Executive Summary

Blockchain games, also called Gaming DAPPs, are an emerging area in the Web3 space. With their gameplay, Gaming DAPPs currently cannot compete with the 3D AAA games and they mostly resemble 2D Hyper Casual Mobile games. Developers often need to make compromises in separating on-chain activity from the actions that are taken off-chain and so the line is blurred between what is truly a blockchain game and what is not. However, strong advantage of these games is that they almost always offer some form of an economic incentive to the players with an opportunity to “own” part of the game in order to influence the game’s further development.

The most widespread model in such games is the “Play to Earn” (P2E) model. In a P2E model, players compete against each other in earning either ERC721 or ERC20 tokens. The mechanism of token distribution varies from game to game but in general it favors those that invest more into the game and so the “Pay to Win” element is introduced. This is currently the most popular model but it has negative connotations of lacking the play aspect and focusing on just the earning potential. Games now try to pivot to what they call “Play and Earn” model with the intention of promoting the community first but also keeping the earning part going. In either way, the economy of such games seems unsustainable as it relies on the never ending stream of new players.

Onboarding process is also almost always a bit complex and pricy. In order to start playing, the player initially needs to bare some cost that is not constant and is decided by the market (i.e. buying ERC721 tokens used in the game).

For the gamers to accept those low quality games, those tradeoffs and a pricy onboarding process there would need to be an economic incentive and whatever the future models will be called the earning aspect needs to be present.

Introduction

Gamers expect interactivity from the games they play. Once an action is taken in the off-chain games, the feedback signal is quick and the gamer knows nearly instantly the consequences of that action. Translating this experience to on-chain games is extremely difficult as every interaction that alters the state involves a transaction. Those transactions cost some amount of tokens native to the chain but they also take some time to process. This affects the whole game design process as the number of transactions that are required to advance the state in the game should be kept to a reasonable number. Having the player’s balance getting continuously decreased while playing the game and breaking the flow of it with the popup needed to sign each transaction will result in that player questioning the whole utility of that game. If the game doesn’t take this seriously into account it will lead to a bad user experience.

Due to current blockchain limitations, games try to cheat their way out of this by having the game logic on the server with players just signing the messages or transactions associated with their in-game actions. Splinterlands¹ is a popular game and an example of such practice. It is developed on the Hive blockchain² that does not support smart contracts and so the rules of the game need to be kept on the server.

Another successful game, Axie Infinity³ takes the approach of having the players register an user account that will be linked to their wallet’s address. Almost every action inside the game is done via the PC or a Mobile client where the player has logged in with their user account. The server is in control of this game’s logic as well, which poses a huge problem because the rules of the game can be altered by the developers.

One important part of the Axie Infinity’s economy relies on ERC721 tokens (called “Axies”) that are used inside the game. Axies can be traded on the Axie Marketplace and are needed when minting new tokens so some players focus purely on this aspect of the game. The main game’s mechanic are “Player vs Player” (PvP) matches. The winner of the match gets some amount of game’s ERC20 tokens that depends on the player’s rank. The higher the rank the bigger the rewards. To increase their rank, players need stronger and better Axies which leads them back to the marketplace. The similar model is used across the industry, like in Pegaxy⁴ which is a competitor game on Polygon that is currently gaining traction.

Because Axie Infinity was one of the first P2E games, is still relevant today and its solutions are being used in other games, this paper takes it as a study case, doing a deep dive into its inner workings.

Goals & Methodology

Being a pioneer in the space, Axie Infinity had to make some compromises. This resulted in combining the actions that are taken on-chain that include trading, minting and paying out the winnings with the off-chain actions like playing a match.

Having a hybrid or completely centralized approach defeats the purpose of a blockchain. The goal of this research is to understand the game’s mechanics and to try to answer the question if this and similar games can be built purely on-chain. It does this by breaking down the rules of the game and then proposing on-chain solutions.

In this research, there are sections dedicated to the basics of the economic model of the game. However, it is not the goal of this paper to delve deep into its problems and possible approaches to solving them. This model and its variations would require a separate research and the intention of those sections is just meant to provide context to the reader.

Results & Discussion

Ronin sidechain

The game was developed on Ethereum in 2018 by the Sky Mavis company located in Vietnam. Due to high Ethereum fees, the game’s creators moved to their own EVM compatible sidechain called Ronin which uses Proof of Authority with validators being chosen by the company⁵ .

