Forums - What really happened with BATS execution?

Forums (http://www.elitetrader.com/vb/index.php)
- Order Execution (http://www.elitetrader.com/vb/forumdisplay.php?forumid=30)
-- What really happened with BATS execution? (http://www.elitetrader.com/vb/showthread.php?threadid=239528)

Posted by Quickless on 03-26-12 01:41 PM:

What really happened with BATS execution?

We have some super sharp traders on ET. Would appreciate some more detail on what happened with BATS on Friday.

I look at it as market orders on the bid side and some big time problems on the ask side. There also had to be some kind of matching algo at work that when there was not a bid - ask match it adjusted the price.

Details from pros welcomed.

Anyone think that BATS was set up?

Posted by syswizard on 03-26-12 06:50 PM:

Likely caused by crappy Indian outsourced software....with a result just like what Marriot experienced:
http://www.thedenverchannel.com/new...287/detail.html

__________________
"Success in the markets runs counter to everything we have been taught...and if you want a friend on Wall Street, get a dog"

Posted by Options12 on 03-28-12 01:14 PM:

BATS Failure

Nanex Research suggests sabotage using an Intermarket Sweep Order:

http://www.nanex.net/aqck/2970.html

Posted by Options12 on 03-28-12 01:58 PM:

ISO's

ET Thread on Intermarket Sweep Orders and the flash crash:

ISO's crashed the market

http://www.elitetrader.com/vb/showt...=6&pagenumber=2

Posted by ronin266 on 03-30-12 04:56 AM:

If this is how Nanex shows, and I think it is. What where the motives to do that?. I mean, people want make money not halting stocks :/.

Maybe there is some connection between this and the lately antiHFT activity around. Anyway, if this is the case, thanks to the guys that performed this "hack". One more reason to review HFT practices and limit them

Posted by EulerFourier on 03-30-12 05:28 AM:

Quote from ronin266:

If this is how Nanex shows, and I think it is. What where the motives to do that?. I mean, people want make money not halting stocks :/.

Maybe there is some connection between this and the lately antiHFT activity around. Anyway, if this is the case, thanks to the guys that performed this "hack". One more reason to review HFT practices and limit them

http://www.zerohedge.com/news/skyne...-destroyed-bats

Posted by ronin266 on 03-30-12 07:17 AM:

Thanks.

Let the antiHFT wars begins.

Looks that BATS knew that something like this was possible and let it happen to reveal to the crowds what´s happening there, and get some support during the next IPO.
Far better decision than letting their bots fight against the others.
Anyway, as I said before, this will attract more public attention to the HFT problem.

Posted by syswizard on 03-31-12 02:25 PM:

re: "that article"....What a huge load of B.S.

1) The Waddell and Reed Flash Crash occurred at the CME because the stupid people there didn't have a limit on the number of contracts PER ENTITY. It was specified per TRADER. Had the proper order limit been PROGRAMMED (it's the software !!!), the orders would have been cancelled....and the flash crash would never have occurred.
In that case, they had good software, but people had established the "rules"....and the rules were faulty.

2) In the case of BATS, they thought they had sophisticated software, but they DID NOT. There was no circuit breaker logic.
Once again, the software was operational, but it was "dumb". The designers were at fault here. Also, where was the testing ? The testers were at fault as well. Finally this says nothing good about the BATS management team given the fact that EVERYONE in the industry is keenly aware of the power of the algos today. This was just a release of beta software, that's all....similar to Windows Vista. Microsoft was able to recover from that debacle, but can BATS survive after this one ?

__________________
"Success in the markets runs counter to everything we have been taught...and if you want a friend on Wall Street, get a dog"

Posted by Quickless on 03-31-12 04:49 PM:

I want to thank everyone who replied to this post.

The link to the Nanex Report blew me away. Can't say I understood everything going on in that report but the amount of detail was certainly impressive.

I had no idea so much was going on in the background.

Posted by Options12 on 03-31-12 05:50 PM:

BATS & Flash Crash

syswizard, what (if any) is the ISO role in these cases?

Posted by wilburbear on 03-31-12 06:15 PM:

Now the BATS website is down.

You know the powers that be are pissed off at an exchange that started in a strip mall in Kansas. They had much lower costs and greater efficiencies. They killed margins in equity trading on Wall Street.

Trading professionals have to drill down very deeply into the micro-structure of the market they trade. A BATS-enemy probably found a weakness and exploited it.

But who could pull this off? Any firm that could do this, ran it through their lawyers.

Or do you think, a firm with the expertise (and incentive) to hurt BATS left themselves wide open to be punished in some facet of this incident? No way.

In the end, it's gonna be one of those unsatisfying stalemates we see very often, lately. There won't be bright lines drawn around the exact players involved and their motives. And the SEC will let everyone off the hook without admitting or denying guilt.

Posted by Options12 on 03-31-12 07:05 PM:

BATS site down

Quote from wilburbear:

Now the BATS website is down.

Could be a repeat of the February DDoS:

http://www.securitynewsdaily.com/15...asdaq-bats.html

Posted by Landis82 on 04-01-12 05:01 AM:

Quote from ronin266:

Thanks.

