CryptoKitties is one of the most popular games in the Blockchain network. It is based on the smart contracts of the Ethereum blockchain network. An example of a collection game with kitties has shown the value of ERC-721 tokens to the community and the possibilities of Ethereum as a playground.
In this research a team of analytics from Chaingraph.io attempted to conduct an extensive analysis of the decentralized application CryptoKitties, using the transactions extracted from the Ethereum network. The main analysis tool used is the smart contract analytical platform Chaingraph.io.
It is also worth adding that analytics is not just about the analysis of transactions, but also an attempt to understand and predict user behavior. The basis and purpose of the analysis is business, and the analytics should work for it. That is why many solutions from this report can be used for other decentralized applications of the crypto-game business sector.
For this study, the key issues were the increase of the audience for the DApp on the part of the developers and making a profit from the gameplay on the part of the player.
The contract that was analyzed is the main KittyCore contract
(0x06012c8cf97bead5deae237070f9587f8e7a266d). We also used the data from the SaleClockAuction (0xb1690c08e213a35ed9bab7b318de14420fb57d8c) and SiringClockAuction (0xc7af99fe5513eb6710e6d5f44f9989da40f27f26) contracts.
Besides Chaingraph.io, the following services were used: Opensea.io, Bloxy.info and Dappradar.com. Data may have some inaccuracy.
Market Cap: 40,000 ETH
Number of users (7d): 1,300
Number of users (24h): 450
Number of token holders: 70,000
Number of tokens in circulation: 1,500,000
Turnover: 57,000 ETH
Turnover (24h): 130 ETH
Number of smart contracts: 4
Twitter followers: 27,000
Each kitty is a unique token that belongs to one holder. The uniqueness of a kitty is defined by its characteristics, or “cattributes”. The color of the paws, the shape of the whiskers, the length of the tail – all of this depends on the attributes obtained at birth. Besides the normal kitties with a set of certain characteristics there are also unique ones that look “not like everyone else” and possess exclusive arts/traits
Normal – ordinary cats with a number of standard characteristics
Fancy – unique cats with a limited number in circulation.They can be created by cross-breeding using the right combination of genes at a certain time (alchemy). Such cats do not possess standard characteristics
Exclusive – these cats exist in the game only in one copy and they can not be bred. Every Exclusive cat is specially introduced into the game by the developers. Do not possess standard characteristics.
Special Edition – similar to Exclusive Cats in that they have unique art, but they are released in larger numbers and may be bought. Also do not have standard characteristics.
There are also some additional unique features:
Prestige is a special characteristic cat obtained by cross-breeding by certain recipe.
Wild is a special wild element
Genetic Anomaly happens when the cat has a hidden gene that can suddenly appear in offspring
Average price for normal cat: 0.1 ETH
Average price for fancy cat: 0.26 ETH
Average price for exclusive cat: 32 ETH
Average price for special cat: 0.4 ETH
Average price for prestige cat: 0.17 ETH
Average price for wild cat: 0.07 ETH
As you can see Exclusive cats are ahead of the rest. And the wild characteristic, on the contrary, is not decisive in defining the price for normal cats. It can be concluded that, despite its rarity, the wild characteristic is not very valuable.
Cats are cross-bred and have such an option as a generation.
As you can see, the cats of the latest generations belong to one and the same player nicknamed Solidox, who is really far ahead. It looks like he wouldn’t be overtaken soon, since with the birth of nextcat’s generation, the cooldown after the breed-crossing increases. That is why the one who first has occupied a niche and is not going to yield it, is hardly probable to let somebody in.
Now let us consider the dependence of the average selling price on the generation of the cat, and also show the number of cats in each group.
Dependence of the number of cats on the generation
As can be seen, after the 20th generation the number of kitties is significantly reducing, and with a three-digit generation their numbers tend to units.
Now let’s see the dependence on the selling price.
Dependence of the medium selling price on the cat generation
As you can see, after the 1000th generation the sales of cats are on the decline. Does it mean that after a fixed number nobody purchases? Let us study the sales of kitties of maximum generations.
The price of a cat depends on the cat itself but the generation does not really matter in its formation. The hypothesis about the guy who breeds the cats of the largest generation to earn seems to have failed. Unless he does not sell out them in some other way. It may be worth looking at buying crossings using the SiringAuction contract (the cats are not put up for sale, but one-time crossing with a cat :)).