To start playing a player needs to :

create their own Ronin wallet
create their user account on the Axie Marketplace and connect their wallet to it
transfer some amount of ETH and buy at least 3 Axies with a floor price of ~42 dollars per Axie
install a PC or Mobile app and login with their user account

Disregarding the difficulty of the onboarding process, the whole point of a blockchain game is to have the players take actions on the blockchain. Some arguments could be made that at the time the game was developed this was needed but with the introduction of Ronin it is unclear why the whole game was not ported to it.

The only actions that are taken on the Ronin sidechain are dedicated to trading - buying, minting and gifting of Axies. There are no fees on Ronin, but the number of transactions per wallet address per day is limited. Deploying on Ronin, also, requires company’s permission so the development of new games and the whole Ronin ecosystem is slowed down.

Scholarships

Market decides on the price of Axies and with a high entry cost of starting with the game, a new model of onboarding has emerged. Newcomers (scholars) can “rent out” the Axies for a certain period of time and negotiate with the owners (managers) the terms of the profit distribution. This is not a official in-game feature and is enabled by having the managers controlling the Ronin wallet and scholars controlling the user account associated with it. This leads to a bad position for the scholar as the profits are claimed by the manager and then the scholar’s share is sent to the their wallet’s address. Scholars are essentially at the mercy of the managers as their earnings and the scholarship itself can be revoked at any time.

More fairer way of enabling Scholarships, would be to have a smart contract that would have complete control of the Axies and through which the scholars would make in-game actions. The profits would go to the contract’s address that would perform a fair split. The terms of the agreement (the minimum amount of profits to be earned by the scholar) and the scholarship’s time period would be embedded in the contract. If both parties agree, the terms could be changed later on.

Game Breakdown

The game itself is organized in 1v1 matches where players do not have any influence against who they will get matched. Matching is done by the server based on players’ Match Making Rankings (MMR) which is determined by the win/lose ratio of those players. All of this could be implemented in a smart contract which would keep track of all of the players’ MMRs and update them after each match. Players could start a new match or join an existing one if the absolute difference in their MMRs is under a certain threshold. The benefits of this approach is that they could also choose what match they will join or challenge a specific player.

Winner of each match gets some amount of game’s “Smooth Love Potion” (SLP) tokens. That amount is dependent on the MMR of that player (the higher the MMR the more SLP tokens they will win). SLP is an inflationary ERC20 token that gets minted after each match. Even though there are SLP burning mechanisms through some in-game actions, most players opt to cash out their winnings so this might not be a viable economic model. One alternative would be to have both players stake some amount of tokens in a match with the winner taking the sum of those stakes. The problem is that it introduces betting connotations and games go to extreme lengths in order to not be considered a betting game as it potentially introduces regulation.

Inside the game, there is also “Axie Infinity Shard” (AXS) token. AXS is an ERC20 token which has a fixed total supply. The company behind the game has roughly 20% of the total AXS supply and small amounts of AXS is distributed to the top players of the month.

The game makes heavy use of Axies which have certain characteristics. Players do not completely own Axies as it was discovered that the company can freeze them, making them useless. Once an Axie is frozen, it cannot be used in the game nor it can be traded on the Axie Marketplace. This possess a major concern as the players’ assets are constantly under a threat of those players being banned from the game by the Sky Mavis company. Full list of violations that will result in a ban can be seen here⁶ .

Each Axie has one of the 9 classes, 4 statistics and 4 cards associated with it. Classes are grouped into three groups that form a “rock paper scissor” relationship. Meaning that, group G1 does 15% extra damage to group G2 but takes 15% extra damage when attacked by group G3. Statistics determine the Health, Skill, Speed and Morale of an Axie. This statistics affect the matches and their state transitions as the Speed for example determines the order of attacks. Cards can have positive or negative effects on an Axie as well additional effects that affect the players. Four cards associated with an Axie are added to the player’s deck when that Axie is used inside a match. More information about Axies is provided in Appendix A and the way new Axies are created is provided in the Breeding section of the Whitepaper⁷ .

Matches

One match consists out of:

Both players choosing a team of 3 Axies and deciding on their positions on the field (creating team formations)
Rounds being played out until one of the players doesn’t have a standing Axie

Picture 1 : Axie Infinity Match View

Team formations

The positioning of an Axie matters because it determines what Axie will take the damage from the enemy. Each player has 5 rows where an Axie can be positioned. The closest row to the enemy lines will be attacked first. The Axies cannot change their position and they stay where they initially were until they get knocked out. Choosing of a team formation would require two transactions per player. Those transactions would be organized in the commit-reveal scheme so that the players wouldn’t have an advantage against the opponent that naively sent the transaction revealing their team and their positions.

