You execute manually right? Click on a button? Then 250ms are irrelevant. You should anyway check into how you can get those spreads/baskets/combos executed via IB algorithm so the legs get filled for you
@GRULSTMRNN, it sounds like when they pulse the data, it has a smoothing effect. Most people care about trades which are not affected by pulsing, but OP cares about quotes.
250ms is forever if your arbitrage strategy relies on capitalizing on inefficiencies that disappear in a flash.
Yeah before IB i had 700 micro second connection to the broker, but i switched because of the commission. what do you guys think of optimusfutures.com?
Median reaction time of humans is around 270-280ms. Given the inspection and waiting time paradox, a reasonable estimate of the time passed between the occurance of a real time quote update and the arrival of an IB snapshot is around 250ms / 2 = 125ms. Hence, it takes a human more than twice as long to react to an occurring snapshot than the average delay caused by the snapshot. Screams to me like a candidate for automation if your strategy is so latency sensitive but you can certainly not convince me that this snapshot delay impacts manual order submission in any meaningful way.
I never tried to say that. I know for normal traders like trend or swing traders 250 ms is nothing. I do put "manual" "pending" order but i rely on tws to execute it. Now if the deviation last for less then 100ms, the order wont triggered. I dont wanna say i am going HFT (because they are much faster than me) but i am definitely faster than other retail traders.
I trade SPX spreads/baskets/combos and have not found an IB algo to improve execution. If you know of an IB algo that does this, please let me know.
Apologies if I was opaque. What I mean was a basket order where there are settings that dictate how each leg is filled. Example: short 300 AMZN, LONG 560 MSFT, filling both legs in 20% increments, meaning if 20% of MSFT get filled first wait until 20% of AMZN gets filled, and so on. Now, whether IB offers such algo I am not sure, I have not used them. But I implemented such basket order in my own architecture. Only downside with this order that rests on my own machine and not on IB servers is a) risk of disconnection or other disruption and b) for small order sizes each child order (lingo in DMA trading is parent order for the main basket order which then splits off child orders that get sent to IB and the exchange matching engine) will be charge the minimum commission rate at least which may in sum exceed the commission charged for the parent order was it submitted all at once. But as I generally submit larger orders point b is moot. Check with IB whether they offer such algorithm.