News for hackers. Stuff that matters.
Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by lordnacho on Tuesday December 18, 2018 @01:59PM (#18709444)
The article is missing some critical details.

What does the fund do? The article suggests they do arbitrage. Those opportunities actually tend to increase when things are volatile. Also while you can lose money doing it you wouldn't expect a precipitous collapse in NAV. If you discover you're slower than everyone else you can shut down.

Volumes aren't necessarily correlated to price either, so that isn't entirely convincing.

I've also heard that plenty of arb guys are doing fine.

If he's speculating and not just running arbs, what is he doing? If he's just punting the cryptos that would seem more in line with what's happened, but it's not clear what he's up to from this.

Also, with arbitrage it's limited how much capital you need. A lot of HFTs use very little. If you're getting money like a hedge fund you have to be sure it can be put to use.

Ex HFT and fund manager.

Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by bhaak on Tuesday December 18, 2018 @02:45PM (#18709832)
This interview[1] sounds like they called bottom at 6k and then the floor fell out: "We thought it was a bear market. I went into it thinking in the long run crypto is going to be a real structural shift in the world and I can just hedge my portfolio. And to be fair, we did a really great job not losing money the first 60 percent down. What you forget is that a market like Bitcoin that’s down 84 percent has dropped 60 percent—and then another 60 percent. That’s where the pain happens. You start buying Ether again, because it’s only $400 after being at $1,300. But then it drops to $100, and you’ve lost 75 percent of your money. We haven’t done horribly in that context, but we’re still down. [...] I did think Bitcoin was going to hold at $6,200. It stayed there for four months. It felt like the selling was finished."

[1] https://www.bloomberg.com/news/articles/2018-12-11/mike-novo...

Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by wpietri on Tuesday December 18, 2018 @03:32PM (#18710270)
"Felt like". That's beautiful. The dude had a feel and lost the equivalent of 2300 years of labor (at the US median income).

There's this bit in Taleb's first book, "Fooled by Randomness". This was before he made a shit-ton of money. He pointed out that a lot of the nominally successful traders he worked around would declare a strategy, make a few bets, make a lot of money, and then act like geniuses for thinking up the strategy. However, they'd never really examine the obvious alternate hypothesis: they got lucky.

I used to work for a proprietary trading firm. We'd get new traders by taking very confident, driven people (e.g., former Olympic athletes), making them be clerks for a while, and then turning them loose in the pits. The CFO kept a very close eye on them early on because their natural confidence combined with early successes lead them to believe they were gods. It wasn't until they got really hammered by a bad bet or two that she'd begin to trust them.

It reminds me of a favorite line: "Every corpse on Mount Everest was once a highly motivated person with a can do-attitude."

Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by _hardwaregeek on Tuesday December 18, 2018 @06:38PM (#18711627)
Honestly the schadenfreude is real. What part of the crypto graph made him believe he could make assumptions about, well, anything?
Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by oblib on Wednesday December 19, 2018 @12:14AM (#18713208)
I hate to admit I agree, but I have to be honest too. That is spot on.
Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by mortenjorck on Tuesday December 18, 2018 @04:40PM (#18710836)
I was pretty sure $6-7k was the floor for BTC, and it’s interesting to see a high-profile fund manager made effectively the same, expensive mistake. The biggest surprise for me, though, was ETH, which given its better fundamentals, I had thought would be where the market went after Bitcoin popped – and yet, it was not only dragged down, but crushed even harder than BTC.
Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by nostrademons on Tuesday December 18, 2018 @05:34PM (#18711231)
It was interesting to watch the crypto subreddits and see so many people make the same mistake. I was surprised that $6K held for so long, because if $6K really was the bottom, it would've been a pretty shallow drop from a very steep bubble. That's only 2/3 down, and it was the point Bitcoin hit in Oct 2017. Bitcoin was < $1k at the start of 2017. Any reasonable rate of return would put its price at under $2k if you ignore the bubble and focus only on fundamental metrics. Plus bubble busts tend to overshoot fair market value, where the underlying asset actually trades at a bargain because so many people got burned buying at the top.

