- It is well known that crypto exchanges have not been model citizens (Mt. Gox, Zaif and many others)
- But it’s hard to catch them engaging in bad behavior; especially in real time
- Golem trading on ZB.com is an exception where it was observed to spike volume by tens of millions each day through dozens of identical market buys and sells
- This pattern of Golem trading is apparent as far back as mid-November 2018
It is not a secret that anomalous and quite likely nefarious behavior happens on crypto exchanges. Other analysis has estimated that OKex, one of the major exchanges, has falsified approximately 93% of the total volume. Sylvaine, the author of the analysis, established a group of exchanges he believes are “clean” and then ran several linear regressions across the broader set in order to arrive at his conclusions. His methods are spelled out in detail in the article and he includes the data at the end for those interested in further experimentation.
Over the past weekend, a different type of anomalous (nefarious?) trading happened on one exchange on the BTC/GNT trading pair. Golem (GNT) has a current market capitalization of $67 million and traded just over $1.3 million in the last 24 hours according to CoinMarketCap. On December 30, that volume spiked to over $40 million. The main contributor to the volume spike was an exchange based out of China called ZB.com. They were highlighted in Cryptocompare’s October report where the discrepancy between visitors to their website and total volume was outside the norm relative to their peers:
This spike in volume is similarly outside of the norm and raises red flags around the legitimacy of the trading volume. In the visual below the volume is shown to spike to nearly 50 million GNT tokens and then falls back down to just over 1,100 GNT. This is not normal market behavior.
Since Saturday, there has been another round of strange behavior on the GNT/BTC market on ZB.com. At 23:22, 23:23, and 23:24 EST on January 2, there were a significant number of buy and sell orders that came through the trade book all of the same size, 2915.9 GNT. It is extremely improbable that the market would naturally execute dozens of orders of exactly the same size within a 3-minute window:
After volume had fallen back to $1.3 million and these suspect market trades started to pick up, volume has inflated back to $12 million or more. This is lower than the previous $42 million mark but still a dramatic increase in a short period of time.
It is important to note the fee structure of ZB.com – the trading fee is listed as a range between .00% and .1%. This means that if the volume is not legitimate, there is a path by which ZB.com could facilitate executing these orders without charging a fee.
This second finding of suspect behavior prompted a deeper look into historical volumes of GNT:
This is also not what normal market behavior looks like. The price is relatively stable while volume oscillates between approximately $1 million and $100 million in daily volume, and after December 10, the daily volume grew larger than the entire market cap. This is a flagrant bad behavior in these markets and ZB.com has accounted for large percentages of the bizarre volume.
- Strange market behavior is visible at the individual trade level through dozens of identically sized orders over a few minutes
- Strange fluctuations of volume on ZB.com
- Even stranger historical volumes since mid-November in the entire GNT market
These crypto markets are home to some of the most flagrant manipulation and abnormal behavior out there. This type of volume manipulation has also been observed in other coin pairs on ZB.com, but it’s not the only exchange where this type of volume inflation and falsification has happened. Retail often pays the price for this, but the institutional folks who do not scrub their data properly to exclude coins like this could be duped by the volume fluctuations. As Ateet and Ali from CoVenture highlight in their recent post, building a quant fund in crypto is hard.
Harry Sudock is working to bring transparency and accountability at Eastwood Analytics. He comes from traditional financial services with a background in building institutional data products, but caught the bitcoin virus in early 2017.