Credit Suisse's Dan Mathisson Doesn't Use Flashed Orders Except When He Does
A new article by Traders Magazine goes on to describe in exquisite detail how brokerages apparently have not only never used Flashed orders, but run like the plague any time the topic is even breached.
Quote Greg Tusar, head of U.S. electronic transaction at GSEC:
"We don't use them in the execution of client orders," Tusar said. "But
we believe it should be a matter of choice--that clients should have
access to them if they choose to. They should be available on an opt-in
basis, client by client."
Greg, while we have your attention, can you also elaborate on this little disclosure taken directly from your March 31, 2008 Street Smart Research Piece:
Not to make a fine point out of it, but isn't it a little presumptuous to say you don't Flash when your entire Sigma X Dark Pool is based on Flashing (granted Goldman is the only beneficiary of the Flashed order data)? If we are not understanding this correctly, please feel free to enlighten us on how we may have misunderstood this: SigmaX@zerohedge.com
Other professionals who chime in include none other than Jatin Suryawanshi, head of global quant strategies at Jefferies & Company:
"Flash orders have the potential to cause information leakage. It's information that was not available on a data feed
that's now available on a data feed." He added that the use of flash
orders, if it's not done purposefully to aggressively take liquidity,
may fly in the face of a broker's best-execution duties.
Well, at least one insider is happy to acknowledge the potential abuse of Flash. Uhm, why did it take all you guys years before coming forward with this knowledge? Does some blog have to disclose all your dirty laundry before you admit that a vast majority of your business models are based on "information leakage?" But then again Jatin's honesty may be attributed to his recent distraction, stemming from the escalating lawsuit with UBS alleging serious quant espionage (NY Civil Supreme, 600909/2009). And here we were thinking that only Sergey Aleynikov is in trouble.
Which bring us to Dan Mathisson, head of Advanced Execution Services at Credit Suisse.
"We don't flash," said Dan Mathisson, head of the Advanced Execution
Services group at Credit Suisse. "The whole reason we exist is to try
to execute [institutional orders] while leaking the smallest possible
amount of information. Flashing orders is not consistent with that
overall strategy." Credit Suisse said it does not send routable orders
from its AES unit to market centers, and therefore doesn't use flash
Yet a little further down, Dan provides this useful clarification:
At the same time, while many brokers don't flash orders, they do feed
flash message information into their smart order routers to improve
executions for their institutional customers. Mathisson of Credit
Suisse said his firm's smart order router reads the flash orders on
Nasdaq's and BATS's proprietary feeds. "That feeds into our heat maps,"
he said. "We may not directly respond to flashed orders, but for the
allotted time, which is some milliseconds, that destination would be
heavily weighted [if we have an order on the opposite side of the
So, Dan, just what happens if you have an order on the opposite side of the flashed order in the allotted time? Do you happen to... retract it? Do you happen to... cancel the order if you notice you are headed against substantial order flow in the allotted 30 millisecond. Pardon my abuse of semantics, but doesn't this statement flagrantly violate your bold claim of "We don't use flash." Let's see if I am getting this right - aiding and abetting information leakage is bad, but if someone else is Flashing (for which you are not at fault, presumably, it is really just evil Direct Edge [or, gasp, SIGMA X, assuming they ever route anything away from Goldman's prying prop hands] who should burn in hell here), you are more than happy to get access to the data and reroute order flow appropriately. Once again, I welcome your rebuttal of my weak attempts at reasoning at SigmaX@zerohedge.com (I am sorry Dan, but you will have to reroute your email packet with all the inbound Sigma X email traffic - I promise not to flash too much of it to the public markets).