,
15 tweets,
5 min read
Read on Twitter
1/
🔍🧐 Cryptocurrency exchange trading volume investigation 🧐🔎
Our team set out to determine whether volumes reported on cryptocurrency exchanges were genuine.
Thread Follows:
The data: docs.google.com/spreadsheets/d…
🔍🧐 Cryptocurrency exchange trading volume investigation 🧐🔎
Our team set out to determine whether volumes reported on cryptocurrency exchanges were genuine.
Thread Follows:
The data: docs.google.com/spreadsheets/d…
2/
We started by pulling a list of the top 100 exchanges by reported trading volume over the last 30 days.
We started by pulling a list of the top 100 exchanges by reported trading volume over the last 30 days.
3/
We then used Similar Web website viewership metrics to calculate the estimated 30 day traffic to each exchange's website. After doing this, we divided reported volume for each exchange by the # of monthly website visits to determine the reported volume per visit.
We then used Similar Web website viewership metrics to calculate the estimated 30 day traffic to each exchange's website. After doing this, we divided reported volume for each exchange by the # of monthly website visits to determine the reported volume per visit.
4/
While website viewership metrics don't account for API or mobile app trading, they do provide a good basis of comparison across exchanges to identify suspicious reported volumes.
For example, while #Binance reported $750 traded per visit, DOBI reported a whopping $356,625
While website viewership metrics don't account for API or mobile app trading, they do provide a good basis of comparison across exchanges to identify suspicious reported volumes.
For example, while #Binance reported $750 traded per visit, DOBI reported a whopping $356,625
6/
We then wanted to create a basis of comparison to calculate what we expected an authentic representation of exchanges' volumes would look like. To do this we took a weighted average of trading volumes across Binance, Coinbase Pro, Poloniex, Gemini, and Kraken.
We then wanted to create a basis of comparison to calculate what we expected an authentic representation of exchanges' volumes would look like. To do this we took a weighted average of trading volumes across Binance, Coinbase Pro, Poloniex, Gemini, and Kraken.
7/
We selected these exchanges because of large usage among institutions, reputation within the market, and because their web viewership appeared consistent with their reported trading volumes.
We found that these exchanges averaged $591 in trading volume per web visit.
We selected these exchanges because of large usage among institutions, reputation within the market, and because their web viewership appeared consistent with their reported trading volumes.
We found that these exchanges averaged $591 in trading volume per web visit.
8/
Next, we multiplied each exchanges' web views by this $591 to calculate what we would expect their volumes to look like.
In the chart below the further to the left and top an exchange is, the higher that exchanges reported vs expected trading volumes.
Next, we multiplied each exchanges' web views by this $591 to calculate what we would expect their volumes to look like.
In the chart below the further to the left and top an exchange is, the higher that exchanges reported vs expected trading volumes.
9/
Some notable outliers were BitMax, Lbank, BW, and ZBG which had extremely high reported volumes, but expected volumes which in some cases were less than 1% of what they reported.
Some notable outliers were BitMax, Lbank, BW, and ZBG which had extremely high reported volumes, but expected volumes which in some cases were less than 1% of what they reported.
10/
When we divided the top 100 exchanges' expected by their reported volumes, we found that 59% of exchanges' reported volumes were over 10 times higher than what we would have expected had they similar volume per visit to Coinbase, Binance, Kraken and others.
When we divided the top 100 exchanges' expected by their reported volumes, we found that 59% of exchanges' reported volumes were over 10 times higher than what we would have expected had they similar volume per visit to Coinbase, Binance, Kraken and others.
11/
75% or exchanges reported trading volumes that were more than double what we would have expected. In the chart below those exchanges are colored red.
The exchanges that had expected volume/reported volume ratios greater than 50% were colored green.
75% or exchanges reported trading volumes that were more than double what we would have expected. In the chart below those exchanges are colored red.
The exchanges that had expected volume/reported volume ratios greater than 50% were colored green.
12/
In total we estimated that 87% of exchanges reported trading volume was potentially suspicious and that 75% of exchanges had some form of suspicious activity occurring on them.
In total we estimated that 87% of exchanges reported trading volume was potentially suspicious and that 75% of exchanges had some form of suspicious activity occurring on them.
13/
If each exchange averaged the volume per visit of CoinbasePro, Gemini, Poloniex, Binance, and Kraken, we would expect the real trading volume among the largest 100 exchanges to equal $2.1B per day. Currently that number is being reported as $15.9B.
If each exchange averaged the volume per visit of CoinbasePro, Gemini, Poloniex, Binance, and Kraken, we would expect the real trading volume among the largest 100 exchanges to equal $2.1B per day. Currently that number is being reported as $15.9B.
14/
There were limitations to this report including some of the aforementioned, but the point of the exercise was to show those exchanges that appear most suspicious and to start a greater conversation around wash trading, transaction mining, and liquidity.
There were limitations to this report including some of the aforementioned, but the point of the exercise was to show those exchanges that appear most suspicious and to start a greater conversation around wash trading, transaction mining, and liquidity.
15/
We made all of the data available for free in a Google Sheet below.
We are actively adjusting pricing, volume and exchanges listed on The TIE to more accurately reflect what we believe to be non-suspicious activity. We appreciate community feedback
docs.google.com/spreadsheets/d…
We made all of the data available for free in a Google Sheet below.
We are actively adjusting pricing, volume and exchanges listed on The TIE to more accurately reflect what we believe to be non-suspicious activity. We appreciate community feedback
docs.google.com/spreadsheets/d…