Average price on crossing depending upon cat’s generation
As shown in the graphic above, the larger the generation is, the lower the price of breeding becomes. Kitties’ youth matters.
Normal cats have a set of cattributes that form their appearance. The direct logic suggests that the rarer any trait occurs, the more expensive the item is. Let us test this assumption.
Here are the rarest cats in the game. For defining the rareness of a kitty, the rareness of each characteristic matters and calculated by secret formula:
The cats above have not been purchased even once 🙂
If we, on the contrary, study the rareness of certain characteristics of kitties, that were marketed at least once, there won’t be any surprises:
The cost of rare kitties tends to the average price of all items.
Cattributes and their rareness are not so obvious for a player. Moreover, the traits that are considered to be the most pricy at the moment should show decrease in rareness, as the kitties with such cattributes will be bred more.
Now let’s take a look at the most expensive cats and at the same time quite rare:
The average rareness index of normal cats is about 3. The cats presented above possess the index 170+ (that is not the highest index, but is closer to maximum 500-800 than the 90% kitties). One may note that high price is formed due to additional rare elements (purrstige), or the combination of such characteristics, that accord kitties the more beautiful and catchy appearance. It can hardly be concluded about certain steady connection between the rareness and the price.
The following charts attest this. By each characteristic we observe approximately one and the same picture, namely sudden upsurge in certain rare trait. It should be mentioned that these spikes affect the price in no way.
For those who are inquired about Cryptokitties game and would like to learn more about the most popular cattributes, the following information is shown below:
The most popular body types:
The rarest body types:
The most popular body types:
The rarest body types:
The most popular colorprimary types:
The rarest colorprimary types:
The most popular colorsecondary types:
The rarest colorsecondary types:
The most popular colortertiary types:
The rarest colortertiary types:
The most popular eyes types:
The rarest eyes types:
The most popular coloreyes types:
The rarest coloreyes types:
The most popular mouth types:
The rarest mouth types:
The most popular pattern types:
The rarest pattern types:
The most popular environment types:
The rarest environment types:
Purrstige – is a unique element of a kitty. There are only 2 species of this element and not many items that possess it.
The most popular wild types:
The rarest wild types:
This cat obtains summarily rarest cattributes:
It is obvious that rareness of a kitty is inverse proportion to number of matings it has had. Let us consider the incidents with body type:
Still, if look at paid cross-breedings, we won’t see any connection between the rareness and the price again. The charts will look like the following:
That is , there are some upsurges with no logical reason to it.
Is there any correlation between the balance of a wallet and the number of kitty tokens that the user possess?
A great number of players who were involved into the game has either changed their accounts or just got rid of their kitties and stopped playing. The quantity of such users is estimated to be over 14000. Speaking about the existing accounts, with the increasing in the number of kitties the amount of players naturally and pretty quickly decreases:
1 CK 8493 users
2 CK 9749 users
3 CK 10004 users
4 CK 6346 users
5 CK 4303 users
6 CK 3265 users
7 CK 2476 users
8 CK 1887 users
9 CK 1437 users
10 CK 1204 users
11 CK 935 users
Account balance dependence on the number of items possessed:
If account balance shows no regularities, then the connection between the number of tokens on Eth turnover is a little more vivid:
Here you may observe that the more kitty tokens the player possesses, the greater Ether turnover he has. Also let us compare the turnover of Kitty tokens and Ether:
The address (0xFC624F8F58DB41BDB95AEDEE1DE3C1CF047105F1) that had the biggest Ether turnover to the moment of research had about 3.5k kitty tokens on balance and its total kitty turnover was about 40k transactions.
It may be suggested that accounts with more than 1000 Eth of turnover will play the game more intensively, than the addresses having 500 Ether turnover.
As we showed above, a separately taken attribute of a Normal cat does not statistically affect the price or the number of purchases of this cat. The difference over a large sample is obvious only between different types of kitty tokens. The histogramme of the dependence of the number of deals on the type of a kitty illustrates this:
As you can see, the majority of deals occur with fancy kitties and ones having wild traits. But, if to separate these parameters to total amount of kitties for each type, then it is noticeable that the most purchased and sold cryptopets are ones from Special Edition.
