ILV, the system then stabilized. Not only that, but also all of the player ETH value continues to flow smoothly to the Stakers and player spending would be unaffected by the shift in ILV price.
The system seems to handle a dump reasonably well by finding a stable point before too long and then maintaining from there.
What can we learn from this? First, by separating the ILV token entirely from the game economy, its value doesn’t depend on in-game functionality. While the developers must continue to live-service the game so that revenue can be consistent, they have abstracted out their token, giving them more flexible control over the game’s economy in a way similar to that of modern F2P game, with the uncapped Fuel tokens functioning as a “hard currency.”
Additionally, by having the Illuvium Vault act as an intermediary rather than giving ETH directly to Stakers as a revenue share, ILV transaction volume will be much higher, as will market upward pressure on the token value. It’s possible that direct ETH distributions would maintain the token price due to these dividends, but Vault purchases will ensure that this is the case. Again, though, that is only if they can maintain a steady income of ETH from participating players.
Splinterlands is a veritable veteran of blockchain games, having started back in 2018. In this card-based game, players battle by selecting a set of cards to meet randomized criteria, and the two teams then battle it out automatically. It has remained quite popular, recently attracting around 330,000 unique monthly wallets.
In the above-mentioned deconstruction, we looked at Splinterlands’ Splintershards (SPS) token and argued that it was unsustainable in its current form. In the flow chart below, we identified two temporary sources (rewards pools) that would feed into the system. SPS has governance properties, but beyond that the only significant utility is purchasing certain limited-edition card packs from the shop. These packs are dynamically priced at $5 USD worth of SPS.
As players collected more of those cards, the demand for, and thus spending on, the cards has reduced, which we saw supported in the data. The result is a continued increase in the supply of SPS with minimal demand, leading to progressive sell-off and decay in value.
We have recreated this model in Machinations to illustrate and test this analysis. This is a more complex system, so we have made a few assumptions:
- We did not include the SPS —> DEC Converter since that would be rare (only if DEC value rises significantly).
- We have a fixed number of card packs in the shop, and as players collect more of those packs there is less demand for new ones. This is modeled linearly with a max demand of 1,000 packs in a step, down to 0 once players own 100% of the packs. Packs are priced at 6 SPS.
- Stakers and the DAO want to maintain some ownership. As such, they only consider selling what they have above 20,000 ILV, and the DAO only considers above 50,000.
- Desire to sell SPS to the DEX is linearly interpolated with a maximum selling 10% of excess ILV when the SPS price is above $0.50 and a minimum of 2% sold when below $0.10.
With all of that said, we can play the model and see what happens to the price of SPS.
SPS Diagram 1
When you press play, you will see players spending SPS on card packs, which goes to the Shop and then the DAO, as well as selling to the DEX. They are being funded by the Reward Pool at the top. The Stakers are also funded by a Reward Pool and are selling portions to the DEX as well.
The player demand for cards reduces over time, and the price of SPS continues to sink, even after the Reward Pools are empty. Outside of potential speculation, there is simply no upward pressure in the system to reverse the change in SPS pricing.
What Splinterlands needs is some way to get value flowing into the system. As you may remember, SPS did see a spike when a new card set that was exclusive to SPS was released. So what happens if Splinterlands were to release another card set?
The diagram below is identical to the previous one with one exception: after 25 steps, another 10,000 card packs are added to the store, and a third set is released after 50 steps. This happens through the time delay nodes on the right, which are depicted by a circle with an hourglass in it.
SPS Diagram 2
When you run this diagram and look at the chart, you will see SPS price drop just like before. However, when the new cards are added to the shop, the spending on cards jumps significantly. By that point the SPS reward pools have emptied and players soon run out of SPS. Additionally, if players want to continue to obtain new cards, they have to buy SPS from the DEX, which is what happens. With the new inflow of USDC, the price of SPS starts to slowly recover, for at least as long as demand for the cards remains.
You may notice a sinusoidal pattern of the SPS price — this is due to players needing to purchase the SPS on one step and then spend it on the next. That two-step process creates alternating spending, which Machinations smooths out into what looks like the oscillations of a diminishing sine wave.
What this suggests is that if Splinterlands can continue to drive a sink for SPS from players, likely through new sets of cards, SPS may be able to sustain or even grow in value. In other words, a steady stream of new, compelling content is required, and with that players are more likely to spend money and support the game’s economy.
We did run one additional experiment. This diagram is identical to the one that releases more card packs over time, but with one exception: instead of being priced at $5 USD worth of SPS, they are priced at a constant 20 SPS. Additionally, players have increased interest in buying packs when SPS is below $0.25, which is the $5 pack price point, and reduced when it is above.
SPS Diagram 3
We were a little surprised to see that the general behavior and ending price of SPS was similar, though we think some additional experimentation would reveal larger differences in some situations. Even here, though, we do see players spent a bit more USDC on packs in the fixed $5 price version, but the difference is fairly small so we hesitate to extrapolate too much from it.
Our biggest takeaway here is that continued new utility needs to be added to a token that is integrated into gameplay, lest the value fall. Splinterlands is certainly adding content over time, but so far very little of that has been priced in SPS. As we discussed in our earlier deconstruction, they have made real progress on recovering their DEC token value, but SPS reward pool inflation and the lack of utility have caused the price to drop.
Additionally, the continued investment value for Stakers is much less clear. They get to vote on the DAO’s behavior and may be able to redistribute revenue from the DAO, but there’s a lot of uncertainty there even by crypto investor standards.
ILV and SPS differ in two primary ways:
- ILV is entirely separate from game utility while SPS can be used to purchase in-game items.
- ILV is automatically and regularly redistributed to ILV Stakers, while spent SPS goes to the Splinterlands DAO with an unknown final destination.
These two differences make ILV a much cleaner investment token, giving the Illuvium team more flexibility with the game economy and ILV Stakers a clear and sustainable source of payouts. Meanwhile, Splinterlands is seeing its SPS token fall in value because it hasn’t offered new content to be purchased with it.
We are not saying Splinterlands’ system is fundamentally unsustainable. If it is able to continue to create new content that players want to spend SPS on, it should be able to bring the value back up.
However, ILV’s value will likely track the success of Illuvium as a whole, while SPS could fall in value even if Splinterlands is showing user growth and steady revenue. If nothing else, we would say Illuvium has set up a system that is easier to maintain, and in the world of blockchain economics that can be a real advantage.