I was a quant prop derivatives trader at an investment bank.
The correlation between MtGox/USD  and GLD daily returns over 12 April 2011 - 30 March 2012 is 0.02; a linear regression produces a 0.151 beta (GLD daily returns independent) with a coefficient of determination of 0.0004. GLD had a period return of 14% with an average daily return (standard deviation) of 1.4% (0.1 percentage points); MtGox/USD had 531% with 10.2% (1.3 percentage points). It thus seems like it could have a place in a portfolio above (in terms of risk) small cap and emerging market stocks (on your worst day in GLD you lost 5.5%; in MtGox/USD 35.8%. Also, GLD delivered 2.7 times the return per unit of risk). Methodology note: since MtGox trades every day and GLD only on trading days, I used closing prices for trading days.
Bitcoin takes the monetary system back essentially a hundred years. We know how to beat that system. In fact, we know how to nuke it for profit. Bitcoin is volatile, inherently deflationary and has no lender of last resort. Cornering and squeezing would work well - they use mass in a finite trading space. Modern predatory algos like bandsaw (testing markets by raising and suddenly dropping prices), sharktooth (electronically front-running orders), and band-burst (creating self-perpetuating volatile equilibria in a leverage-sensitive trading space, e.g. an inherently deflationary one), would rapidly wreak havoc. There is also a part of me that figures regulators will turn a blind eye to Bitcoin shenanigans.
1) If you know how to nuke the bitcoin system for profit, are you doing so? It's not clear to me that you personally would be able to corner and squeeze as easily as you think, because I believe most bitcoins are held by a small group of people with a long term interest in a viable bitcoin economy.
2) HFT is usually justified by its practitioners as improving the stability and liquidity of the market, but here you characterize it as a fundamentally predatory tool if applied to bitcoin. What do you think about it in general?
That actually makes the actions of a serious predatory trader even more dangerous, since the total pool of available bitcoins is smaller due to the number of people just sitting on their bitcoins. In fact, it could be worse, especially if the trader causes the market to drop, which could make the investors bail out of the market, lowering the price of bitcoins even further. The trader then buys up the bitcoins for a song and a dance.
Now, if people weren't hoarding bitcoins, it would be much less of a problem. If the majority of the currency was liquid and actually flowing, it's much harder to corner the market on it.
This isn't necessarily HFT-specific. HFT is constrained by a lot of rules-both regulatory and exchange-specific-that attempt to stop this kind of damage(for example, NASDAQ has the right to void any trade chain that takes place if they deem it necessary), as well as a massive amount of competition from other players in the market. With Bitcoins, there aren't any of these rules, either on the exchange level or regulatory. There is not anything to stop a trader, either using HFT algorithms or just manually executing the trades, from applying these techniques.
1) I am not because of what you said - it's tough, more so from a data architecture standpoint (set up the accounts, data scrapes, etc.). Also, at 8 million USD 30-day dollar volume on MtGox, I'm not sure if there's enough depth yet to make it worth it (still toying with the idea, though).
A currency's value isn't a function of how much of it is outstanding but how much is transacted in it. If "most bitcoins are held by...people" holding on speculation, they contribute to holding its value down (don't think of stocks with dividends, think of fx). They also make the market smaller and less liquid.
2) The quote cycling HF market making algos use to discover prices disappears in a liquid market. If done in penny stocks, on the other hand, it would scare the hell out of everyone when a market order executes against a discovery bid way out on Pluto.
When these slips happen the HF guys get hurt - that's why they pull their servers causing the liquidity cuts they're criticised for. Thus, they work to predict when that kind of slip-up will happen. I'm reversing the logic and saying I'll force the slip-up and take advantage of the volatility right after, something with uncertain pay-off in liquid markets (you burn yourself in spawning the unstable equilibria) but more definite pay-off in illiquid ones.
Note that I wouldn't recommend trading Bitcoins at a high frequency for the same reason that HFs don't mess with penny stocks. The pay-off is too low compared with the infrastructure investment required. I'd probably also blow my cover with the exchanges.
EDIT: MtGox/USD returns 0.12 points per unit of standardised risk in a simplified environment; GLD 0.05 (both these numbers are very low). Thus, Bitcoin delivers 2.7 times the return per unit of risk as GLD; I got it backwards in my post. Sorry about that.
Or the silver markets . Difference is the SEC doesn't protect Bitcoin exchanges . Thus, while both roll back time in terms of the asset as a currency, only one rolls back the clock in terms of regulation. A loose analogy would be a cat-and-mouse game with primordial mice.
wouldn't it be difficult to cause such HFT havoc on a currency that is not that frequently traded (relative to big exchanges)? After all, someone has to be on the other side of all those HF trades. Also, since it's heavily decentralized wouldn't that make it more difficult to execute trades in high quantities?
How do you "corner and squeeze" a market that can't be shorted? You can manipulate GLD down by issuing naked shorts (essentially counterfeited against collateral). But you can't counterfeit Bitcoin. If you want to drop the market price of Bitcoin, you have to sell actual BC that you had before and don't have now.
Then, to actually profit, you have to find a way to buy those BC back for less than you sold for. This isn't going to happen unless your smart algo finds an equal and opposite stupid algo that it can rape. Who gets raped and why?
(If you ask me, it's BC that's a well-designed market and Wall Street that's a funky, broken anachronism. A lot of stupid pointless volatility would go away if we eliminated synthetic securities and lenders of last resort, and replaced the ancient order-matching system with a central limit-order book. Then, you'd have to actually add information to the system in order to profit - rather than profiting by raping other algos, and more plausibly retail traders, with "sharktooth and band-burst.")
No - seriously - bitcoin has nothing to recommend it! How many times must people who know what they are talking about poke holes in the bitcoin fantasy? Why are self-identified "tech" people clinging to bitcoin? If one is really hell-bent on designing an alternative currency learn from the problems that bitcoin has, and move on to something else. Ignoring the analyses offered by experienced traders, economists, and their like will result in more bitcoin-related mishaps.
There's no other decentralized e-currency. This means bitcoin is the only electronic payments and storage system where your funds are not subject to seizure by a central authority. You won't wake up one day and find your account frozen (cf paypal, bitcoins are safe in the blockchain as long as you have your wallet) or be totally helpless if the domain name of some fly-by-night startup goes down (cf e-gold, bitcoins are safe as long as there are nodes in the network).
As for the economic critiques, of course bitcoins are no panacea. These critiques are relevant to any market; the punchline of the article was even "I'd say the same about the Euro"
Right, so bitcoin (and its alternative blockchains). But the fact that there's already plenty of alt-chains that the market values at neglible fractions of a bitcoin shows that copycat crypto-currencies won't easily disrupt the "network effect" of bitcoin and dilute its value, as many detractors used to argue.