Thus, if player wants to make money off virtual pets, he should pay more attention to Special and Fancy items.
Widely recognizable chart illustrating the number of calls to KittyCore contract with the splash in the end of 2017 is almost identical to the stats of all deals, with the exception of New Year 2018.
If we take the stats by days of a week, then Wednesday is a bit ahead, though these data may be deceptive.
When we look at this option with a limited number of users that had only 10 Kitty turnover or less ( i. e. to remove the accounts of avid players, markets, shops, etc), then the 1st position goes to Monday and Tuesday:
Almost the same picture we get from Ether turnover for the SaleClockAuction contract. For all addresses who play Cryptokitties game:
For addresses whose turnover is less than 10 Cryptokitties:
In addition, to answer the question correctly we should remember about the difference in time zones. However, we will come back to days of week further on.
As we have already found out the rareness of a cat and its cattributes do not affect the price so much as some unique “off-game” characteristics (token ID, name, etc) do. To confirm this, let us look at Kitty tokens for which difference between purchase price and selling price exceeds the amount of 100 Eth (towards an increase).
As you can observe, the most profit-making Kitties’ IDs are under 100. These are the so- called “founders’ cats”, that firstly were bought for 9 Ether and then sold for 100+ Ether.
Let us have a look at a list with “younger” cryptopets ( ID > 100):
id 114684 ~ name FRAIDSYCAT
id 2238 ~ name 战斗猫
id 35068 ~ name Dracula | Earliest Gen Vampire
And the Kitty purchased for 600 Eth you are already familiar with!
Here are the most requested smart-contracts by Cryptokitties users for the latest 3 months ( the titles for smart contracts were extracted from bloxy.info). Requests from user accounts to smart contracts:
Requests from smart contracts addresses to user accounts:
Recognizable addresses of exchanges (EtherDelta, IDEX) are diluted with some other interesting accounts. One may observe a nice scope of cryptogames: 0xUniverse, Axie Infinity, KryptoWar (ranked third, titleless), MLB Champions and a number of some other apps. Also the presence of casino smart contracts( Dice2Win, FckDice) and other addresses belonging to gambling services in the list is interesting . Also several auctions can be found. All these DApps should be cared in mind when to run advertising campaign or use any other means to attract the target audience (airdrop, for example). Auctionity is the first, but there is some surprise in these stats. More details are further on:)
The average interval between two calls from one and the same player is less than 2 minutes. Still, this parameter is interesting for short periods of time.
Let us take the Retention parameter for first months after Cryptokitties were launched and compare it to the number of calls for that period:
23 November, 2017 – 1 May, 2018
23 November, 2017 – 1 December, 2017 (before the upsurge)
December, 2017 (upsurge)
You see during the most active periods the chart is shifted to the right. This may indicate that since the users have made several transactions within a couple of minutes, they do not leave the application, but remain in the game for a longer period: 2-3 hours or more.
Thus, by using this metric, it is possible to track the interest to a new introduced game event or advertising campaign. First, we focus on the number of transactions and the influx of unique addresses – this is the first step of analysis. Then we can assess the quality of the audience that came: whether they only started the DApp,made some test transactions and quitted or joined the existing players of a game.
Still, it should be noted that the explanation provided above does not cover all the cases. If for Cryptokitties we have a dragging out when activity is high, therefore the understanding of crypto casinos logic should be quite different.
Have a look at the chart describing the user retention of Dice2Win (0xD1CEeeeee83F8bCF3BEDad437202b6154E9F5405) casino for the latest 3 months:
Another example of a Cryptokitties-like game is Dragonereum:
November,2018 – December, 2018 (the upsurge in activity when everyone bought and unpacked the eggs):
December, 2018 – January, 2019 (after the peak, when battles and trading began):
If take the general stats, then most of the users seem to come on Monday. However, considering time zones and if to request the data for separate months, then the most active days of a week are Saturday and Sunday, followed by Friday and Monday. Conversely, the most passive days are Tuesday, Wednesday and Thursday.
These days are good to initiate certain gaming activities or start advertising campaigns.
Here are the list of smart contracts ( total number is 225) that interacted with KittyCore Contract. The descriptions are taken from bloxy.info.
