Do you know anyone using ASUROUX who has never seen this behavior or that could verify otherwise? I've never addressed the broker as other people I have spoken to -- who use other brokers, have had similar behavior and attributed it to asuroux specifically. Other people in our office using it for over an year now, say it has always been that way. Lately Ive not been using it..so it isn't top priority, but will verify with DirectEdge/broker once it comes to it. I'll look up quik/quid for you, will post in weekend. Best regards
A big part of my work has to do with verifying fees charged/paid for ECN executions, and upon finding something like an EDGA flag W that charged a fee different than the one clearly stated in their website, I would have to give monies back to affected parties and fix the system bug that generated the erroneous charge. The problem is that liquidity flags are what tell a FIX system what fee to charge for a given execution. There are some exceptions to these rule, such as sub-dollar, over dollar fees; or ECN's that charge different fees depending not on the liq flag but on the order that was originally used (such as NYFIX); but this cannot be the case in the examples you cited, since the executions are done with the same order type and at similar prices. You could also have situations where an ECN charges 2 fees for one liquidity flag. They do this by submitting an aditional tag (usually 9621) that states the total monetary charge for the execution in question, but as far as I know only BATS and TRAC use this technology in the US.
Very interesting. So clearly then, from your experience, what I cited is in fact a broker's mistake? (IOW, not an issue of Direct Edge) Would you also happen to know, if it is possible to tell, or at least make a learned estimation as to what liquidity tag the EDGA family is gonna give you? I'm talking about remove-only with conditions such as in UYG for instance, very good liquidity, serial execution i.e. id say max 3 sec from one exec to another, distribution of tags seems random (I checked for 0 charge, and it came out roughly 40% of executions like I said. didn't check for tags though) IOW, if I send 10 orders one after another to ASUROUX is there any way of telling whether there will be more N/W or I/X? (notice: only in cases such as UYG or FAS nowdays)
Thanks a lot. I will start digging into the liquidity tag issue as well on EDGA I also need to see if we are configured to use their mulitple routing strategies. How fast is the turnaround where there is no liquidity? I know that EDGX is very fast, BATS a bit slower, and GETCO is slowest (out of the fast ones, however). My timeframes range - depending on the conditions - but large positions are typically 1 - 3 hours. We are trying as hard as possible to shorten the trades to several seconds -however the liquidity seems to be an issue. We dont really trade ETFs all that much for spreads, but we do a lot of ADDING in ETFs.
It sounds like a software bug to me... I don't think it is possible to tell which flag you're getting in advance (unless you add liq), since you can't see the ELP's, so you never know when you're going to hit them...
A good way to go when you're dealing with large positions is to spread them out across several pools... and then rebalance to give more orders to the ones that are giving you the most fills...
thats not a bad idea, we are already thinking of implementing something of this sort. right now we send to open venue - left over shares to another, leftover shares to another, until a fill. there is an inherent problem with the way we are doing it, so we are working on splitting up before and recombining later for the unfilled portions.
I was trying to read on up this. So GETCO is looking to hit the dark liquity been the spread right? and you say that it hit's it 1 in 4 times for dow 30 stocks? But on the BATS web site, it says if it can't hit the dark liquity, it then goes to the BATS book, which I'm assuming tries to hit orders on BATS. From how you guys are talking it seems like the order tries to hit the dark liquity, then if it doesn't it cancels. Can someone clarify what exactly this order does Thanks
Im not sure I understand what you are saying. Anyway - GETCO now hits for me with much greater rate and is becoming quite usable. I think what you are talking of regarding BATS is BATS Dark Scan which will go to DLP and then fill remaining form BATS book. Which is also quite usable for me - however I cannot ascertain that the amount BATS will fill from dark is greater or less that what GETCO will fill (even that on BATS site it says that GETCO is partner etc). So I still prefer to use GETCO in particular cases and rather not get filled than use BATSDS and and pay 25cent/1000shares on what is not filled by the DLPs. hope that helps.
Yes that does help. If the current b /a is 20.50 / 20.53 A buy mkt order what are all the possibilities where will GETCO fill? Does it go all the way up to the NBO including the NBO, or not including the NBO? Thanks