Assume we look at all passive fills going through a security. How do we benchmark how good each passive fill is (relative to each other)?

We expect the edge to be (assuming midprice is our theoretical value) to be the spread between midprice and our fill, and our expected profit would take into account adverse selection & market impact as well.

That being said, given the last 100 fills (all executed with the same 'edge'), what is a good metric to rank those fills?

Also -- how can we quantify adverse selection?

  • $\begingroup$ Difficult question. From a trader's pt of view the worst limit orders are the ones where you never got in (opportunity cost), the second worst are the fills where you got in but the price at close of day was even lower (adverse selection, negative news coming out for example). $\endgroup$
    – noob2
    Apr 25 '17 at 15:45
  • 1
    $\begingroup$ Without knowing what someone's intention was before the trade how could you know whether it was a quality fill or not and then rank it? All parties have different reasons for trading and different levels of urgency associated with each trade. In the absence of an arbitrage, there is no 'edge' on either side of the trade...unless you know what is about to happen next. :) $\endgroup$
    – amdopt
    Apr 25 '17 at 15:55
  • $\begingroup$ Well, there edge -- assuming the market is a martingale (and no adverse selection), then you would be collecting the spread as edge on every passive fill, no? $\endgroup$
    – homebee
    Apr 25 '17 at 16:39
  • $\begingroup$ @homebee I suppose if you can accept that assumption and you are a market maker. That's a stretch for me though. $\endgroup$
    – amdopt
    Apr 25 '17 at 17:15
  • 1
    $\begingroup$ @amdopt I'm asking this question from a market-maker point of view. $\endgroup$
    – homebee
    Apr 25 '17 at 17:54

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.