Australian (ASX) Stock Market Forum

Alan Kohler's Take on High Frequency Trading

Don't know why I even bother arguing when I'm obviously right. Wasting my time trying to make people understand and all.

Oh, we are sooooooo lucky to have such an experienced and knowledgeable member to grace us and set us fools straight.

Imagine how lucky we will be when you start trading. Hope its in the futs. :D
 
TH, LOL. Made me spill my coffee.

SCM did say when he first joined ASF that he worked for an engineering firm in Sydney City somewhere. He posted this information in the NBN forum. Not sure if he was full time, part time, casual or doing a school work experience week :D

I posted that vimeo clip for a basic overall concept of HFT.

There was a great article a number of years ago in regards to this and how technology is changing the markets.
Some firms with their algorithms are/were making a few k at the open when the bid/ask orders were getting processed. I'll had a dig for it, but it's highly unlikely l'll ever find it. Fascinating stuff and faster is better in this instance.
 
Who cares

Lots of experts looking pretty ordinary.
You guys must be bored s$&tless.
 
You guys must be bored s$&tless.

Or, you know, some of us were kind of hoping to have a legit discussion about a trading strategy that constitutes half (or more) of global all-market volume and an even higher proportion of equity bourse volume? You know, considering it's coming to Aus in a big way very soon?

Not bored ****less enough to spew derailing stuff about HFT into a personal thread about value investing like you did, however.
 
Not bored ****less enough to spew derailing stuff about HFT into a personal thread about value investing like you did, however.

I'm gutted!
Argument pointing to a changing coal face is seen as derailing!
I thought---Thought provoking.
Dinosaurs-----
 
some of us were kind of hoping to have a legit discussion about a trading strategy that constitutes half (or more) of global all-market volume and an even higher proportion of equity bourse volume?

But whats the problem with it. The flash orders stuff is a beat up. In the futs when it happens its ignored if not its hit into to by another bot or big dick. It just doesn't work. especially with the ASX where the markets spends most of the time asleep anyway.

Order stuffing is the same. The instrument get to where its going anyway. Can be frustrating fighting for a fill but its just an evolution of the markets. And I'm talking about fills when you are an active trader where a tick or two per trade means big money. Not Joe retail who can get all the volume he needs with a hit on the Comsec market button.
You know, considering it's coming to Aus in a big way very soon?

The biggest effect I see is ironically the slowing down of markets yet the Muppets think its their enemy. The ASX top 20 and futs have greatly reduced volatility from pre bots days. To the point that most Prop now doesn't touch it. Its arbed to within an inch of it life. Its already here.
 
I'm gutted!
Argument pointing to a changing coal face is seen as derailing!
I thought---Thought provoking.
Dinosaurs-----

No issues with the "argument", just where you decided to dump it and your attitude towards people in the "argument".

But whats the problem with it. The flash orders stuff is a beat up. In the futs when it happens its ignored if not its hit into to by another bot or big dick. It just doesn't work.

Right no problem, I wonder why real sugar traders complaining sugar contracts not representative since ICE killed their "Implied Engine" in an effort to attract HFT?

Order stuffing is the same. The instrument get to where its going anyway.

lol, yeah instrument gets to where it's going, like the BATS IPO which went to 0?

No issues with order stuffing except all the NYSE "we will cancel trades below/above 30% flash crash/smash"? Like NASDAQ, Direct Edge, Deutsche Borse, Borse Italiana all moving to penalise quote stuffing? Like the SEC finally moving to regulate HFT code which can bring down a bourse?

http://www.ft.com/cms/s/0/8d3aead4-689b-11e1-b803-00144feabdc0.html
Though rapid quote cancelling is seen as a way for marketmakers to minimise the risks of offering prices to other traders, keeping trading costs low, there is growing concern that a portion of such traffic is inefficient and raises costs to other investors who must invest in greater data bandwidth.
Nasdaq’s measures will penalise traders that send over 1m messages per day – which will include quotes, cancellations and replacements – but generate fewer than 1 trade per 100 messages.
The penalties will initially range from 0.01 cents to 1 cent per trade, with higher charges for messages further from the market price. This is called the weighted order-to-trade ratio. The charges could change, however, if they prove not be effective at curbing messaging.
“We don’t expect there to be a large number of customers affected, but it should cause people to focus on efficient quoting behaviour,” said Eric Noll, executive vice president at Nasdaq OMX. “In an ideal world, no one would pay anything.”

Can be frustrating fighting for a fill but its just an evolution of the markets. And I'm talking about fills when you are an active trader where a tick or two per trade means big money. Not Joe retail who can get all the volume he needs with a hit on the Comsec market button.