The ETH drop is fairly easy to explain - it was the base asset for a very interesting double-pyramid-scheme on the way up, which meant that that leverage works in reverse on the way down. The ICO boom worked because many of the people investing in ICOs had bought their ETH for pennies and already recouped their costs; throwing $40M into a shitcoin makes a lot more rational sense when that ETH actually cost about $4K in real money, and the factor of 10,000 difference is the price appreciation of the asset you're putting in. That led to eye-popping dollar valuations for many of these companies (even though they were actually raising in ETH), which brought more newbies into the Ethereum world, which pushed the price even higher, which meant later ICOs were raising even more in dollar terms, even though the actual dollar amounts traded were tiny.

On the way down, this cycle works in reverse. Those ICOs don't actually hold $40M in a company bank account; they hold 33K ETH which was worth $40M at the peak. Now that ETH is down to $94, they hold more like $3M, and they're getting scared about being able to pay salaries. So every ICO that still holds ETH is trying to liquidate it for dollars, but they're fighting over the much-reduced pot of money that is coming into crypto now, which drives the price down every time somebody wants to sell.

Personally I still think we're not at the bottom yet (for either ETH or BTC), but we may be getting close. We're at the bottom when companies start going bankrupt, Solidity engineers start getting laid off, and people start going to jail. Assuming it's not all smoke and mirrors and some useful infrastructure was actually built, it'll be the buying opportunity of the lifetime then, because everyone will think this whole crypto fad will be totally over and you'll be able to buy up that infrastructure at really cheap prices.

Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by solveit on Tuesday December 18, 2018 @10:57PM (#18712886)
Re: your last paragraph

But if everyone thinks that (and most people who think crypto isn't all shit do think that), the bottom will be a lot softer than what you describe (modulo people going to jail, I think this is mostly independent of the value of crypto).

It's the efficient market hypothesis all over again.

Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by elhudy on Tuesday December 18, 2018 @05:18PM (#18711121)
What "fundamentals" exist for Eth that allow it to be valued at a specific price point? Eth has no earnings, no assets, and no liabilities. Its intrinsic value is close to 0. It could be totally wiped off the planet tommorow and almost nobody besides crypto traders would care. Crypto trading should be based on technical analysis not "fundamentals".
Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by nostrademons on Tuesday December 18, 2018 @05:36PM (#18711248)
Adoption and projects using it. Like any currency, ETH becomes more valuable when you can spend it on more things. It's the base currency for a whole infrastructure of distributed trust & settlement; if useful economic activity occurs on top of that infrastructure, ETH gains in value proportionally.
Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by elhudy on Tuesday December 18, 2018 @05:42PM (#18711291)
Ok, but the whole idea behind fundamental analysis is that you generate a specific value of worth based on your analysis. So what formula are you using to calculate a price point for Eth based on "adoption and projects using it"? Are you tracking all projects using it, and assigning it a dollar value based on a per project usage and cross-checking that against the number of ways you can use it? Or are you just guessing? Because if you're just guessing, that's not a fundamental analysis.
Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by nostrademons on Tuesday December 18, 2018 @05:53PM (#18711369)
I actually am, to most of those points. I'm indexing every forum & blog on the web, then cross-referencing it with a big long list of crypto projects, then using that to infer which projects have an active, involved, growing community and which are scams or dead. Sentiment, activity, and viral mentions aren't a perfect proxy for adoption, but it's pretty close, and a lot more robust data source than technical analysis or previous prices, which by their nature will always be a lagging indicator.

Still in closed beta with a very limited set of testers - I've been working on a lot of the kinks in the data acquisition, which as you could imagine is a little tricky. If you want to get on the wait list, signup is here:

http://www.cryptolazza.com/

Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by JoachimSchipper on Wednesday December 19, 2018 @01:00AM (#18713433)
That is very interesting, but it does sound like you’re measuring engagement (or, less charitably, the effectiveness of the self-organizing distributed boiler room of patio11 fame) rather than, say, assets and earnings.

I’d call the latter “fundamentals”, if only because I don’t see why engagement would provide a price floor at e.g. 1000 $/BTC rather than at 10k $/BTC or 100 $/BTC?

Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by tim333 on Wednesday December 19, 2018 @12:10PM (#18717189)
While engagement as in just tweeting doesn't provide a price you can do an estimate by (number of fans) x (amount they'd like to have invested) / (number of coins) = (estimated price)

eg 20 million BTC investors want to have $10k in BTC, 20 mil coins available => the price should be $10k

obviously some guess work in there

Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by elhudy on Tuesday December 18, 2018 @06:04PM (#18711438)
That's fair; props on all of the effort you've put into that! I'd imagine one of the toughest challenges is subtracting the volatile "layman demand" and gamblers to get to that fundamental value derived from actual users.
Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by nostrademons on Tuesday December 18, 2018 @06:09PM (#18711477)
The nice thing about indexing the whole web is that you also have a complete per-author picture of everything they've written. Shills often post the exact same message on many different sites, or talk exclusively about one project and nothing else. Gamblers & laymen tend to pop up abruptly when a project gets hot, and then quickly either disappear or move on to the next hot thing. When a veteran user with a reputation for detailed, often-quoted posts starts talking about a particular project, though, and keeps talking about it, usually there's something real there.
Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by blackstrips on Tuesday December 18, 2018 @07:19PM (#18711883)
What are the chances those projects will generate income and pay for ETH?

Can those projects work without a blockchain? Would it be cheaper to just handle stuff with (digital) signatures and old fashion contracts?

Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by nostrademons on Wednesday December 19, 2018 @01:17AM (#18713509)
My investing thesis for crypto is this:

The cryptocurrency projects that will survive and thrive will be those that let people do things that are insane, and yet they want to do anyway. It'll be new markets where currently no economic activity is taking place, because the participants cannot trust each other (or make use of the legal system to trust each other) enough for any rational actor to consider transacting. Cryptocurrencies, of course, alter that rational calculus by letting you put trust in an anonymous network of worldwide miners to secure your transactions.

Think of some of the biggest companies created in Web 2.0. Facebook - who in their right mind would give Mark Zuckerburg all of their personal data so they could hook up with that hot girl in the dorm across campus? AirBnB - people actually invite strangers into their homes to stay with them? Even the founder and first investor call that "The worst idea that actually worked." Uber - you're going to get in a car with a complete stranger and pay them to drive you places?

(Interestingly, all of these are markets that are well-positioned to be disrupted by cryptocurrencies. The key elements of an EBay/AirBnB/Uber-type marketplace are 1.) ability to connect latent demand for a service with people who can supply that demand 2.) ability to receive payment and 3.) a reputation system to ensure that the service was performed reliably. Right now, the UI & scalability properties of crypto are not good enough to let them replace these centralized marketplaces, but all of the information involved could easily be stored on a blockchain rather than a database, and with current Ethereum transaction costs at 0.07c and most of these tech companies taking a 20% cut, at some point the economic incentives to replicate them will become huge.)

We've actually had a couple of these killer apps for cryptocurrencies - buying drugs off the Internet, and ICOs. In both cases, any normal, rational person looking at the behavior is going to be like "What? Are you crazy?" In both cases, people do it anyway, presumably because they really really want to buy drugs without having to meet in person, and because they really really want to invest in startups but otherwise can't.

If you can service a market with databases, digital signatures, and old fashion contracts, you should service it that way. Blockchain is useless for anything that people are doing now, because if they're doing it now, there's already a way to accomplish it.

Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by tim333 on Wednesday December 19, 2018 @12:47PM (#18717595)
Not sure I buy the insane bit. Airbnb, Facebook, Uber, drugs on the net and ICOs were proceeded by traditional B&Bs[0], showing holiday snaps, taxis, traditional drug dealing and traditional iffy investments. It's just new tech either did things better or allowed skirting of regulations, licences and the like.

I think there may be a future in security token offerings, a more regulated less scammy version of ICOs, kind of like traditional stock listings but cheaper and less Sarbanes–Oxleyed.

[0] https://www.streetdirectory.com/travel_guide/212726/travel_a...

Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by wmeredith on Tuesday December 18, 2018 @10:43PM (#18712841)
Sounds like you’ve invented coinrank. I dig it.
Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by JohnJamesRambo on Tuesday December 18, 2018 @05:27PM (#18711192)
6k was the bottom and then a black swan event happened, at the worst possible time, when Bitcoin Cash forked again and the owners of the two forks had to childishly sell tons of bitcoin to fund their hashpower war. I don't think that was a predictable event.
Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by airstrike on Tuesday December 18, 2018 @05:50PM (#18711352)
And which event is really predictable in the world of speculating?
Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by Animats on Tuesday December 18, 2018 @04:52PM (#18710917)
The article suggests they do arbitrage.

