I am surprised people aren't talking about the interview on CNBC with the Virtu CEO. First, I don't know much about the company, but after the interview, one thing is clear[er] to me than every before - people misuse the word HFT. Here is the distinction that _I_ use between a MMing firm and an HFT firm: They are identical or close to identical in their leaning on extreme low latency tech. The difference is simple. MMers take home inventory. HFTs are flat EOD. This dramatically affects strategy. The CEO disclosed that essentially they are a market making firm, transferring risk from natural buyers to natural sellers, and acting as a middleman liquidity provider in thousands of instruments around the world 24x5. Sure, they make use of tons of technology behind the scenes, but imo there were three key statements made that everyone should take note of: On most days, they make 10% of the volume in GE. On one of their biggest days ever yesterday, they made about $1000. LOL! But multiply that by 10,000 instruments around the world a day... Over an enormous amount of volume, their win:loss ratio is 51:49. So this is Thorp's blackjack system applied to markets. They lose tens of thousands if not hundreds of thousands of times a day. How much of the profit in equities are due to rebates? Risk Management is the key to whole operation To me, the last one is the most interesting of all. Is the risk management done simply by the huge amount of volume that is statistically offset by each other in so many different instruments? Or, are they hedging at some interval with some future or index with some frequency? How much of the money they make are in the stock, or in the tax treatment of futures versus stocks? Do they take inventory? (real MMers do. Otherwise they truly are an HFT firm) I will leave you with a chart from a while back: http://www.zerohedge.com/news/2014-...irm-reveals-1-losing-trading-day-1238-days-tr
yes, I heard an interview a few months back, for them a $600 profit in GE is a good day. Almost half of their trades a losers. They want to be the tightest spread out there. They very ratrely have a losing day, like maybe 3 or 4 a year.
I know of several of them personally, but those are more arbitrage than MMing. Even Arbitrage can be HFT or non HFT, statistical or not. Most arb firms today are statistical. (Funny, I am pretty sure I was called by VIRTU, but I had very little interest at this point in my career) My own firm is statistical arbitrage leaning on extreme low latency tech (ELLT), although it can carry inventory therefore it is not HFT. But it is not MMing either. I don't post that stuff here. The point is that many styles of trading rely on ELLT and machines, and the term "HFT" is being misapplied all over the place.
That's earwash for their institutional shareholders. Obviously the key is their short-term α, without which there's no point being in business at all.
My firm has a name. It is an army of one, yours truly . I do everything, from raising cash, network administration, programming, strategy development, research, crying. Don't take it the wrong way, I just prefer to keep things to myself. As to firms that I know that do this sort of thing, I have worked for [some of] them and I feel it is not in my right to disseminate their business here.
virtu is just the digital version of the old market maker. For about a hundred years market makers have been both praised and villified depending on how it opened. If it opened up, not a word. If it opened down, the mm could face a stiff fine (because he was the one buying and unloaded when the buy orders started coming in at a profit.)
Yeah, perhaps so. I agree that these are all neutral words that can mean anything. The devil is in the details.