Rounds

Each Round is carried out in the following order:

Players randomly draw 3 cards* from their own decks (consisting of 24 Cards**) and decide on what cards they will play
Cards are revealed and their affects are applied to Axies
Battle of the Round takes place
If one of the players doesn’t have a standing Axie then the match is over

* Exception is that in the first Round, players draw 6 cards.

** There are 3 Axies per team, each Axie adds two copies for each of the 4 cards. So in total there are 3*2*4 = 24 cards in one player’s deck.

Drawing of Cards

Once the teams have been revealed both players know each other decks. However, they should not know what cards the opponent has in their hands and so they should not know what cards they have yet to draw.

One scheme that could be applied is the following:

Have both players choose their secret “random” number and commit to it with a hash
Have the game’s contract ask for a random number from Chainlink’s VRF⁸
Once the randomness has been fulfilled, both players know what is the order of cards they will be drawing as the seed for shuffling their decks will be some function of their own number and the received number from Chainlink’s VRF

The drawing order of cards in the deck is now known only to the players. During the match, each player can play any card inside their deck and until a match ends, players trust each other that the played card was in the opponent’s hand. However, when the match ends both players would need to reveal their secret numbers as the contract needs to verify if they honored the drawing order and if the card they played at a certain moment was one of the cards they were holding. The first discrepancy would end the verification process and the cheater would be penalized.

One additional subproblem is that after each round the remaining cards inside the deck should be reshuffled. This could be done with requesting another random number which would help form a seed for random shuffling of the remaining cards. The verification process would need to be modified to support this.

Note: after all of the cards have been drawn, the deck resets.

Reveal of the Cards

The reveal of the cards would also need two transactions per player so that one player wouldn’t just wait for the opponent to reveal their cards and then change their strategy accordingly.

Once cards have been revealed, before the battle takes place, the effects of the played Cards are applied. In total there are 3 constant effects (Attack damage, Defensive points and the cost of playing that Card) and 19 additional effects that a Card may have. For example, one of those additional effects are an increase/decrease in one of the Axie statistics. When multiple Cards that affect the same Axie are played, their effects are accumulated. More information about cards is provided in Appendix A .

Battle of the Round

Battle rules:

the order of attacks is determined by the highest Speed statistic (if there is a draw then it is decided by the lowest Health and then by lowest ID of an Axie)
closest row with a standing Axie to the enemy lines will take the damage
if two Axies are inside the row taking the damage there’s a 50/50 chance which Axie will absorb the damage

All of this logic could be kept inside a contract. One of that things that would need to be taken into account is that Axies attack in the pre-defined order. So during a Round, an Axie that has not yet attacked could be knocked out of the game and so the Cards associated with it, that were played in the Round, should lose their effects. Also, when an Axie gets knocked out, all of their cards should be removed from the deck.

Summary of proposed matched progression

Inital proposal

Step	Phase	Description	Number of transactions	Done by
1	Match Set Up	Starting of a Match	1	Player 1
2	Match Set Up	Joining a Match	1	Player 2
3	Match Set Up	Committing to a team formation and setting of a secret number	2*1	Player 1 and Player 2
4	Match Set Up	Revealing team formations	2*1	Player 1 and Player 2
5	Round	Committing to Cards that will be played in this Round	2*1	Player 1 and Player 2
6	Round	Revealing Cards of the Round	2*1	Player 1 and Player 2
7	Match Wrap Up	Revealing secret numbers after the match has ended	2*1	Player 1 and Player 2

Total = (1+1+2+2) + N*(2+2) + (2) = 8 + 4*N ; where N is the number of rounds in the match

Even with just 3 rounds the number of transactions per player would be 16, so this is a problem.

Improved Match Progression

Something that could be done is the introduction of “fair play” where for one round’s commit stage, one player would transmit both of the commit messages to the chain after which the other player would transmit both reveal messages.

If the player transmitting commitment messages decides to not transmit them, then the match is stuck. However, the game can be structured in a way such that it is in the interest of both players to resolve the match regardless of the outcome. This could be done by staking some amount of tokens that the players will receive at the end of the match.

