So bottom line, the people paid millions of dollars per year by banks to understand the market have to ask IT for an explanation of the market? Hmmm...
I was getting the same impression. Surely these traders couldn't be so dense?
This isn't even necessarily a HFT issue. If I go look at the quote for a stock on Google Finance, the only thing I'm completely sure of is that the value displayed is not in fact the accurate price. It's a reflection of the past.
What I don't understand is why the markets doesn't create a new order type, a delayed one that is created with a set trigger time. That way bulk sellers and buyers could set up big distributed trades in advance without having to resort to the same game as the arbitrage players are doing with locations close to the market and whatnot.
The biggest obstacle to this is that the Exchange's primary customers are the electronic trading firms. A firm like NASDAQ makes the majority of its revenue from co-location services and high-speed market data lines.
It's difficult for a new exchange to break into the space (or an existing one to go out on its own) with rules that disadvantage high speed traders. IEX at 0.9% market share for example has been working on solving the problems with a slow SIP feed for a couple years now.
I like these kids of ideas, but a bit of regulation couldn't hurt to speed things up...
Even if an order type like that existed, there's still the concept of "there's no such thing as a free lunch." A cost of doing business in trading is the spread between the bid and the offer. If liquidity providers take too much adverse selection, they'll just widen the spread tomorrow. So, sure, you help the liquidity taker today at the expensive the maker, and tomorrow the maker will increase his prices.
Why are these large investment banks, with quarterly revenues that dwarf every HFT player in the market, painted as victims? Since when is it unfair to be fast?
The exchanges offer co-location sites with access to GPS clock synchronization. Market players could easily coordinate their moves.
I hope I didn't give the impression that banks are the victims here. The situation most fixed income departnents are in is entirely their own fault because they haven't invested in their people or technology. Heck, most don't even talk to their equities departments about how they adapted 15 years ago.
> Heck, most don't even talk to their equities departments about how they adapted 15 years ago.
I'm curious to know who you are referring too here. As far as I'm concerned, FI desks have always been more forward thinking than equities, both asset managers and FI brokerage houses.
When an algo puts a bid in and pulls it quickly if the bids start getting hit. It makes the order book seem thick but if there is any decent order flow there is no liquidity to be found.
If orders are pulled the moment someone actually tries to match the market price, it's a clear indication the parties putting up the offers were not willing to accept the price they had put up.
I think that's fair. It would mean much less liquidity available though as your risk taking capability was reduced. Would traders actually be willing to accept the much thinner books if some kind of economic or regulatory enforcement of this was in effect? Maybe if they knew it was 'higher' quality liquidity and they could depend on it being there.
On the one hand, if I really wanted my resting orders to fill, of course I would list them at all possible venues, and once it filled in one place, I would remove the order with all haste at the other venues.
On the other hand, if I really wanted my orders to fill against others' resting orders, I would time the sending to each venue so that they arrived at roughly the same time; if I'm in Chicago, that means send to the New York venues first, wait about 20 ms and send to the Chicago venues. Staggered sending should work pretty well as long as venues are far enough apart; I don't know how many venues are located in the same city.
21 comments
[ 4.5 ms ] story [ 72.4 ms ] threadThis isn't even necessarily a HFT issue. If I go look at the quote for a stock on Google Finance, the only thing I'm completely sure of is that the value displayed is not in fact the accurate price. It's a reflection of the past.
It's difficult for a new exchange to break into the space (or an existing one to go out on its own) with rules that disadvantage high speed traders. IEX at 0.9% market share for example has been working on solving the problems with a slow SIP feed for a couple years now.
I like these kids of ideas, but a bit of regulation couldn't hurt to speed things up...
I'm curious to know who you are referring too here. As far as I'm concerned, FI desks have always been more forward thinking than equities, both asset managers and FI brokerage houses.
Charge a cancellation fee for any orders cancelled under 150ms.
That way it's no longer advantageous to quote stuff or float out fake liquidity.
If orders are pulled the moment someone actually tries to match the market price, it's a clear indication the parties putting up the offers were not willing to accept the price they had put up.
On the other hand, if I really wanted my orders to fill against others' resting orders, I would time the sending to each venue so that they arrived at roughly the same time; if I'm in Chicago, that means send to the New York venues first, wait about 20 ms and send to the Chicago venues. Staggered sending should work pretty well as long as venues are far enough apart; I don't know how many venues are located in the same city.