I am wondering if you were trading the ES or USDJPY on May 6 2010 like I was during the NY session. If not, I am not sure you really understand until it happens to you. I didn't lose money that night, but it opened my eyes.

The biggest effect I see is ironically the slowing down of markets yet the Muppets think its their enemy. The ASX top 20 and futs have greatly reduced volatility from pre bots days. To the point that most Prop now doesn't touch it. Its arbed to within an inch of it life. Its already here.

I am talking specifically about comparative trading volumes, 10% for the ASX versus 60-70% for NYSE, Euronext, LSE, etc. When I say 'coming' I mean the increase of this participation which moves in line with other major developed world markets.
 
Right no problem, I wonder why real sugar traders complaining sugar contracts not representative since ICE implemented their "Implied Engine"? As FT pointed out at the time

http://www.ft.com/cms/s/0/05ba0b60-33d8-11e0-b1ed-00144feabdc0.html
And, as you know the world is stuffed full of funny money looking for a place to put it. Just because it is moving more does that mean its solely the fault of bots?

No issues with order stuffing except all the NYSE "we will cancel trades below/above 30% flash crash/smash"? Like NASDAQ, Direct Edge, Deutsche Borse, Borse Italiana all moving to penalise quote stuffing? Like the SEC finally moving to regulate HFT code which can bring down a bourse?

http://www.ft.com/cms/s/0/8d3aead4-689b-11e1-b803-00144feabdc0.html
And just like all other regulatory moves associated with the market they fix the "problem" long after the "market" has moved on from it. No problems with regulating dangerous code that F's up a market but thats exactly the same problem they had to deal with when electronic access started 20-30 years ago.

I am wondering if you were trading the ES or USDJPY on May 6 2010 like I was. Thankfully I was short USDJPY but trying to get out of the trade was fkn hell.
Just another fun day. :D Come over to a thin and fast market like Honkers if you wanna know hell!!

I am talking specifically about comparative trading volumes, 10% for the ASX versus 60-70% for NYSE, Euronext, LSE, etc. When I say 'coming' I mean the increase of this participation which moves in line with other major developed world markets.

But thats the thing. Yes undoubtedly there has been stuff ups with "fatfinger" bots. But it use to be every problem was blamed on the evil short sellers. Now its the evil HFT. But they have a legit reason to exist. Its a market. We are in a Bear market of sorts and that leads to short and sharp moves. I cannot see anyway around regulating markets not to move when they are thin?

As for this original reason for the topic its all a beat up. Kohler is a muppet who took the existence of Co-location servers as the existence of evil flash order, front running, manipulating Bots. BS. They are Oppy arb bots, Insto arb Futs bots and Insto execution bots.
 
And, as you know the world is stuffed full of funny money looking for a place to put it. Just because it is moving more does that mean its solely the fault of bots?

All I was pointing out is that there are futs out there where what you said happens isn't happening, big dicks didn't ignore it or hit it, they stopped participating in the contract! They live in the real world, their access to production capital from the bank for the next year depends on a properly functioning futures market, if the contract series doesn't reflect reality then it is no good to them. If it's no good to the producers, the economic function it serves is more akin to casino than futs market.

So there are clearly issues with capital formation and efficient allocation of capital here.

I personally think the move by many long-only instos like pension funds to execute in dark pools indicates a similar issue.

And just like all other regulatory moves associated with the market they fix the "problem" long after the "market" has moved on from it. No problems with regulating dangerous code that F's up a market but thats exactly the same problem they had to deal with when electronic access started 20-30 years ago.

Respectfully disagree that this problem is anything like the one faced by regulators in 1982.

If you have no problem with the idea that there is potential for code to completely balls an exchange and that needs to be dealt with, then we agree in principle that an issue exists. I think the BATS IPO is a pretty good example.

Just another fun day. :D Come over to a thin and fast market like Honkers if you wanna know hell!!

That comparison with HK is sort of the point isn't it. The normal size sitting 10 pips on either side of the book in USDJPY is something like a couple of billion. Normally, those are real bids and offers sitting in EBS or Reuters3000 from very strong hands. Experiencing a HK type move in USDJPY shows serious systematic issues.

I cannot see anyway around regulating markets not to move when they are thin?

I would propose a very simple rule which is my understanding used to apply to the previous iteration of market-makers called "specialists". That is only this: if you get special exchange privilege (re latency, informational advantage, order types, etc) then you have to fulfil your role as a MM even in conditions aren't totally favourable to MM strategies. You don't wanna be a MM some days? Too bad! Take the trade or get the f off the floor.

