Most Bitcoin Trading Faked by Unregulated Exchanges, Study Finds | Hacker News Reader

Most Bitcoin Trading Faked by Unregulated Exchanges, Study Finds

wsj.com

The mention of 81 exchanges left me thinking: If none of the trading was fake, but then I created a single exchange literally named "ScamCo Faketrading Exchange" which claimed 1 petatrades per second were happening on its platform, then would this analysis say that >99% of all trades were fake?

I think it would.

I don't think any of the exchanges that I could have named off the top of my head were on their naughty list (paywall is keeping me from loading it again to check), ... so it seems to me that this report might be a bit overstated.

You're right, few of the mentioned exchanges are actually well-known. They seem to be using coinmarketcap.com as their source for the 'top exchanges' which is basically going off of reported volume with no validation.

Of course there are going to be scammer/spam exchanges reporting ridiculous numbers. How about looking at Coinbase, Kraken, Bitstamp, Bitfinex, Gemini. At least those have been around awhile.

The actual volumes on different exchanges are not anything close to that disparate, so while your hypothetical does raise a real way that such a claim could be misleading, it has nothing to do with the actual situation at hand.

The presentation (linked in another comment here) lists a number of exchanges which they believe are presenting fake data:

CoinBene, RightBTC, CHAOEX, IDAX, LBANK, BitForex, Exrates

I don't think I've heard of any of these, and there are zero hits in my email.

They go on to identify pretty much every exchange I've ever seen discussed on HN (for example) as having legitimate volume-- coinbase, bitstamp, bitfinex, gemini, ...

I think the headline should instead be "coinmarketcap volume claims are nonsense" -- which I think most people in the bitcoin space would say "well, duh". At least my view purpose of coinmarketcap and similar sites seems largely to direct people to the latest token pumps, as evidence by the absurd "market caps" they display for illiquid pre-mined tokens...

What set off alarms for me is that there is no way that you get to 81 exchanges without including some really sketchy or obscure stuff.

But a site that wants to list sketchy tokens needs to include the sketchy exchanges because for the most part the non-sketchy exchanges won't list them.

Bingo. I started using Messari/onchainfx and it’s a much better alternative.

Not quite that extreme, but the difference is quite disparate. Coinbase Pro (legit) has a 24h volume of 53,234,129 whereas Coineal is reporting 740,643,873. A quick google / google trends search shows that Coineal is basically unknown compared to Coinbase.

Article seems to be going by:

https://coinmarketcap.com/rankings/exchanges/

Here's the slide deck presented by Bitwise, who did the study: https://www.sec.gov/comments/sr-nysearca-2019-01/srnysearca2...

Another summary article (non-paywalled) from CNBC: https://www.cnbc.com/2019/03/22/majority-of-bitcoin-trading-...

EDIT: This is a very slick presentation. The weakest point I saw was a bit of handwaving around the key signing ceremony, specifically about how sharded keys are distributed. It's odd that they go into such detail about how to provision a clean laptop but then just sort of say "well, take your private key and mask it into seven pieces".

The slide deck is worth reading. There are only 10 real Bitcoin exchanges doing real transactions. The others are all clearly faking it, just from observing the overall trading patterns.

Something the slide deck doesn't address is how much all that fake activity affects the real price. The slide deck demonstrates that the real exchanges track each other closely. But what's the fake activity doing to the price? It's much bigger than the real activity; it must have some effect.

This new ETF looks better than the rejected Winkelvoss ETF. This output proposes third-party custody and insurance from major insurance companies. The Winkelvoss ETF tried to unload all the risk of theft on their customers.

Page 61 has the 10 exchanges that appear to have real volume of the 81 exchanges analyzed: Binance, Bitfinex, Kraken, Bitstamp, Coinbase, bitFlyer, Gemini, itBit, Bittrex, and Poloniex. They estimate that the total ADV for the market is $273M, less than the $6B reported by https://coinmarketcap.com.

>Page 61 has the 10 exchanges that appear to have real volume of the 81 exchanges analyzed: Binance, Bitfinex, Kraken, Bitstamp, Coinbase, bitFlyer, Gemini, itBit, Bittrex, and Poloniex.

I didn't expect Bitfinex to be on that list, considering all the USDT shenanigans that they were caught up in.

I was a little surprised by that as well, but my own analyses of Bitfinex data (together with the USDT/USD market on Kraken) brought me to a similar conclusion that they are likely operating above board. That said, there have been previous studies challenging that conclusion so I wouldn't be surprised if they found more signals on a deeper inspection.

The slide deck does go into the USDT/USD differences as well, and even mentions the large Bitfinex hack.