Let the antiHFT wars begins.

Looks that BATS knew that something like this was possible and let it happen to reveal to the crowds what´s happening there, and get some support during the next IPO.

You are out of your mind.
The last thing that they wanted was to have to CXL their IPO and come back to the market place a second time . . . and (most likely) at a much lower offering price.

Posted by syswizard on 04-01-12 01:12 PM:

Re: BATS & Flash Crash

Quote from Options12:

syswizard, what (if any) is the ISO role in these cases?

I don't think there are any standards related to this type of software. I mean anyone could write a order-matching engine.

In the case of BATS, someone knew there was no circuit breaker logic. I wonder who put out the word ? Some underpaid programmer ?

__________________
"Success in the markets runs counter to everything we have been taught...and if you want a friend on Wall Street, get a dog"

Posted by Rationalize on 04-01-12 01:15 PM:

Re: Re: BATS & Flash Crash

Quote from syswizard:

I don't think there are any standards related to this type of software. I mean anyone could write a order-matching engine.

In the case of BATS, someone knew there was no circuit breaker logic. I wonder who put out the word ? Some underpaid programmer ?

I think you're overestimating the complexity, and adding too much conspiracy.

Route to somewhere less liquid, and it takes less volume to have more impact.

Posted by syswizard on 04-01-12 03:23 PM:

Re: Re: Re: BATS & Flash Crash

Quote from Rationalize:

Route to somewhere less liquid, and it takes less volume to have more impact.

Yes but liquidity is a huge factor and should have been taken into consideration by the software. I mean if the algos are offering at 10x the bid size at 1 tick under the last print, you've got to reroute or freeze those orders, right ? Otherwise, the bid drops, and then the algos cancel and replace at the next lower offer, and this continues until the bid is zero.
That's how it likely happened in this case. Now if the exchanges started to charge for cancel and replace orders, that would thwart this behavior as well.
And so it's happening:
http://www.tradersmagazine.com/news...s-109872-1.html

__________________
"Success in the markets runs counter to everything we have been taught...and if you want a friend on Wall Street, get a dog"

Posted by Rationalize on 04-01-12 03:29 PM:

Re: Re: Re: Re: BATS & Flash Crash

Quote from syswizard:

Yes but liquidity is a huge factor and should have been taken into consideration by the software. I mean if the algos are offering at 10x the bid size at 1 tick under the last print, you've got to reroute or freeze those orders, right ? Otherwise, the bid drops, and then the algos cancel and replace at the next lower offer, and this continues until the bid is zero.
That's how it likely happened in this case. Now if the exchanges started to charge for cancel and replace orders, that would thwart this behavior as well.
And so it's happening:
http://www.tradersmagazine.com/news...s-109872-1.html

If there's a gap in the book, then sell through what's above it, to buy back below it.

This is not fkn rocket science ...

Posted by Options12 on 04-01-12 03:38 PM:

Re: Re: Re: BATS & Flash Crash

Quote from Rationalize:

I think you're overestimating the complexity, and adding too much conspiracy.

Route to somewhere less liquid, and it takes less volume to have more impact.

Rationalize, do you think Nanex is correct in their analysis here:

So the use of an ISO order is not only inappropriate, but highly suspect. We have seen ISO orders abused like this before, such as the seemingly intentional halt in CSCO back on July 29, 2010 when AMEX first started trading Nasdaq listed stocks.

http://www.nanex.net/aqck/2970.html

Posted by Rationalize on 04-01-12 04:35 PM:

Re: Re: Re: Re: BATS & Flash Crash

Quote from Options12:

Rationalize, do you think Nanex is correct in their analysis here:

So the use of an ISO order is not only inappropriate, but highly suspect. We have seen ISO orders abused like this before, such as the seemingly intentional halt in CSCO back on July 29, 2010 when AMEX first started trading Nasdaq listed stocks.

http://www.nanex.net/aqck/2970.html

I don't follow their analysis.

If I send an ISO, I want to consume liquidly from all markets irrespective of best price. No fancy routing algos involved.

Nanex is saying the only liquidity was on BATS.

So any order would have arrived at BATS anyway.

Posted by ronin266 on 04-01-12 07:09 PM:

Quote from Landis82:

You are out of your mind.
The last thing that they wanted was to have to CXL their IPO and come back to the market place a second time . . . and (most likely) at a much lower offering price.

Time will show the truth wathever it is. I just give one variable.
Doubt that the price will be much lower.

Posted by Zr1Trader on 04-01-12 09:05 PM:

How A Nasdaq Algo Destroyed BATS

http://www.zerohedge.com/news/skyne...-destroyed-bats

Posted by ronin266 on 04-01-12 09:12 PM:

Quote from Zr1Trader:

How A Nasdaq Algo Destroyed BATS

http://www.zerohedge.com/news/skyne...-destroyed-bats

read the whole thread at least. The link was posted at the begginin lol

Posted by syswizard on 04-01-12 10:57 PM:

ISO orders can be devastating without circuit-breaker logic.
Well, that sums it up.

