Search results

they were stacking the book with size in order to manipulate algo's and other traders who react to to size. i think every trader i know has done that at one point in order to try and attract some liquidity. looks like they made it their strategy though, or in the very least part of an order...

i'm referencing any relative value type trading with a market making bent. so, could be pairs, could be larger baskets, etc. i don't integrate s/r, but i'm sure you could.

i have nothing against holding metals, but that article is idiotic.

@risktaker... no, the unwind was precisely why intervention was needed. the carry-trade involves selling yen (borrowing it) for cheap, and using the money to invest in other assets (a lot have gone into treasuries). may 6th (flash crash) marked the beginning of the unwind of that trade...

carry trade unwind... writing's on the wall.

why does, 'the public' need to trade with low latency? have you ever attempted to quantify latency costs? for you? for the public? i'd recommend answering those questions first, before assessing fairness. (hint: for 99% of traders and the public in general, it's not significant.)

i agree with you... my point was the scope of algorithmic trading is huge and what people are supposing to be hft can be something as simple as the peg example i gave.

while not 100k/day, i've only had 2 losing days this year. a couple other guys i trade with are similarly consistent. we don't do dark-book arb or any of the other super low-latency stuff. it's pretty much vanilla market making on stat arb books. consistency in trading is about frequency...

let me ask you a question lights. please answer specifically. 1) exactly how should market makers (hft's whatever you want to call them) have handled may 6th with market orders coming down the pipe on ALL their stocks sweeping the book down 10-90%+ in seconds? 2) do you know what an ISO...

it wasn't an analogy. 87 is a great reference for pre-algorithmic market crash dynamics within recent memory. the dynamics don't change regardless of mechanics: if you stuff more than the market will bear, you will crash it. period. the recovery speed has little relevance, but since you...

liquidity exists because people have an incentive to provide it (profits). take away that incentive and force them to make a market at any cost 'for the greater good', and you no longer have a market. think about it for a second. you can't force altruism in place of basic personal survival...

this is bullshit. talk to traders who were around in 87... ask them how many stocks had a bid. the only difference between then and now is there was a human there to see the obvious idiocy of hitting a specialist quote 99% outside of the market. everything else was the same: excessive selling...

yep, exactly. the other side of the carry trade was corp debt. if you look up the corp debt names on the 6th, just like the yen, the move started much before the drop and really started throwing off the etfs, which in turn, which in turn... pointing the fingers at hft's is a cop out. the...

well, seeing as how this is the automated trading forum... the finer points of manual trader reaction times? really? talk about meaningless.

not too much different than how you wrote to file: double myData; ifstream infile("data.bin", ios::in|ios::binary); infile.read((char*)&myData, sizeof myData); @bigD: like pulling teeth?... care to show a simpler version using read()?

short answer: yes. ;-) google linux rtos extensions and network stack configurability and you'll have your answer.

@bellman yep, that looks good. buffer it based on resources... 2k ticks sounds fine though. you can also wrap it up in a low priority thread.

hi atticus, would be interested in hearing you reasons for wanting to be short dispersion vs long it in this environemnt? also, low betas, yes of course, but, low betas today != low betas tomorrow. so, how do you manage bankruptcy/merger risk? i would imagine flirting with this scenario...

if it's opening that late it sounds like the nyse. there's a lot of rules for that, you can call the nyse trade desk and ask... but the short answer is that they can hold the open until there's enough volume to offset whatever market orders are on their book, or if there's a crowd order...

other option is to start him on something small, and then build from it when you verify his work.