0
$\begingroup$

I'm using LOBSTER data to observe order execution. I'm taking the message data and grouping events by order ID to trace their life cycle.

I've noticed that, for example, if a buy order for 1000 is submitted at 100 and then that ID has 4 subsequent executions of 250 shares, the prices for those 4 execution events are all 100.

This raises the question: does this data not account for differences between limit order price and actual executed price? Does each event (other than submission) that is tied to a given ID just use the submission price?

My understanding is that the expected execution price (limit order price) vs the realized execution price can often be expected to differ.

Any thoughts? Cheers!

$\endgroup$
1
  • $\begingroup$ This is a known phenomenon. If you look at HFT papers that measure the intensity of MOs (how far from the mid price $d_{\text{buy}} = p_a- p_m$), they will sometimes decide to exclude data points when $d<0$, because it happens quite a bit in HF data. Funnily, when I was university, I used that exact data set and learned about this exact phenomenon for the first time too. $\endgroup$ Commented Jun 19 at 13:17

1 Answer 1

1
$\begingroup$

There are some senarios where what you described could happen:

  1. The buy order 100@1000 is a passive limit order, then it was executed when hit by the four aggressive orders with prices <= 100.

  2. There four passive limit ask orders each with 100@250 resting on the order book, and the buy order 100@1000 hit all of them.

In both cases the execution price for the bid order is 100.

There are definitely cases where the order price is different than the execution price. For example, if the ask side of the book is 99@1000, then the bid order with price 100 will have a execution price of 99.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.