__________________
"Success in the markets runs counter to everything we have been taught...and if you want a friend on Wall Street, get a dog"

Posted by Options12 on 04-02-12 02:09 AM:

ISO's

Any reason other than algo-error or attempted sabotage for someone to deploy ISO's in such an illiquid market?

If algo-error, then the timing was phenomenally bad for BATS.

Posted by bwolinsky on 04-03-12 12:23 PM:

BATS Market Structure and plan for the IPO was inherently flawed.

If you consider most NYSE Brokers with seats wanted to see it fail, and probably so did NASDAQ Market Makers wanting to get there's, too, it's no wonder it would go to a half cent if their plan after the IPO was to give all of their balance sheet away that had just been raised.

Not only constant selling, but imbalances by Market Structure, the impetus was motivated by abusive short selling, and the fact that BATS publicly announced that the funds of the IPO in their entirety were going to be given away as a one time dividend to older shareholders.

This would grab anybody's attention, because their stock exchange had not been properly tested, but more importantly it was a target by much bigger players that would love to have seen this exchange fail.

Faulty wiring, and a flawed plan of what to do post-ipo lead to this debacle, and I'm personally undettered by the outcomes because I know how to write market structure so that things like this do not happen...at least, not so often that it happens the first day.

__________________
HOW MUCH IS ENOUGH?

Bud Fox

Wall Street

Posted by Options12 on 04-03-12 02:24 PM:

BATS and ISO Sell Orders

Quote from bwolinsky:

This would grab anybody's attention, because their stock exchange had not been properly tested, but more importantly it was a target by much bigger players that would love to have seen this exchange fail.

Couldn't regulators investigate the origination of the ISO sell orders and identify who engineered this attack?

Posted by Rationalize on 04-03-12 02:42 PM:

Quote from syswizard:

ISO orders can be devastating without circuit-breaker logic.
Well, that sums it up.

Any large order can move price without circuit-breaker logic.

It all depends on the liquidity on the other side.

Posted by Rationalize on 04-03-12 02:45 PM:

Re: ISO's

Quote from Options12:

Any reason other than algo-error or attempted sabotage for someone to deploy ISO's in such an illiquid market?

If algo-error, then the timing was phenomenally bad for BATS.

Someone wanted to sell first, sell at any price, and didn’t care which market.

Posted by Options12 on 04-03-12 02:52 PM:

Re: Re: ISO's

Quote from Rationalize:

Someone wanted to sell first, sell at any price, and didn’t care which market.

I would add: it appears from the Nanex report that they wanted to sell at any price in all markets.

The use of the ISO would ensure that all liquidity would be absorbed at any price. Therefore the price would be driven to 0.

Is this correct? Thanks in advance.

Posted by Rationalize on 04-03-12 02:57 PM:

Re: Re: Re: ISO's

Quote from Options12:

I would add: it appears from the Nanex report that they wanted to sell at any price in all markets.

The use of the ISO would ensure that all liquidity would be absorbed at any price. Therefore the price would be driven to 0.

Is this correct? Thanks in advance.

If the market sell order was bigger than all the buy limit orders in all books, then yes.

This is not rocket science.

Edit: Strictly speaking, the last trade would be at whatever the limit price of the last bid consumed was. So not actually 0.

Posted by Options12 on 04-03-12 03:10 PM:

BATS & Flash Crash

Good point.

It's all so similar to the flash crash reports that I am wondering if this attack could have occurred on another exchange post-flash crash or only at BATS due to bad circuit-breaker logic.

http://www.minyanville.com/business...2#ixzz1qzCyL2M6

But ISO trades can also be used for more nefarious predatory trading tactics such as combining ISO with short selling to suck liquidity out of an illiquid/troubled market or security, or fine-tuning an attack on a known weak hand or anticipated liquidation.

Both the NASDAQ and NASDAQ-BATS declared self-help against NYSE-ARCA in the minutes preceding the flash crash. Was that the equivalent of “blood in the water” alerting predators of weakness?

Posted by Rationalize on 04-03-12 03:14 PM:

If there's not a lot of strength on the bid side, then why would anyone post there?

Just wait, and buy later, at a lower price.

...

If there's no interest in a product at a store, come back later when it's on sale.

No conspiracy.

Posted by bwolinsky on 04-05-12 02:04 AM:

Re: BATS and ISO Sell Orders

Quote from Options12:

Couldn't regulators investigate the origination of the ISO sell orders and identify who engineered this attack?

I doubt it was anything coordinated. Just a bunch of good old boys ready to get shorties with 50 cals the new market couldn't possibly have handled.

__________________
HOW MUCH IS ENOUGH?

Bud Fox

Wall Street

Posted by bwolinsky on 04-05-12 02:21 AM:

Re: BATS and ISO Sell Orders

Quote from Options12:

Couldn't regulators investigate the origination of the ISO sell orders and identify who engineered this attack?

The other problem is that I've never heard of such investigations, and I believe they probably would have publicly announced this rather than blaming software.

__________________
HOW MUCH IS ENOUGH?

Bud Fox

Wall Street

All times are GMT. The time now is 02:52 PM.