Probably no surprise to most people at least moderately familiar with the Bitcoin trading world. I would've named 8 of those off the top of my head as exchanges with legitimate volume, and I'm not a day trader or all that involved, just a bit of a holder and followed along the 2017 craze.

Fascinating slide deck. They work out a scoring system to identify exchanges with fake volume.

This has a number of consequences. For example, the exchange-traded volume starts to look more realistic given Bitcoin's aggregate value. Similarly, futures volume starts to look more comparable to exchange traded volume.

The main benefit, however, is to surmount the SEC's objection to previous ETF applications: bitcoin market manipulation. If they succeed, Bitwise stands to become the first Bitcoin ETF in the US.

I think the one-cent spread on Coinbase Pro mentioned in that article is partly just an artifact of their fee structure. Maker orders which are placed and sit on the order book have no fee, whereas orders which match against the order book are charged a fee. So unless people are desperate to trade, they often save on fees by place a post-only order at 1 cent away from the other side of the orderbook. You can often see this as a pair of walls 1 cent apart on either side of the orderbook.

I spent last summer algo trading crypto. This study doesn't surprise me at all. On one market, I consistently saw limit orders that would appear inside the spread and then get instantly filled.

If you do the math on how much you need to spend on exchange fees to get a coin to get coin to the top 100 in volume, it's only a few grand. That's assuming you don't own the exchange.

Company: "All our competitors are fraudulent. We're on the level, we promise"

Seems like a good thing. Look at the volume for cryptocurrencies. It seems to be insane / ridiculous, as if the sum total of all coins for almost any given cryptocurrency are exchanging hands weekly (I know this is not exactly what volume means, but it's close enough).

Looking at bitcoin volume on coinmarketcap right now, it's about $9B 24h volume on a $71B market cap. In just 8 days the entire market cap of bitcoin will have 'exchanged hands' (again, I know this is not exactly what volume means). Meanwhile, most people are not using bitcoin as a medium of exchange. These numbers just seem absurd to me. There's no way that people are using bitcoin as a medium or exchange OR a speculative trade that frequently, unless these traders are fickle idiots.

That said, I know little about how currencies work, so maybe I'm making some invalid assumption. But I don't see how this data being faked is at all meaningful unless you're using this data to try to estimate the "value" of bitcoin, etc. If so, yeah this is bad news for you. But how did you not know that bitcoin, etc. aren't really being used to exchange value?

You say:

"Looking at bitcoin volume on coinmarketcap right now, it's about $9B 24h volume on a $71B market cap. In just 8 days the entire market cap of bitcoin will have 'exchanged hands' (again, I know this is not exactly what volume means)."

And, from the deck in the article:

"Despite its widespread use, the CoinMarketCap.com data is wrong. It includes a large amount of fake and/or non-economic trading volume, thereby giving a fundamentally mistaken impression of the true size and nature of the bitcoin market."

(Emphasis mine)

What exactly is your point? That I am repeating the point in a paywalled article? I have always thought the market data on cmc makes no sense, and am pointing out a simple experiment you personally can do to check.

I just am saying that basing any assumptions on the data from CoinMarketCap is to make those assumptions meaningless, because the data at CoinMarketCap is false. Apologies if I'm misreading your comment.

Yes the massive volume of fake transactions is actually good for Bitcoin.

I appreciate the snark, but my point is that the massive volume data on coinmarketcap is so unbelievable that I prefer knowing that it's actually lower.

Do you have anything else to contribute? Would it make you happy if actually bitcoin exchange volume were $17 TRILLION dollars, or if it were reported as such?

Really the snark is enough.

High frequency trading perhaps?

Possible, but I don't think it's likely. The volume on something like BTC is higher than AAPL. It doesn't make sense and never made sense. I don't think there are enough people who wrote and confidently run their own crypto-HFT bots to compete with volume in the stock market.

It does make sense if you look at how change addresses work with Bitcoin. Best-practice is to generate a new address for the 'change' from each transaction.

Suppose an exchange has a hot wallet with 10btc. Each time a user requests a withdrawal, say 0.01 BTC, the exchange generates a txn that sends 0.01 to the user, and the remaining 9.99 btc to a newly generated hot wallet address. On the blockchain, we can't tell if that was a transaction for 9.99 btc or 0.01 btc and most of these 'analysis' services are considering both to be volume.

This amplifies apparent volume even if no one is doing anything questionable.

To use the AAPL example, this type of blockchain volume analysis is like adding each trader's private brokerage account value to the total AAPL trade volume each time they buy/sell a share.

Reply on news.ycombinator.com