You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
ONE HFT algo made up 4% of all quote traffic last week.
- Thread starter Grandluxe
- Start date
Quote from Random.Capital:
Anybody can have resting orders in place to catch algos trying to do this.
This is fundamental issue: resting where? There are many places - last number I have seen is 80 where they can rest.
The problem with those quotes is that they are posted and controlled locally and will be cancelled before any incoming order would cross them. It would have to be resting order at the exchange/ecn where this particular HFT operator is posting it and he knows that there are none. So it is risk free because he can cancel faster.
That is why once order is posted it should be in centralized queue and order cancellation should be queued as well with lower priority than executing orders. Due to liquidity and competition concerns multiple exchanges/ecn's should exist and provide final execution for these orders based on some predetermined protocol. So if they want to interfere they would interfere with their own quotes and would think twice before slowing down quotation system.
And what happened to trade-at rule and order internalization issue.
well, it's justified because on average it will only be a tiny minority of a firms actual marketable or "useful" quotes. exchanges allow the testing because these are the firms that make up most of the quotes in our markets. they're their biggest customers, and without the ability to get a good read on their network capabilities they wouldn't route there. i mean, this isn't happening all day everyday. hunsander's allegations that this is quote stuffing is idiotic. exchanges will tolerate port testing, but they're not stupid, they're not going to tolerate continuous dns attacks so one trader can screw everyone else. no one would do business there if they did.Quote from Occam:
If Hunsader is correct and this did constitute such a huge number of useless quotes, then how would "latency profiling" (or any other non-trading activity for that matter) justify spamming the entire market that badly?
One firm's testing process shouldn't impact so many others; if they need to test their systems, they should do so offline, or face significant penalties.
also, the only way you can test this stuff is live. latency is dynamic. if you're going to be floating hundreds of millions of $ in quotes, and you're latency/queue sensitive (which most market makers are), then a test port isn't going to cut it. that's why those patterns pop up.
the idea that this is "impacting so many others", eh, not really. i mean 4%? that's not even noise. the reality is, everyone whose business relies on them processing the whole feed, understands that the increase in quote volumes are a cost of doing business and allocate for it. it's the reality of automation in our markets and while that's a cost, the overall benefits are why you don't hear any firms up in arms about it. well, except nanex.
here's the deal, if nanex were proving manipulation is consistently taking place without a shadow of a doubt, then i'd be all ears. but, they're just not. the vast majority of the time they're posting charts of quote patterns they don't understand and try and infer foul play by default. just the fact that they post something, wide quotes, a big market order, a bunch of ms cancels, must be hft manipulators. it's a joke. no one with a serious business would print unproven allegations like that. and one you thing you never see are those types of emotional complaints from firms doing well, or, i would argue, with any shred of ethics.
Quote from propseeker:
also, the only way you can test this stuff is live. latency is dynamic. if you're going to be floating hundreds of millions of $ in quotes, and you're latency/queue sensitive (which most market makers are), then a test port isn't going to cut it. that's why those patterns pop up.
the idea that this is "impacting so many others", eh, not really. i mean 4%? that's not even noise.
Latency can be mesaured and provided by the exchange on continuous basis as "technical stuff" including how fast their MM queues are cleared, hardware characteristics and actual live performance in various segments of their network. There should not be a need for customers testing it during market hours live. They could get some base read off hours and test on dummy server and then trade live real quotes and money.
4% is one guy and is borderline noise add few more guys and it is not.
Nanex anti HFT stance is puzzling but maybe it is not unbiased as per your comments. Unfortunately everybody is talking their book and there is so much confusion about HFT including what it really is. It is used as buzzword to scare people and it works. So maybe it is time to do something smart about it without hurting HFT and the markets.
latency profiling does not automatically mean increasing latency for everyone else. it means, measuring the latency of a network. in this case, how a large qty of orders on a large qty of stocks affects internal/exchange networks.Quote from vicirek:
Although I see many advantages of HFT, latency profiling by increasing latency thus interfering with market information dissemination is illegal. They have no intention to fill the posted order and some of that activity could be classified as market manipulation by inducing trading activity by others or interfering with trading by other market participants.
I think that re-labelling certain activities to appear as legitimate is not the way of addressing HFT. It does not help HFT and honest discussion about it would help to stop bashing HFT.
as far as honesty, you have to ask yourself, if you repeat the allegation that there is market manipulation going on with these quotes, wihtout proof, without question, is that honest? just because nanex said it?
to me, it's obvious profiling going on by a larger firm testing a basket or portfolio they'll be quoting. just by the size of the basket, it would be a HUGE red flag to the exchange if there wasn't testing going on. i've gotten calls from nasdaq for much less. i do similar tests, much smaller, on single stocks, and from experience, that's what testing looks like. i'm not trying to spin things, it's honesty based on experience.
now, nanex, is claiming manipulation. i've not seen them show any proof or even a breakdown of where the manipulation happened. a barrage of quotes on a bunch of stocks, doesn't automatically suggest that. i've not seen them try and get any hft industry opinion on what happened, or for any of their past allegations for that matter. they just assume they're right, without proof, without industry opinion, without open debate, without peer review, and most times based purely on conjecture.
so, if we're going to talk about intellectual dishonesty, really, it should go both ways.
good idea. pitch it to the exchanges. currently these tools do not exist. hence the need and practice of firms profiling live.Quote from vicirek:
Latency can be mesaured and provided by the exchange on continuous basis as "technical stuff" including how fast their MM queues are cleared, hardware characteristics and actual live performance in various segments of their network.
Quote from vicirek:
It would have to be resting order at the exchange/ecn where this particular HFT operator is posting it and he knows that there are none. So it is risk free because he can cancel faster.
It would have to be resting - yes.
It would not have to be visible, and is therefore not "risk free".
You have access to the same toys they do.
Quote from vicirek:
There should not be a need for customers testing it during market hours live.
There sure can be. Latency is dependent on things like order of trade flow and mix of order types, amongst numerous other things. Only way to really know what's going on with parameters that matter to "you" is to regularly/continuously probe and measure.
Test servers are useless for this - it has to be a live port, during the actual trading hours of interest.
...currently these tools do not exist.
Yes, they do. They're just not especially useful, outside of giving general numbers.
Whether they are willing to share the measurements is, of course, a separate issue.
Yeah, Nanex is full of it, as usual, and just talking up hype and FUD since its good for their business... selling Fear. Think about it - how can you have manipulation without making a single trade? I mean maybe the rest of the market reacts to you (although I assume these quotes were far from the market price or they would have traded at least a little), but if you don't trade you don't make any money. Why would you take all that risk for no money? It makes no sense, and it seems much more likely that all these quotes at stupid low or high prices just got ignored by everyone... Except Nanex, for whom it was a slow news day.Quote from propseeker:
now, nanex, is claiming manipulation. i've not seen them show any proof or even a breakdown of where the manipulation happened. a barrage of quotes on a bunch of stocks, doesn't automatically suggest that. i've not seen them try and get any hft industry opinion on what happened, or for any of their past allegations for that matter. they just assume they're right, without proof, without industry opinion, without open debate, without peer review, and most times based purely on conjecture.
so, if we're going to talk about intellectual dishonesty, really, it should go both ways. [/B]