Apart from 7 dozen autowallets the full list contains several quite interesting smart-contracts that attempted to create autonomous contracts over the Kittycore contract. For example, the battles between cats 🙂 Unfortunately, none of them took off. EtherDelta is just a throw off (two transactions over a year old)
The majority of fail transactions happened when a player wanted to call a certain method when he had no permission to do this. For example, to sell someone else’s kitty. If we look at chart of fail transactions, we will see that most of them are concentrated in December, 2017, when the project was just launched. It can be assumed that at the first stage there were some bugs when forming a transaction on the UI, that were later fixed.
On 25 March, 2018 the address 0x68B42E44079D1D0A4A037E8C6ECD62C48967E69F made 3330 failed transactions trying to call the breedWithAuto method.
Here is a list of top-5 addresses that have had most fail requests:
0x6B9AA4087E8643E672CAAD55CEB8A78D547189EA 13667 fail txs
0x1F35924AE9B34BBD1DFA7E527964C430D5AFAB92 9174 fail txs
0x49C4460CC2366BD93D24974FF6813EF7FF4FB649 5644 fail txs
0xE1AB3DF91C8D871D59AE48D1DF18664173CB6297 4865 fail txs
0x4B15E04EBA58A46FB44829BBFE0D9CD82C149745 4830 fail txs
The list below enumerates top 10 players that have gained the most revenue playing the game. The parameter called “Kitties_amount” is the number of cryptopets, that the user has had for the all period of gaming. Apart from the developers’ account the list includes the address of already mentioned player (ranked second) who has bought a kitty from himself for 600 Eth
These are accounts that are running at the most deficit
The main unit of a decentralized system is token. The more it is in demand the more success the DApp has. The situation of such interdependence of supply and demand is a pyramid. The more users and assets are involved in the game, the more valuable this game becomes. Thus, the main goal for developers is to create an environment in which new technologies and new players are constantly involved.
The first thing to do is to study when generally new accounts start using the DApp and find out the reason of such behavior. If to speak about Cryptokitties, the vivid interest is observed during the New Year time: New Year 2018 was the launch of the project and then that all-known “kitty boom”. A year after, New Year 2019 the game celebrated the first anniversary and the team used it as a good moment for giving-out of cryptocutties. Advertising campaigns and special events related to special occasions to which users react through the application, allow to transform a passive player into a more involved one. Having the figures, one can build a marketing strategy and understand what time is the best to start it. Statistics in its turn allows to monitor user reaction and behavior during a particular campaign: which methods are called most often, the intense of user retention to the application, etc.
The second to be done is to analyze the periods ( time intervals, days of week) when the audience of the DApp is the most active and make prediction on further activities. These data allow to make some conclusions on user profile and behavior. For example, region, evening activity or during working hours, weekends or weekdays, on which days the turnover is higher, and which days is traffic. These metrics are of interest not only for the team of the project, but also for partners and competitors. If you understand the users of your business, then you better realize what kind of product they need the most.
The third is to single out the most active accounts of your competitors’ products ( the biggest profit, the most transactions, etc). They are most likely to have interest to your DApp, that’s why they are highly desirable to be airdropped from you. One more moment not-to-forget is to study the behavior of such users and analyze their nearest environment using the graph.
All in all, the number of transactions depends on two reasons – the size of the audience and the intensity of the gaming process. Regarding the first factor we have already talked about above – this is primarily analytics, marketing and business strategy – then the second one is mostly connected with the developer of a smart contract.
Also you need to get to know why the players like this or that game. Such data can be obtained by analyzing the statistics on the methods called. If a bet was made, say, on the crossing of cats, but the audience is interested in the auction part of the DApp, then it may be worthwhile to adjust the logic of the game. Chaingraph.io easily provides you with these data : one may calculate not only the overall retention to the smart contract, but also to the each of the contract methods.
The lack of demand for some methods can be associated with problems on user interface, errors when making a transaction, and even the amount of gas needed. Therefore, it is necessary to have constantly updated relevant statistics on all technical aspects of the decentralized application: trace not only external calls but internal ones as well, event logs, gas statistics, etc.
It is almost impossible to foresee all the vulnerabilities and loopholes of your application, but you can minimize the likelihood of their occurrence and rapidly track the possible failure if happens . The first may be implemented by the Security Check service, which detects possible bugs in the smart contract code. For the second item you need a tool to monitor user activity. All these instruments are at the disposal as soon as you start using Chaingraph.io.