They blamed their failure on "increased competition for arbitrage opportunities." Properly, arbitrage is low risk exploitation of differences between two markets in the same thing. Cryptocurrencies used to look like they had arbitrage opportunities, with differing prices between exchanges. Mostly, that was because it was so hard to pry cash out of the underfinanced cryptocurrency exchanges. They always had some excuse for delaying paying out. Arbitrage requires the ability to move cash quickly from one exchange to another.

As that situation improved, the spreads between exchanges narrowed. Opportunities for low-risk arbitrage disappeared. These traders reacted to this by going into higher-risk forms of trading. Which is about typical for traders. It didn't end well.

Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by startupdiscuss on Tuesday December 18, 2018 @03:07PM (#18710061)
You need volume for liquidity which you need to execute the arb.

So, for instance, cryptoA trades for $99 on exchange 1, but $101 on exchange 2. Quick, buy on exchange 1 and sell on exchange 2!

Congrats you made $2 on each trade x 5 cryptos. That is hardly enough volume to make it happen.

Now imagine that the volume is so thin that after you bought the first one on exchange 1, the price gap closed, and by the time you sold on exchange 2 it collapsed.

Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by koolba on Tuesday December 18, 2018 @06:04PM (#18711437)
Don't forget to throw in significant counter party risk in dealing with fly by night exchanges. There's a non-zero chance that your entire balance held at an exchange disappears overnight.
Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by GreaterFool on Tuesday December 18, 2018 @02:31PM (#18709738)
I also hear arb is doing fine.

I wonder how it works though. Most exchange APIs are terrible and it is simply not technically feasible to do anything using those APIs.

Also prices at too many exchanges move in absolute lockstep.

Must be private, privileged access to better (internal) APIs, profit sharing, rip the face of "retail" investor kind of deals.

After all the "exchanges" are not really exchanges as we know them in other markets. They're more like forex trading.

I think the crypto market would benefit from a (semi-)regulated impartial exchange where other firms can become members and act as brokers to direct retail flow.

But unless there's some sort of regulation about it it would just be a private business that could kick out any member for any reason. To much risk for the members.

Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by module0000 on Tuesday December 18, 2018 @03:47PM (#18710382)
Forget the 3rd party exchanges with their homegrown APIs. If you are dealing with size(1k+), you go straight to the CME and trade crypto futures contracts. Dealing with the CME also adds the regulation necessary(ie CFTC) for your investors.
Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by polkapolka on Tuesday December 18, 2018 @03:27PM (#18710227)
You bypass the API completely and fill orders "manually".
Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by bhaak on Tuesday December 18, 2018 @04:30PM (#18710767)
What do you mean by "manually"?

The general web page of an exchange is horrible slow compared to their APIs and often breaks during high volume periods.

Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by polkapolka on Tuesday December 18, 2018 @04:39PM (#18710827)
Did you run into latency problems doing arbitrage this way? Or do you rely on API's?
Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by bhaak on Tuesday December 18, 2018 @05:02PM (#18710996)
I didn't do arbitrage but running bots for statistical data via API. I think latency is not even the big problem, although of course you have to account for it when doing arbitrage. The latency can be quite high even during calm periods (also don't forget that exchanges can be on different sides of the globe, that already adds quite a bit of latency).

During times of high volumes as in the period of the last rally, the websites of many crypto exchanges break easily and even the API are not that stable.

You can also see that on reddit. Every time the price spiked, you got people complaining about exchanges breaking. There's still a long way to go for crypto exchanges.

Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by lend000 on Tuesday December 18, 2018 @03:23PM (#18710197)
Any arb strategy requires holding reserves, so while the amount of the asset you can afford may increase by 4% per month, it's not a great business to be in if you're measuring from the start of a bear market. Same goes for a lot of stat-arb funds in the space, which are currently measuring their performance against the market, but which are significantly down YTD.
Re:Novogratz's Crypto Trading Desk Lost $136M in Nine Months by revel on Tuesday December 18, 2018 @04:53PM (#18710934)
I think you're completely right that his fund isn't just performing straight up arb. In fact, I don't think it can generating most of its P&L from arb since (from what I remember at least), Galaxy Digital simply has too much capital to put it to use with an arb strategy alone.

My guess is that his fund got locked into a levered position on one of the exchanges like BitMex for trying to trade too much then couldn't get out for 24 hours. This kind of thing is more common than you might expect. There are also very few exchanges that allow margin trading, leverage or shorting, and all the ones that do have, uh, questionable compliance practices.