If the player transmitting the reveal messages realizes he will lose and decides to not transmit them, then the first player could transmit just their own reveal message that would start a countdown for the other player to transmit their reveal message. If the countdown is reached then the player that didn’t act in the spirit of fair play is penalized in some way.

Step	Phase	Description	Number of transactions	Done by
1	Match Set Up	Starting of a Match	1	Player 1
2	Match Set Up	Joining a Match	1	Player 2
3	Match Set Up	Transmitting commitments to the team formations and setting of a secret number	1	Player 1
4	Match Set Up	Transmitting the Reveal of team formations	1	Player 2
5	Round	Transmitting Commitment Messages	1	Player 1
6	Round	Transmitting Reveal Messages	1	Player 2
7	Match Wrap Up	Revealing secret numbers after the match has ended	1	Player 1

This scheme would give the total number of transactions of:

Total = (1+1+2) + N*(2) + (1) = 5 + 2*N ; where N is still the number of rounds in the match

With the number of rounds equal to 3, the number of transactions per player is less or equal to 6.

This is an improvement but maybe it could be brought down further. Also, signing 6 transactions per match would still break the flow of the game, so the players would have to have a substantial financial interest in continuing to play the current match and the game itself.

Conclusion

Developers should keep realistic expectations on what games can be built purely on the blockchain and they should convey those expectations transparently to the end users. Proposed on-chain solutions from this research can be improved and adjusted to fit similar games. The current bottleneck of all Gaming DAPPS is the number and speed of transactions needed, so the choice of a blockchain on which the games will be built, is extremely important. Hybrid approaches, in my opinion, should not be considered, but at the end of the day, the players will decide what games will be played. The most important thing is that the rules of the game cannot change by the will of the developers. The game’s logic should be kept in the smart contracts with the source code being public. This will keep everyone honest and the players would not be at the mercy of the developers.

Games are meant at meant to be fun but players ought to have an additional financial benefit from blockchain games because at this moment their user experience is hindered by the current blockchain limits as well as the lower quality graphics and player involvement. Everyone involved in the P2E model should view it as an investment first. Investments can be either good or bad and they depend on every party involved contributing to the end result. Once the investment has turned bad, investors should cut their losses if there is no indication that the situation will change. In this context that would result with the players quitting the game and moving to other games which leads to a bad word of mouth and it damages the whole ecosystem. A sustainable economic model needs to be embedded in the game itself so there should not be unbound minting of new tokens. However, how exactly should that future model look like would require serious thought and it should be a topic of a separate research.

Appendices

Appendix A

Each Axie has a fixed set of attributes - a class, 4 stats and 4 cards.

Picture 2: Axie characteristics

Axie Classes

9 of them (Plant, Reptile, Dusk, Beast, Bug, Mech, Aqua, Bird, Dawn)
6 main classes (Plant, Beast, Bug, Reptile, Aqua, Bird, Dawn)
3 remaining are “secret” (and generally weaker in the game)
to make the strengths of classes balanced, they are grouped in 3 groups, forming a “rock paper scissor” relationship

Axie Stats

can be divided into “base” and “additional” Stats
total of 140 points is distributed between 4 of the base Stats (with each class having its own base distribution)
- Health — amount of damage Axie can take before getting knocked out
- Speed — affects the order in which Axies attack in a match (higher Speeds attack first)
- Skill — increases damage dealt when the Axie performs multiple cards/moves (a.k.a. combo)
- Morale — increases chance to land a critical hit, as well as entering “last stand” which allows them to attack a few more times before getting knocked out

Base Stats Distribution

Class	Health	Speed	Skill	Morale
Aqua	39	39	35	27
Beast	31	35	31	43
Birds	27	43	35	35
Bug	35	31	35	39
Plant	61	31	31	41
Reptile	39	35	31	35

additional Stats depend on Axie’s Body Parts

Axie Body Parts

there are 6 of them (Eyes, Ears, Horns, Mouth, Back, Tail)
only Horns, Mouth, Back, Tail have an associated card with it (one card per body part)
In total, there are:
- 4*6 types of Mouth
- 6*6 types of Horns
- 6*6 types of Back
- 6*6 types of Tail
inside groups of those types each of the 6 main classes is equally represented
if the class of an Axie matches with the type of a body part, Axie’s stats will increase

Additional Stats Increase

Class/Type	Health	Speed	Morale
Aqua	+1	+3	+0
Beast	0	+1	+3
Birds	0	+3	+1
Bug	+1	0	+3
Plant	+3	0	+1
Reptile	+3	+1	0

An Axie is a ‘pure breed’ when its class and all of its body parts are of the same type.