Chaingraph.io: Security Check Tool
Chaingraph.io: Live Section
Some examples of atypical user behavior on the example of CryptoKitties:
Buying and selling kitty from one and the same account may testify on a bubble of statistics and kitty price. Without analyzing the situation, a third-party user can observe a picture of active trades of a certain token, but this can only be a game to create the illusion of its “demand”. Such action can lead the local market of the application to undesirable consequences and the developer, at least, should be aware of the current economic situation in order to react adequately to it.
The addresses that were involved into a transaction for 600 Eth
Next is one more illustration of cheating statistics, this time the number of sales. More than a dozen kittytokens are far ahead from the nearest top (the gap is 20-30 times). If to go deeper it becomes evident that all 16 Kitties that have approximately the same number of transfers (about 300) were transmitted through the Auctionity smart contract for the same price. The representative of this crypto-auction commented that those actions were made to demonstrate the capabilities of the service, to boost initial stats and are not real sales. Perhaps this is a good example that statistics alone sometimes is not enough, it should be digested first.
The gap in the statistics on the number of cat sales
The answer to this question is mainly got from two previous ones. With the stable game process, the increase of the audience and the interest to the DApp, apart from gaining more transactions you also get the increase in value for your token. Also, bear in mind that cryptogames are about money. Therefore, the trading volume of your token depends on the economic success of the entire project, that is composed of many aspects. Competent analysis gives an understanding to which parts of the project you should pay attention foremost.
Live Section that allows for tracking activity in the real mode
These are several cases:
- Buy a rare (exclusive, special) cat, preferably at some in-game event.
- Monitor the graphs of user activity (the number of transactions, the number of accounts calling the smart contract).
- When bursts of activity happen, put up a cat for sale at an inflated price.
- Gain profit.
- Monitor the occurrence of the event on the crossing of Fancy cats.
- Buy the maximum number of cats with the required genes.
- Quickly crossbreed new fancy cats.
- All kitties that have the right genes that remained after receiving fancy cats put up for sale a little more expensive than the buying price.
- A couple of fancy cats sell immediately at an average price in the market.
- The rest leave on the Long Play.
- Gain profit
- Create a number of Ethereum addresses – these will be pseudo players.
- Trade kitties between them, breedcross, etc, i.e. create an ecosystem with a local economy.
- Choose one cat or more for promotion – preferably with some special feature (interesting name, description, body part, etc).
- Create around it an artificial picture of demand: resell it for big money, boost the price, offer for sale on Opensea.io.
- The bigger and more unexpected the numbers are – the more interesting. It is advisable the Kitty to be covered in the crypto-media.
- Sell a cat at a high price.
- Gain profit.
- Get statistics on the most active players of other decentralized games that have not played cats yet.
- Send them an egg with a cat.
- Analyze the first environment of CryptoKitties users.
- Detect which other decentralized applications the audience uses. For example, online casino Dice2Win, cryptokitties-similar-like games Blockchain Cuties or collection game MLB Champions.
- Analyze the smart contracts of competitors and figure out regions with potential audience. In this case, it is even not necessary to determine the region, it is enough to compare the graphs (data shown in UTC +3):
Average transaction time
for January,2019 – May,2019 for CryptoKitties:
Dapp: MLB Champions
From the sample above 0xUniverse matches the best : the region coincides and a large number of CryptoKitties users play this game.
4. Start cross-promotion (preferably from 9 pm till 2 am (UTC +3)).
5. The idea is to exchange tokens if some special conditions are met. For example, if you have certain kittytokens, then you can get a bonus in another game and vice versa.
- By analyzing the activity of the audience, you can hypothesize the time zone of most users. Add information from the application database and get the most active regions. We, in contrast to the developers, are limited in obtaining information about the IP-addresses of players, so we can only assume: the main audience of the game is located in the American region. It is more likely to talk about the United States (especially the eastern coast) and Canada, however, as an option, it may be a country of South America.
- Now you can attach gaming events to events of importance for this region: Christmas (cat-Santa), Halloween (cat-zombie), Thanksgiving (cat-turkey 0_o), Black Friday (cat-discount 0_0), etc.
- Do not forget to monitor the figures and the reaction for each activity and thus, single out the most profitable regions.