May 6 2010 and plenty of other cases have shown that right now HFT supposedly fulfils the role of 'liquidity provider' when things are ok and gets paid good for it, but as soon as there is a sniff of trouble they suddenly become huge consumers of liquidity trying to flatten their book out as fast as possible, causing the most bizarre price action on the way.

That bizarre price action exhibits the characteristics of what we would call 'systematic instability' in compsci land. That the system being measured is exhibiting signs that indicate shifting underlying parameters beyond and point of equilibrium such that the system will either collapse catastrophically or become a new system.

What does the sound of liquidity getting sucked out of the market sound like?


You can hear near the end the 'human price' is reached and JPM steps in a big way to try and hit any offers it can find, shame NYSE and CME cancelled all those trades the next day 'cos that would've been their whole month right there.

As for this original reason for the topic its all a beat up. Kohler is a muppet who took the existence of Co-location servers as the existence of evil flash order, front running, manipulating Bots. BS. They are Oppy arb bots, Insto arb Futs bots and Insto execution bots.

I don't disagree, but happy to use the opportunity to put a different view forward.
 
Last edited by a moderator:
All I was pointing out is that there are futs out there where what you said happens isn't happening, big dicks didn't ignore it or hit it, they stopped participating in the contract! They live in the real world, their access to production capital from the bank for the next year depends on a properly functioning futures market, if the contract series doesn't reflect reality then it is no good to them. If it's no good to the producers, the economic function it serves is more akin to casino than futs market.

So there are clearly issues with capital formation and efficient allocation of capital here.

Just on that point about futs not "working" I have read many articles about how the markets are being pushed and pulled at the expenses of poor ole Mum & Pop yet I have seen the opposite since bots came in. Over my many many years of my futs trading :cool:. Just seriously deep deep order books. As a result when looking at intraday moves they have locked up. With our very own SPI arbbed into almost non existence. After the first 30 min you start getting 5 min bars with a 3 tick range! FFS!!!! Thats what you call locked. And the fav of Mum & Pop, the "blue chips", have followed the same pattern. At the expense of the daytraders.

Even the almighty ES just chews along now days with very little range. Nothing at all like the nice sweeping 1-5 min moves of 10 years ago. I would be more than happy to have all those friggin bots booted out and we all go back to ripping off the dumb traders at insto desks. But after 2008 I said good bye to those good old days and now just trying to again adapt.

On the Sugar contract to take the range in ticks when it was at $8-9 bucks then take the tick range when its moved up to $36 dollars is a little silly. And the hedges and the produces making those nice sugary diabetic inducing products may be p!ssed off that they have to trade a contract now above 30 but the farmers who are in the real world are more than happy to have it there I bet.

Again the "efficient allocation of capital" or inefficient allocation of capital more like it is not the fault of bots IMO rather a much much bigger problem.
 
So effectively everyone who places a trade when there is a HFT action in their direction will win out. Whats the problem then?
And no, they can't "see" a trade before it comes through to the market. They can be fast but not go back in time.

No you're wrong about that. They in fact can see the order before everyone else, hence "see the future". That's why people are up in arms about flash trading don't you get it?

For example on the asx they have a matchtrading system with bids and offers on multiple order books, signals are flashed between the systems to a fill matched to the best price.

eg 1.
Best bid Best offer
TradeMatch 4.05 4.10
Chi-X 4.02 4.06
PureMatch 4.02 4.07

But when the new order comes and you have a fast enough system, its possible to get a fill at a lower price and sell into the higher price. Its all a matter of latency. So when Joe 6 pack or even institutional investors who don't have access to HFT bots want to buy at the lower price, they would find they can't get a fill and have to move into a higher price. Hence you get slippage. (this is all happening in milliseconds btw)

The situation is even worse in volatile conditions like in a buying or selling frenzy where you can get bids above the offer prices or offers below the bidding prices. HFT bots can make tremendous amounts of money just from the spreads.

And this is just one way of doing it. I've seen other ways HFT have the advantage over the public, like using midpoint crossing trades with darkpools sell into rallies without public buyers knowing. This isn't like the dumb limit sell like we have either. They typically have standard deviations built into the system so it even MOVES with the rally, ensuring they get a good price. How come we can't do that? And if the direction goes against them, they will throw up a fake bid wall to manipulate the market into believing there's buying pressure. But when it actually get to them in the queue, they'll pull the order and put it one step back, letting fools who fill in after them take on more of their sells.
 
No you're wrong about that. They in fact can see the order before everyone else, hence "see the future". That's why people are up in arms about flash trading don't you get it?

No flash orders on the ASX. Never was. Happy to be proven wrong though so I can be up in arms over it.