Axie Abilities (Cards)

there are 132 cards in total (first divided by 6 main classes and then by the body parts)
each Axie has 4 cards associated with it
each Card has
- an amount of Energy it costs to play it
- attack points - damage it does to the enemy Axie
- defensive points - forms a shield that takes the damage instead of Axie’s Health
- potentially buffs/debuffs
- additional effects (draw another card, steal some Energy from the opponent,…)
cards realize a “combo” when 2 or more of them are played (for the same Axie)
cards realize a “chain” when 2 or more Axies use 2 or more cards that are from the same class

Axie Card Buffs/Debuffs (Card Effects)

there are 3 buffs (positive effects on the Axie) and 16 debuffs (negative effects on the enemy Axie)
they affect the Axie for one or more turns
they are “stackable” - their effects are accumulated

List of Buffs

Name	Description
Attack+	Increase next attack by 20%
Morale+	Increase Moral by 20% for the following round
Speed+	Increase Speed by 20% for the following round.

List of Debuffs

Name	Description
Aroma	Target priority changes to this Axie until the next round.
Attack-	Decrease next attack by 20%.
Chill	Affected Axie cannot enter Last Stand
Fear	Affected Axie will miss their next attack
Fragile	Affected Axie’s shield will take double damage from the next incoming attack
Jinx	Affected Axie cannot land critical hits
Lethal	Next incoming attack is a guaranteed critical strike
Morale-	Decrease Morale by 20% for the following round
Poison	Affected Axie will lose 2 HP for every card used
Sleep	Next incoming attack will ignore shields
Speed-	Decrease Speed by 20% for the following round
Stench	Affected Axie will lose target priority for the following round
Stun	Affected Axie’s first attack will miss.Next incoming attack will ignore shields
Cannot Be Healed	This Axie cannot be healed or recover health. This Debuff cannot be removed

Bibliography

‘Axie Infinity’, Axie Infinity <https://axieinfinity.com/> [accessed 24 February 2022]

‘Breeding and $SLP’ <https://whitepaper.axieinfinity.com/gameplay/breeding> [accessed 13 February 2022]

‘Introduction to Chainlink VRF | Chainlink Documentation’ <https://docs.chain.link/docs/chainlink-vrf/> [accessed 13 February 2022]

‘Pegaxy Summary’ <https://whitepaper.pegaxy.io/> [accessed 24 February 2022]

‘Ronin Ethereum Sidechain’ <https://whitepaper.axieinfinity.com/technology/ronin-ethereum-sidechain> [accessed 14 February 2022]

‘Terms’, Axie Infinity <https://axieinfinity.com/terms/> [accessed 10 February 2022]

‘The Splintershards (SPS) Token’ <https://sps.splinterlands.com/> [accessed 24 February 2022]

Web3, Hive-The Blockchain & Cryptocurrency for, ‘Hive - The Blockchain & Cryptocurrency for Web3’, Hive - The Blockchain & Cryptocurrency for Web3 <https://hive.io> [accessed 24 February 2022]

Footnotes

‘The Splintershards (SPS) Token’ <<https://sps.splinterlands.com/>> [accessed 24 February 2022].↩︎
Hive-The Blockchain & Cryptocurrency for Web3, ‘Hive - The Blockchain & Cryptocurrency for Web3’, Hive - The Blockchain & Cryptocurrency for Web3 <<https://hive.io>> [accessed 24 February 2022].↩︎
‘Axie Infinity’, Axie Infinity <<https://axieinfinity.com/>> [accessed 24 February 2022].↩︎
‘Pegaxy Summary’ <<https://whitepaper.pegaxy.io/>> [accessed 24 February 2022].↩︎
‘Ronin Ethereum Sidechain’ <<https://whitepaper.axieinfinity.com/technology/ronin-ethereum-sidechain>> [accessed 14 February 2022].↩︎
‘Terms’, Axie Infinity <<https://axieinfinity.com/terms/>> [accessed 10 February 2022].↩︎
‘Breeding and $SLP’ <<https://whitepaper.axieinfinity.com/gameplay/breeding>> [accessed 13 February 2022].↩︎
‘Introduction to Chainlink VRF | Chainlink Documentation’ <<https://docs.chain.link/docs/chainlink-vrf/>> [accessed 13 February 2022].↩︎