Execution speed Limit order vs. market order

[2rosy]

I'm running day-trading quant strategies and I'm dealing with large orders and working to minimize impact. Speed of execution is critical to me I'm trying to get down to a few 1/10 of seconds. I'm wondering whether there would be a time difference in execution of

- a buy limit order at the ask + X ticks (assuming all shares are available at the time of submission)
- and a plain buy market order

It feels like in theory there should not be but Im curious whether it's actually the case or maybe it's insignificant (less than 1/100 second)?

What exchange? Are you asking the execution on the exchange itself after the order has been generated, passed through the broker, an onto exchange?

If the exchange is good, it might have a spec or metrics or even tell you. If the exchange sucks nothing prevents them from writing inefficient software

 

[rb7]

yes it's fully automated im actually tracking the time difference between when the tick that triggers my order appear and when i get my market order fullfilled so that difference also (mostly?) includes computing & network lag (I do a bunch of calculus in between) and I'm able to get a couple 1/10 of a seconds in average but it can vary to less than 1/10s to 1 second in some rare occasions. But good point i'll drill down to see where my lag usually comes from thank you!

Fyi, when I send market orders to the CME, through IB, I'm getting an average of 0.25 secs, between the time the order is sent and the time I receive the fill. Almost never above 0.3 secs.

 

[globalarbtrader]

I'm running day-trading quant strategies and I'm dealing with large orders and working to minimize impact. Speed of execution is critical to me I'm trying to get down to a few 1/10 of seconds. I'm wondering whether there would be a time difference in execution of

- a buy limit order at the ask + X ticks (assuming all shares are available at the time of submission)
- and a plain buy market order

It feels like in theory there should not be but Im curious whether it's actually the case or maybe it's insignificant (less than 1/100 second)?

There might be a very small difference in the execution of these orders on the exchange computer, but as others have said this is a rounding error in the context of your total slippage. If I was programming an exchange computer (which I have never done!) the obvious logic would be to convert a market order into a limit order with some arbitrary offset, so I would imagine the market order is slightly slower, but we're talking like 3 to 4 lines of C and maybe a few dozen clock cycles at most....

This is guesswork on my part though.

Rob