A graphical analysis of the traders that have been active on Ethereum based DEXes
The data that decentralized exchanges leave on the Ethereum chain is written history . The artifacts, log entries, transactions and contract messages, are there to stay.
Of course you can explore DEX trades on our monitoring platform (http://dex.watch/). But as always, there are many ways to look at a data set and we’re constantly exploring the possibilities. One interesting view on trading activities is the the co-currency/percentage matrix of traders. It counts all unique trading addresses that are active on the DEXes. We think it can provide useful insights on user distribution and movement between different platforms.
Dex.watch is currently monitoring 17 DEXes. How many of the traders are solely active on one platform? How many use multiple DEXes? Below you can see a chart showing the number of addresses that were trading on 1 or more platforms. We collected the data on October 17 — on that day, our database contained 6,634,760 trades and 214,391 unique traders across 17 DEXes.
We further excluded 76,034 small traders who only had 2 or less than 2 trades in their history. The distribution of the rest of the traders in the dataset is below:
Fig 1 — Distribution of Traders Using Only 1 DEX, 2 DEXes or Multiple Platforms (excluding small traders with ≤ 2 trades)
The largest group of traders is trading on only 1 exchange. There also is a large portion of traders with orders across multiple exchanges. The reason for this could vary, ranging from arbitrage opportunities on price differences, trading preference or just experimenting the leverage of different mechanisms.
By arranging the DEXes in a matrix and counting the traders each pair of DEXes share, we get a type of heat-map of shared customers.
Fig 2 — Number of Unique Traders Shared by DEXes
- The diagonal represents the number of exclusive users (it matches a DEX with itself). There are different possible interpretations for the values — they can represent loyalty of traders to the platform or just represent a legitimate address generation mechanism or even a wash-trading scheme
- The non-diagonal fields show the user similarity between DEXes by the number of traders shared with each other
- The 0x protocol has a special role in our table. All the DEXes below its entry are DEXes who are based on the 0x protocol. Thus, all the trades will be recorded both in their own platforms as well as for 0x protocol, they are duplicated in the matrix. For example all users of DDEX will also be counted in the 0x row. This explains why all the DEXes beneath 0x have no diagonal element
- By ranking the number of shared traders in each column, the potential direction of adoption of other DEXes can be estimated
This plot only shows the lower-triangle of DEX pairs. The upper-triangle is redundant because it is symmetric to the lower one. Fields with a value of 0, are left blank for better readability.
The Role of Smart Contracts
One unique feature of decentralized exchanges is the ability to create smart contracts that interact with them. This can be, for example, for arbitrage, batching orders or market making. Among the ~200k addresses, 1,300 are contract accounts. Most these use the DSProxy contract , which may work for harvesting arbitrage either within one platform or across multiple platforms.
OasisDex has the largest number of different contract accounts in its diagonal element (1091 ), these exclusively interact with OasisDex. The second highest is EtherDelta , having 114 contracts solely trading on its platform.
Fig 3 — Number of Smart Contracts Interacting with DEXes
A closer look on unique traders
IDEX has the highest number of unique traders on their platform. One should interpret this finding with caution, however, as it may misrepresent the actual user size due to many potential factors behind. But the number is taken directly from the blockchain and shows that there have been 119,910 unique addresses that exclusively appear on IDEX and never use any other DEX for trading tokens. The second position is Etherdelta with 39,591. Because IDEXis such an outlier it strikes out as an extreme dark block compared to all the others, which reminds us that this count should be put in proportion with the volume that was traded.
In most cases, there is an incentive for exchange operators to try and increase the observed volume on an exchange. Higher volume signifies higher liquidity and trust and will attract more users. This incentive can be different for fully decentralized exchanges but for most existing ones it is an important factor to consider. Regarding fake or wash trades, the chances are high that they are fully included in the diagonal counts. The platform operators/advocates will only generate addresses and submit orders on their own platforms, rather than go to other platforms to also do fake trades.
Below, you can see the matrix from above, now showing percentages of shared users instead of absolute values. For each DEX the column sum is 100%. The percentage matrix is not symmetric because the elements in opposite positions now are attached with different meannings — e.g. Element E_12 and E_21 both means the users who use both Airswap and ETHREC, but are divided by different user base size of Airswap and ETHREC respectively, so they have different meanings to different DEXes, according to the column they stay.
By looking at the columns, we can see the percentage of trader accounts shared with the other platforms. The darkest block by column shows the max value of each column. The percentage table reveals the make up of trader account type on all DEXes.
Fig 4 — Percentage of Unique Traders trading on 2 DEXes
On this percentage matrix, the diagonal values should be understood in a different way: differently from absolute value, the percentage can be seen as a metrics to estimate loyalty of users for each platforms, as the higher the percentage of exclusive users it has, the more likely that users are not going to move around and adapt other DEXes, but stick to the one only.
Reminder: the seven platforms showing highest shared user base with 0x are all using 0x as their protocol, thus the records are duplicated. For them, those darkest elements are the exclusive user number.
By removing the diagonal from the matrix the chart allows us to focus on the interaction between decentralized exchanges by proxy of their users.
Fig 5 — Co-currency Matrix Masking Diagonals
One could interpret the counts as the closeness between each pair of platforms. The darker the elements are, the closer the two DEXes are, as they share a larger group of users. There can be various reasons, it can be because the platforms share available token pairs, or they operate similarly in order book management or UX, that a group of traders favors. As these are also major aspects that users would consider when choosing platforms, we believe the value also implies the likelihood of users adoption/moving direction — for DEXes sharing higher number of users, they are more likely to be the substitutes for each other.
We will continue to analyze activities on the evolving ecosystem of decentralized exchanges. Open topics are
- Contracts and smart contract systems that use DEXes
- Wash trading and market makers
- New types of DEXes like DutchX
Feedback is always welcome!
Watch this space for more Ethereum analytics.
Disclaimer: As the available data ( # of listed DEXes, the length of trades history we cached, etc.) is limited to what can be seen on chain and the mechanisms of different DEXes vary diversely, we readily admit that our research is incomplete. We made assumptions and clearly labeled them, as well as we were able to. Alethio has a strong commitment to staying neutral by providing facts and best judgments based on objective and/or verified information. This article should never be used as guide for any malicious practice or trading suggestion.