I've been doing some research on electronic Limit Order Books (mainly equities) and I was wondering if anyone has seen a paper on how to compute competitive limit order prices. By competitive, I mean placing limit orders in a way that the probability of execution is 'relatively' high.

This, in my humble opinion, is purely a forecasting question using an online algorithm, and I was wondering if, in order to be consistent with the Order Book dynamics, one must also include a component to take into account this latter (something like probability of upward move-weighted averaging ..)

Thanks !

  • $\begingroup$ There are a lot of fancy sounding papers out like this: arxiv.org/pdf/1911.05892.pdf ... some of them give decent overviews of the basic approaches before the fancy steps in. I'm not sure how much "fancy" is really deployed. $\endgroup$ – mathtick Apr 3 at 12:31
  • $\begingroup$ Will take a look at it. I doubt that a Reinforcement Learning Algo would be efficient in this case though, RL algos need a big load of data which is not convenient in this case. $\endgroup$ – New quant Apr 6 at 11:20
  • $\begingroup$ Yes, ignore the RL stuff but the setup describes the inputs. If I remember correctly they barely use RL and basically do some agent-agent play to tune a single parameter via RL. $\endgroup$ – mathtick Apr 6 at 14:52
  • $\begingroup$ Thanks @mathtick , unfortunately it's not quite what I was looking for. I'm more interested in short-term online price prediction that would allow to place limit order in an efficient way. $\endgroup$ – New quant Apr 7 at 11:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.