For example on the asx they have a matchtrading system with bids and offers on multiple order books, signals are flashed between the systems to a fill matched to the best price.

eg 1.
Best bid Best offer
TradeMatch 4.05 4.10
Chi-X 4.02 4.06
PureMatch 4.02 4.07

Orders are not 'flashed' between them. It will be a consolidated market depth so it'll show 4.05 bid, 4.06 ask. All aussie brokers sent out a best execution policy a while ago dealing wit this specific issue. Non crossing volume is still 99%+ TM though.
But when the new order comes and you have a fast enough system, its possible to get a fill at a lower price and sell into the higher price. Its all a matter of latency. So when Joe 6 pack or even institutional investors who don't have access to HFT bots want to buy at the lower price, they would find they can't get a fill and have to move into a higher price. Hence you get slippage. (this is all happening in milliseconds btw)

The situation is even worse in volatile conditions like in a buying or selling frenzy where you can get bids above the offer prices or offers below the bidding prices. HFT bots can make tremendous amounts of money just from the spreads.

And this is just one way of doing it. I've seen other ways HFT have the advantage over the public, like using midpoint crossing trades with darkpools sell into rallies without public buyers knowing. This isn't like the dumb limit sell like we have either. They typically have standard deviations built into the system so it even MOVES with the rally, ensuring they get a good price. How come we can't do that? And if the direction goes against them, they will throw up a fake bid wall to manipulate the market into believing there's buying pressure. But when it actually get to them in the queue, they'll pull the order and put it one step back, letting fools who fill in after them take on more of their sells.

Agree with most of this, and you raise a point with a example of a sell order that sells into rallies and spoofs bids.
That would be more of a regular algo than a HFT.
 
Gee Iro you have been busy. Shame you didn't manage to understand this post,

Gee I wonder why. You never answered my very specific questions, didn't provided any helpful information, and were very condescending so I moved onto other sources thanks.

And how did you see what I posted before I edited it? Are you actually a mod in disguise?
 
No flash orders on the ASX. Never was. Happy to be proven wrong though so I can be up in arms over it.

Heh they may not call it that but I understand you can still pay asx a hefty fee to have access to order book level information, hence be able to see orders develop before everyone else and "see the future". It doesn't have to be like the Nasdaq style flash orders where for 30 milliseconds the prices are frozen while they try to find a seller/buyer on the other side following a quote. On the press releases asx say they run some sort of "preorder filter" through their order books to prevent "market manipulation and front-running" but this seems more like a PR exercise to booster retail investor's confidence in the integrity of the market rather than any real effort because similar claims have been made by other exchanges with very little difference and if they really want to put a stop to this sort of behavior they could simply impose a minimum size threshold to screw up their algorithms and that would make this sort of behavior uneconomical. As long as there's differences in and between the orderbooks, HFT firms will find a way to profit from it, and I doubt asx will do much about it since there's a conflict of interest in that HFT firms are such a gold mine in revenue.

I guess the real proof is in the pudding, in that HFT firms around the world are still in very much in profit even after all the negative publicity after the flash crash and all that rhetoric about putting bans on flash trading. It just means the firms aren't using strategies that take advantage of the 30 millisecond window and have moved onto something else. Another thing to consider is that even though the flash crash happened in 2009, on asx statistics since then the average order size have been going down while the actual number of trades have been shooting up (just like before), suggesting even more players have been hopping on this bandwagon and that the volume is big enough to noticeably influence the asx statistics on ALL trades.

Orders are not 'flashed' between them. It will be a consolidated market depth so it'll show 4.05 bid, 4.06 ask. All aussie brokers sent out a best execution policy a while ago dealing wit this specific issue. Non crossing volume is still 99%+ TM though.

By consolidated market depth 4.05 bid, 4.06 ask, I assume you mean the NBBO? That's just the aggregate. Orderbook level pricing occurs before this phase and in fact alot of HFC strategies involve manipulating the orderbook bid/ask price to influence the NBBO.

Agree with most of this, and you raise a point with a example of a sell order that sells into rallies and spoofs bids.
That would be more of a regular algo than a HFT.

I would say what you define as regular algo bots and HFT bots are basically the same thing. Since they both manipulate the prices before humans can react, HFT bots have algorithms too, algo bots benefit lower latency setups and they're not mutually exclusive imho.
 
Gee I wonder why. You never answered my very specific questions, didn't provided any helpful information, and were very condescending so I moved onto other sources thanks.

And how did you see what I posted before I edited it? Are you actually a mod in disguise?

I wasn't condescending. I did answer it with great detail. Then realised you had no clue and were miles away from having to worry about it anyway because you didn't even understand position sizing. So I then tried to correct your huge hole in the most basic of trading understanding.

But clearly I have wasted my time.
 
Top